The Art of the Deal

Chapter 1: The Ordinary Account

At eight forty-five on Monday morning, Sophie Bennett was arguing about a delivery note with a man who had never seen the delivery note. He was a buyer at a small chain of garden centres, and he had been charged for six replacement tablets Northbridge Solutions had sent after a courier damaged the originals. Sophie opened the account, found the courier report, and explained that the invoice had been raised automatically before the credit was approved. She promised to send the corrected statement by lunchtime, then waited while he decided whether to believe her.

“You’re very calm about it,” he said at last.

“I’m trying to be useful,” Sophie replied. It was true, though not the whole truth. She was also watching the clock. Her meeting with James Whitmore, Northbridge’s commercial director, was in ten minutes, and she had spent the previous evening rehearsing what she would say about her next career step. She had prepared a short list of accounts she could take on. Meridian Retail Group was not on it. Meridian was not the sort of account anyone took on casually.

James’s office had glass walls and a view of the canal, which made private conversations feel faintly theatrical. He did not ask about her presentation. He passed her a folder and told her that Eleanor Shaw, the senior account director responsible for Meridian, was going to be away for several months. The renewal was due in fourteen weeks. Sophie would lead the account from that morning.

“Lead it or keep the wheels turning?” she asked.

“Lead it,” James said. “With support. Meridian is worth about £4.8 million in annual revenue to us. It is also one of our most visible contracts. If we retain it on poor terms, we can be worse off than if we lose it. I need someone organised, credible and prepared to listen.”

Sophie felt the pleasure first, then the fear. “Why me?”

“Because you know how to get answers from people who are cross. And because you’ve been asking for a bigger account.” James watched her take that in. “This is the bigger account. It is not a reward. It is a test.”

At her desk, Sophie read the contract summary twice. Northbridge supplied Meridian with store-support software, a service desk, field engineers and data integrations across 312 shops and three distribution centres. The current agreement had been signed for three years, with annual indexation and a service-credit schedule. Meridian’s renewal letter requested a substantial reduction in annual charges, tighter response and restoration targets, and greater flexibility to change the number of supported sites.

Sophie copied the requests into a spreadsheet and labelled the last column “Possible concessions”. She meant it as a working heading, but the phrase looked more decisive than she felt. She had not yet spoken to Meridian’s operations team, reviewed the incident history or asked finance what the contract actually earned. Still, a blank document had a way of inviting premature certainty.

Across the aisle, Oliver Grant, a junior analyst, was trying to get a report out of the service database. Rachel Singh from legal passed behind him with a marked-up contract under her arm. Martin Cole, Northbridge’s service-delivery manager, was on a call, explaining to someone that an engineer could not be in Leeds and Bristol at the same time. The company was not a single voice. Sophie knew that. She also knew that, in fourteen weeks, she would be expected to make it sound like one.

Chapter 2: The First Position

Meridian’s offices occupied two floors above a flagship shop near King’s Cross. Sophie arrived early enough to review her notes in the reception area, where a wall display showed live sales figures in carefully rounded millions. She had brought a service report, a price history and a draft agenda. James had advised her to ask questions before defending anything. She had written this at the top of the agenda in capitals.

Helen Carter, Meridian’s procurement director, greeted her with a brisk handshake. David Chen, operations director, joined them in a meeting room with a view of the railway lines. Helen placed a printed copy of Meridian’s renewal request on the table.

“We want a twenty per cent reduction in the annual charge,” she said. “We also need stronger service guarantees and the ability to flex the estate without a change-control exercise every time a shop opens or closes.”

Sophie nodded. “That’s a clear starting position.” She heard herself say it too quickly, as if clarity made the demand reasonable.

Helen explained that Meridian was under pressure to simplify supplier costs. Its retail estate had shifted since the pandemic: a handful of smaller sites were closing, while several larger formats were being tested. Procurement had been asked to find savings across every major technology contract. “We can’t pay for a fixed footprint when the business itself is changing,” she said.

Sophie opened her report. “The current service has delivered 99.84 per cent availability across the core platform, and the annual charge reflects the support coverage, integrations and on-site response model. A twenty per cent reduction would take us below the cost of delivering the service as specified.”

“Then the service as specified may not be what we need,” Helen said. “We’re not asking you to perform the same work for less. We’re asking for a different deal.”

It was a useful distinction. Sophie recognised it, but her attention had fixed on the percentage. She began listing Northbridge’s costs: the service desk, the regional engineer rota, the overnight monitoring, the third-party hosting fees. Each item was accurate. Together, they sounded like a defence of Northbridge rather than an enquiry into Meridian.

David leaned forward. “The headline availability figure is fine. What concerns me is recovery. In November, the Midlands distribution centre lost the dispatch interface for three hours. The system came back, but we had people reconciling orders manually for most of the afternoon.”

Sophie looked down at her report. It showed the incident as a priority-one event, restored within the contractual target. “The service report records a two-hour restoration.”

“The interface was restored,” David said. “The operation wasn’t. That’s the difference I’m trying to get at.”

Sophie asked whether the issue was the response time, the restoration time or the recovery process. David described the additional work his team had carried out and the lack of a clear status update during the incident. Helen watched Sophie closely. For the first time, Sophie wondered whether Meridian’s demand for stronger guarantees was less about punishing Northbridge than about being unable to explain operational risk to its own senior team.

“We can review the incident evidence,” she said. “And look at a more useful measure than platform availability alone.” Then, uncomfortable with the silence that followed, she added: “But we would need to price any material change.”

Helen closed the folder. “We haven’t agreed what material means, or what you would measure.”

Sophie said she would return with options. Helen said she expected a commercial response, not another account of Northbridge’s costs. David asked for the incident timeline and the recovery notes. The meeting ended after forty minutes without a price movement, an agreed service measure or a date for the next discussion.

On the train back, Sophie replayed the conversation and heard how often she had spoken. She had protected Northbridge’s position, but she had not established what Meridian needed to be able to do. A lower price might be a position; predictable operating cost, flexibility and confidence in recovery might be the interests beneath it. She opened her notebook and wrote: “Ask what they will trade. Do not assume the number is the problem.”

Then she added, after a pause: “Find out whether they would accept a smaller reduction for a different service design.”

Chapter 3: The Cost of Keeping It

James read Sophie’s meeting notes without interrupting. When she finished, he tapped the page where she had written “20%”.

“That’s their opening position,” he said. “It isn’t yet their reservation point, and it certainly isn’t our answer. What did they say they needed?”

Sophie recounted the changing store estate, the distribution-centre incident and the need to explain service risk internally. James asked what she had learned about Meridian’s budget cycle, who signed off operational changes, and whether the request applied to all suppliers or only Northbridge. Sophie knew the renewal deadline and the names in the room. She did not know the rest.

“You don’t have to solve the negotiation in one meeting,” James said. “But you do have to stop treating a demand as a diagnosis.”

He had brought in Priya Nair from finance and Martin Cole from service delivery. Oliver joined to take notes and pull the relevant data. Rachel, legal counsel, had sent her comments in advance and appeared by video for the first half-hour.

Priya displayed the current contract economics. Meridian’s annual revenue was £4.8 million, but revenue was not profit. Direct delivery costs, hosting, licences and allocated support came to about £3.75 million. The resulting contribution margin was just under twenty-two per cent, before central overhead. A twenty per cent price reduction with the existing service scope would remove nearly all that contribution.

“That doesn’t mean every discount is impossible,” Priya said. “It means a discount must be attached to a change in cost, risk or commitment. If we cut price and keep the same obligations, we are paying Meridian to remain our customer.”

Martin said the current agreement was expensive to serve in some areas and efficient in others. The engineer rota had to cover all regions even when site visits were rare. The service desk took a high volume of low-complexity requests, while a small number of integration failures created disproportionate disruption. He could redesign the model, but not safely by guessing.

Rachel’s voice came through the speaker. “Flexibility also has a cost. If Meridian can remove sites at short notice while Northbridge keeps the same dedicated capacity, the risk sits entirely with us. We can offer a quarterly adjustment mechanism, but it needs notice, bands and a minimum commitment.”

Sophie wrote down “reciprocal”. She had been thinking of concessions as things Northbridge might give to win the renewal. Rachel was describing them as exchanges. Better flexibility could be worth something to Meridian, but it would also require Meridian to commit to forecast accuracy, transition periods or minimum volumes.

James asked what alternatives each side had. Northbridge could reallocate some staff if Meridian left, but the transition would be disruptive and the lost revenue would matter. Meridian could extend parts of its existing arrangement, use another provider, or bring selected support functions in-house. None of those choices was effortless. Neither side was negotiating from a position of helplessness.

“Our alternative has a cost,” Priya said. “So does theirs. We should know both, without pretending either one is a threat.”

The team argued over service levels. Martin thought a guaranteed response time was feasible; a guaranteed restoration time for every incident was not, because Meridian’s own systems and third-party networks could be involved. David’s concern, Sophie said, was that the existing measures rewarded technical restoration while leaving operational recovery unmeasured.

They agreed to investigate a two-part service model: response and restoration targets for Northbridge-controlled issues, plus a defined incident-management process for wider failures. Service credits would apply only to failures attributable to Northbridge and would be capped. Martin wanted clear exclusions; Sophie insisted the reporting had to show the customer what happened, not merely why a credit did not apply.

When the meeting ended, Sophie had a list of tasks rather than an offer. She felt an unwelcome but useful shift in her thinking. Winning the renewal could not mean giving Helen every visible concession and hoping the numbers worked later. It would mean finding a structure both organisations could operate—and making sure each improvement had a counterpart.

Chapter 4: What the Numbers Conceal

For the next three days Sophie treated preparation as if it were a form of control. She asked Oliver for twenty-four months of incident records, then checked his classifications against Martin’s operational logs. She requested the invoices for third-party hosting and licences. She read the contract’s change-control clauses and marked Rachel’s comments in the margins. At home, she built a spreadsheet with three commercial scenarios and colour-coded the cells until the figures seemed almost reassuring.

Oliver found that the headline availability figure concealed a pattern. Most incidents were brief and resolved within target, but five had caused meaningful disruption to Meridian’s distribution or store operations. In three cases, the platform had been available while a connected process remained impaired. The contractual report did not capture the full recovery period.

“The data isn’t wrong,” Oliver said, “but it answers a narrower question than David was asking.”

Sophie thanked him and asked him to add a column for operational impact, with the source of each estimate clearly stated. She was learning that a useful report did not merely make Northbridge look good.

At their next internal review, Priya challenged Sophie’s first scenario. It offered a twelve per cent reduction in annual fees, paired with reduced on-site coverage and a two-year extension.

“What does Meridian give us in return?” Priya asked.

“A longer term.”

“A term is not automatically valuable if they can still remove sites whenever they want.” Priya pointed to the draft. “And we haven’t costed the transition between service tiers. You’re offering a discount before you’ve established the baseline.”

Sophie felt the old instinct to defend her work rise, then set it aside. She replaced the single percentage with a pricing framework: a fixed core charge, a variable site component, and a limited adjustment window each quarter. The fee would fall if the supported estate contracted, but only within agreed bands and with notice. Larger changes would trigger a review and a transition plan.

Martin was less comfortable with the service proposal. Sophie had written a target for restoration of priority-one incidents and an improved service-credit schedule. He pointed out that a numerical target could be misleading if an issue depended on Meridian’s network or a software supplier.

“Then define the clock,” she said. “And define what happens when the clock is paused. But don’t make the reporting so full of exceptions that David cannot use it.”

They settled on separate measures for acknowledgement, response and restoration, with a joint incident review for severe events. Northbridge would provide regular updates during a priority-one incident and a written cause-and-recovery report within five working days. The draft did not promise that every operational consequence could be eliminated; it promised clearer ownership and evidence.

Rachel reviewed the service-credit language and warned that credits were a remedy, not an insurance policy. Sophie agreed. She did not want a generous credit that concealed a weak service design, or a guarantee so narrow that Meridian would regard it as theatre.

James asked her to rehearse the commercial conversation. Sophie began with the cost of the current arrangement, then stopped herself.

“I’m explaining our position again,” she said.

“You are,” James replied. “Try asking what a workable change looks like to them.”

Sophie practised: “If we separate the fixed platform and integration costs from the variable site support, would that help you manage an changing estate? What commitment could Meridian make in exchange for greater flexibility?” She replaced “an changing” with “a changing”, and James let the correction pass without comment.

She also prepared questions about the twenty per cent: whether it was a target, a budget requirement or an instruction from the board; whether Meridian would consider a lower reduction in year one if the service measures improved; and what evidence David needed to trust those measures.

At the end of the session, James told her she had enough to take a proposal into the room, but not enough to present it as the answer. Sophie gathered the papers. Her carefully coloured spreadsheet no longer looked like control. It looked like a map of what she still needed to learn.

Chapter 5: A Better Question

Sophie returned to Meridian with no completed offer. She had brought a one-page outline, the verified incident analysis, and a table showing how different levels of site flexibility would affect price and delivery cost. She had also left several figures blank. James had advised her not to fill them merely because an empty cell made a meeting feel unfinished.

Helen glanced at the outline. “I thought you were coming back with a commercial response.”

“I am,” Sophie said. “But I don’t want to put a price against a service model you haven’t agreed is useful. I’d like to test the structure first, then send a priced proposal within two working days.”

Helen studied her, as if deciding whether this was discipline or delay. David opened the incident analysis. “This is the first report I’ve seen that separates technical restoration from operational recovery.”

They discussed the November outage in detail. Sophie did not argue that Northbridge had met its contractual target. She explained where the service desk had identified the failure, where updates had been delayed, and how the handover between the platform team and Meridian’s distribution staff had lacked a clear owner. David said the technical team had been responsive once engaged, but Meridian had not known who was coordinating the wider recovery.

“We can put a named incident lead in place for priority-one events,” Sophie said. “And set an update rhythm, even if the cause sits outside Northbridge’s control.”

“Would that be included?” Helen asked.

“For the contracted priority-one response, yes. A dedicated resource around the clock would be different, and I wouldn’t want to imply otherwise.”

That answer seemed to reassure David more than an unqualified promise would have done. He asked for the proposed reporting format and the definition of a priority-one event. Sophie wrote both down.

Helen returned to the price. The twenty per cent reduction, she said, was the ambition her leadership team had given procurement. It was not a simple instruction to accept any service cut. The group had to absorb a rise in logistics costs and needed predictable spend. “A reduction that disappears when the estate changes is not a saving I can defend,” she said.

Sophie asked whether predictable spend mattered more than the precise year-one reduction. Helen paused. “It matters alongside it. I still need a material reduction. I also need to show that any flexibility is not a blank cheque.”

The exchange made the interests clearer without making them easier. Meridian wanted lower cost and room to alter its estate; it would not accept new charges every time a shop closed. Northbridge needed a viable margin and enough notice to plan capacity. David wanted operational confidence that the existing service measures did not provide.

Sophie proposed separating the fee into a fixed platform and integration charge, and a variable site-support charge. Meridian could adjust the number of sites within agreed quarterly bands, with advance notice and a minimum annual commitment. If the estate changed beyond the bands, the parties would review the service model rather than trigger an automatic price cut.

Helen said the mechanism might work, but asked for the numbers. Sophie did not offer the twelve per cent she had written in her first draft. That figure had been built on assumptions about reduced coverage and an extension Meridian had not promised. Instead, she said Northbridge could model a material first-year reduction against the current fee if Meridian accepted the defined bands, a minimum commitment and a two-year renewal. The actual reduction would depend on the confirmed site baseline and the chosen coverage.

“So you’re not offering the twenty per cent,” Helen said.

“Not for the current scope, no. I am saying we can test a different scope and cost structure. If the reduction is the only acceptable outcome, we should say so now rather than trade figures without a basis.”

David asked whether the service-credit cap would rise. Sophie said Northbridge could review the cap in exchange for the revised incident process and a clear distinction between failures it controlled and those it did not. Helen wanted the credit schedule to carry real weight. They did not settle the detail, but they agreed that a credit would not replace the operational review.

When the meeting ended, Helen had not accepted a proposal and Sophie had not secured a renewal. They had, however, agreed that Sophie would send two priced options, with the assumptions and trade-offs stated plainly, by Friday. David would provide Meridian’s forecast of site changes; Helen would confirm the budget range she could take to her leadership team.

Back at Northbridge, Sophie sent James a short account of the meeting. She did not call it progress in the subject line. She wrote that the parties had identified a possible exchange: lower and more predictable pricing against minimum commitment and notice, with stronger incident ownership against a service model that could be costed. Then she opened the spreadsheet and began again, this time leaving room for Meridian to say no.

Chapter 6: The Shape of the Evidence

By half past eight, the meeting room at Northbridge Solutions had acquired the air of a small control centre. Priya had put the support-ticket data on one screen and the contract’s service schedule on another. Sophie stood between them with a notebook open, while James poured coffee nobody had asked for.

“We have three years of tickets,” Priya said. “Or rather, three years of records that people have agreed to call tickets.”

James glanced at her. “That sounds reassuring.”

“It isn’t meant to.” Priya enlarged a column of figures. “Meridian logged 1,842 incidents last year. Our system shows 1,316. The gap is mostly duplicate entries, requests sent to the wrong queue, and cases closed by Meridian before we could verify the fix.”

Sophie read the labels. Some discrepancies were plainly procedural; others were not. “Can we separate them without making the categories do more work than the evidence?”

“I’ve started. Of the 526 unmatched cases, 214 have a clear duplicate or routing explanation. For 137, there’s no reliable record beyond an email thread. The rest are mixed.” Priya paused. “I can defend the method. I can’t claim the result is perfect.”

“Good,” Sophie said. “We shouldn’t present uncertainty as a saving.”

James leaned over the table. “But the overall trend is favourable. Northbridge response times improved by eleven per cent.”

“Against our own clock,” Priya said. “Meridian’s clock starts when the store reports a problem. Ours starts when the case reaches the correct support queue. The difference matters.”

Sophie looked again at the two screens. In the earlier sessions, both sides had argued from headline figures: Meridian’s demand for a lower annual fee, Northbridge’s insistence that the existing price was already tight. Here, under the labels, were several separate disputes—what counted as an incident, who owned the first response, when the clock began, and what a timely resolution meant. They could not be settled by choosing the more persuasive total.

“Let’s make a reconciliation table,” she said. “Show the agreed records, the disputed records, and the reason for each dispute. Then we can talk about what the service actually needs, rather than whose spreadsheet wins.”

James gave a small nod. “And the cost position?”

Priya switched windows. “The account revenue is £4.8 million a year. Direct delivery costs are about £3.94 million. That leaves £860,000 before central overheads, which is an eighteen per cent contribution margin. Not eighteen per cent profit.”

“The distinction will be lost if we put it in a slide,” James said.

“Then we should put it in the slide,” Sophie replied.

The contract renewal was due in four months. Meridian had signalled that its board expected a reduction in operating costs, and Helen had repeated that expectation in the previous session without specifying a target. Northbridge had already offered to hold its annual fee at £4.8 million for a further year, provided the scope remained stable. Meridian wanted a lower price and clearer accountability; Northbridge wanted predictable demand and fewer avoidable incidents. Neither position was the same as the interest beneath it.

“There’s one more finding,” Priya said. “Seventy-three per cent of the tickets classed as integration failures arrived during the overnight stock update. But we can’t establish whether the fault sits in our connector, Meridian’s data feed, or the hand-off between them.”

“Can you establish the cost?” Sophie asked.

“Not cleanly. We estimate £96,000 a year in extra support and manual checking. There’s a range of about forty thousand either way.”

James whistled softly. “That would make a useful efficiency number.”

“It would make a useful hypothesis,” Priya said. “If we call it a saving before we know the cause, we’ll be arguing about a number we invented.”

Sophie wrote: validate before pricing. “We can propose a joint diagnostic, with access to both systems and a named owner on each side. If it finds a Northbridge fault, we fix it within the service plan. If the cause is outside our scope, we agree a change before anyone starts charging.”

“And if it finds both?” James asked.

“Then we allocate responsibility by evidence, not by which team has the louder voice.”

She could see that James was considering whether this sounded too cautious. He had spent the first two meetings defending Northbridge’s competence. Sophie understood the impulse. Yet confidence that ignored an uncertain hand-off would become brittle as soon as Meridian produced a different record.

“What do we show Helen?” James asked.

“The trend, with its definition. The gap, with its explanation. And the uncertainty, without pretending it is a concession.”

Priya smiled. “That may be the least glamorous slide deck I’ve ever made.”

“It may also be the first one they trust.”

Before the meeting, Sophie sent Helen a short note attaching the proposed reconciliation categories. She did not claim that the data proved Northbridge had met every obligation. She said it supported improvement in response times, while leaving the timing and ownership of a significant group of incidents unresolved. She asked whether Meridian would nominate someone to take part in a joint review.

Helen replied ten minutes later: “Rachel can join. Please keep this separate from the commercial offer.”

Sophie read the sentence twice. It was not a refusal. It was a boundary, carefully drawn. She forwarded it to Priya and James.

“Separate for now,” she said. “Not irrelevant.”

At eleven, David joined them to check the commercial model. He had spent the morning reviewing the assumptions behind the renewal price and had marked three lines in red.

“If we hold the fee and keep the existing scope, the contribution stays at eighteen per cent,” he said. “If we absorb the proposed extended weekend cover without changing the fee, it drops below fifteen. That is before any further volume growth.”

“Can we offer a lower fee against a narrower service?” Sophie asked.

“Possibly. But the saving must be real. Removing low-volume activity does not help much if the staffing rota remains unchanged.”

James folded his arms. “Meridian will say the new automation should reduce our cost.”

“It might,” David said. “If it works, if the data is clean, and if the teams stop doing the same checks manually. Until then, it is a forecast, not margin.”

The distinction settled into the room. Meridian’s likely request was for a visible price reduction. Northbridge could not make one safely by relabelling its existing margin, nor could it promise that an untested technical change would pay for itself. But the data showed places where both sides might reduce work: duplicated tickets, repeated checks, and incidents whose ownership was unclear.

Sophie drew three columns on the whiteboard: price, scope, performance. “We’ve been treating them as one argument. We should not. A price reduction needs a reciprocal change—less scope, more predictable volume, or a measurable efficiency. A service improvement needs an agreed baseline. And a concession on one dimension should not quietly become an entitlement on all three.”

“That’s tidy in the room,” James said. “Less tidy at the table.”

“Then we need to be clear before we sit down.”

Priya saved the reconciliation draft. “I’ll mark the records we can verify and the ones we can’t. No blended total.”

Sophie closed her notebook. The evidence had not delivered an easy answer. It had done something more useful: it had made it harder to mistake a position for a fact, or an estimate for a commitment.

Outside, the office was filling with the ordinary sounds of work—calls, footsteps, the lift announcing itself. Sophie looked at Helen’s message once more. Rachel could join the review. That meant Meridian was willing to examine the problem. It did not mean Helen was ready to move on price.

For the next meeting, Sophie would need to understand what the demand for savings was protecting, and what Northbridge could offer in return without giving away the service that made the contract worth renewing.

Chapter 7: The Number Behind the Number

Helen accepted Sophie’s invitation for a short call on Thursday, but asked that they keep the discussion “at working level”. Sophie understood the distinction. No formal offer, no recorded commitment—just enough room to say something that might later matter.

They met in a small room at Meridian’s head office. Helen arrived with a folder and her phone face down on the table. She was neatly composed, though a faint crease between her eyebrows suggested the morning had already made demands of her.

“Your note was more careful than most,” she said.

“The data is more complicated than most headlines,” Sophie replied.

“I appreciate that. I still need a commercial outcome I can take to the board.”

“Can you tell me what the board needs to see?”

Helen turned the phone over, then back again. “A reduction in the cost of the service, or a credible route to one. The business has a wider savings programme. Procurement has been asked to show results, not merely avoid increases.”

“Is there a particular figure attached to this contract?”

“There’s a target across the operating portfolio. It’s not a number I can allocate neatly to one supplier.”

Sophie heard the answer as both information and boundary. Helen had not said there was no target. She had said the target was not assigned to Northbridge. That could mean the figure remained under discussion, or that Helen had no authority to disclose it, or that she wanted Northbridge to make the first move.

“If we offered a reduction in annual price,” Sophie said, “what would Meridian be prepared to change in return? For example, a firmer volume forecast, a longer term, or a defined reduction in scope?”

Helen’s expression did not alter. “I can’t promise a longer term in return for a concession you haven’t made.”

“That’s fair. I’m not asking you to promise now. I’m trying to find out which forms of commitment are possible.”

“The business has not agreed to lock its volumes. Store activity changes quickly. A three-year commitment would be difficult.”

“Would a two-year term be possible if the first year had a review point?”

“Possibly. But I don’t want to trade a longer term for a price that remains above the benchmark.”

Sophie let the silence stand. Helen aligned the edge of her folder with the table. It was a small action, but it gave Sophie a moment to resist the temptation to fill the pause with a discount.

“You’re asking for a benchmark,” Sophie said at last. “Can you share what it includes?”

Helen looked directly at her. “The comparison isn’t like for like. One provider has quoted a lower headline fee. Their offer assumes a narrower support window and treats some integration work as chargeable.”

“So the headline is lower, but the total cost may not be.”

“It may be. We haven’t completed the assessment.”

“And switching?”

“There would be a transition cost. It is not impossible, but it is not free.” Helen’s tone remained neutral. “We do have a credible alternative. I don’t want you to assume that the difficulty of changing supplier means Meridian will stay.”

Sophie nodded. “I wouldn’t. Equally, I don’t want us to assume a lower headline fee means a lower cost to Meridian.”

For a moment, Helen’s mouth softened into something close to a smile. “That is why I’m having this conversation with you.”

The admission was modest, but it changed the temperature in the room. Meridian’s alternative was real; so was the cost and risk of using it. Northbridge’s alternative was not an empty threat either. It could decline a renewal that required wider coverage at a price below its delivery cost. Neither side could afford to confuse a viable walk-away option with a costless one.

Sophie opened her notebook. “Would it help if we separated the current fee from optional cover and from the work that might be avoided through better integration? Then we could compare the offers on the same basis.”

“Yes, if you don’t present every potential efficiency as money already saved.”

“Agreed. We would need a baseline and a test period. Any saving would be measured after the change, not promised in advance.”

Helen considered this. “That may help me show a route to savings. But the board still expects a result this year.”

“Would a capped service credit count?”

“A credit is not a saving if the underlying service is poor.”

“No. And we wouldn’t propose one as a substitute for fixing the service. But if the baseline is agreed, a limited credit could give Meridian protection while we test the integration work.”

“That would depend on the threshold, the cap, and what happens if a system outside your control causes the failure.”

“Of course.”

Helen closed the folder. “I’m not agreeing to a credit. I’m saying it could be considered as part of a complete package.”

“Understood.” Sophie wrote that down—not the possibility alone, but its limits. She would not take it back to Northbridge as approval.

When they returned to the savings programme, Helen’s caution became more personal without becoming confessional.

“I have to show that procurement has challenged the cost,” she said. “There are people who think renewal is simply the easiest option. If I recommend Northbridge at the current rate, I need to explain why that is better value than the alternative, and why the risk of transition is being priced sensibly rather than used as an excuse.”

Sophie recognised the pressure. “And if you recommend a change, you have to explain why the apparent saving justifies the transition risk.”

“Yes.” Helen’s answer came quickly. “I don’t want a failed migration on my record because everyone liked a lower first-page number.”

There it was: not a hidden wish for a discount, but a need for an account she could defend. The contract was important to both companies, but their internal audiences were different. Northbridge’s finance team would test margin and delivery exposure. Meridian’s board would ask for savings, accountability and evidence that alternatives had been considered.

“Then perhaps the package needs to give you two things,” Sophie said. “A defined near-term reduction where we can identify work that is genuinely removable, and a measurable route to further savings if the integration review succeeds.”

“That sounds more credible than a broad promise.” Helen stood, then paused. “But don’t send me a proposal built on best-case assumptions.”

“I won’t.”

Outside the room, Sophie called James. She told him that Meridian’s benchmark was lower but incomplete, the transition was feasible but costly, and Helen needed a defensible result this year. She did not say that Helen had accepted a service credit, because she had not. Nor did she say that Meridian would agree to a two-year term. Helen had described it as possible, not promised.

James listened. “What can we offer without moving the floor?”

“A narrow reduction tied to a defined removal of low-value checks, if Meridian provides a stable forecast and lets us validate the workload. Then a joint review of the integration issue, with a credit mechanism only if we agree the baseline and exclusions.”

“That is a lot of conditions.”

“They are reciprocal commitments. If we want the benefit of predictable work, we have to say what Meridian receives for giving it. If we want to promise a credit, we have to know what event triggers it.”

James was quiet. “Oliver thinks we should make the pressure visible. Tell Helen we can’t meet the benchmark and put the choice back on her.”

“Has he spoken to Helen?”

“No. He thinks she’s testing our resolve.”

Sophie looked back through the glass at Helen’s empty meeting room. Perhaps she was testing it. But the information she had shared was not a bluff; it was a map of the decision she had to take inside Meridian.

“I’ll think about the wording,” Sophie said.

As she ended the call, she knew she would have to choose between a posture that looked strong and a question that might reveal more. Oliver’s method had an appeal: it reduced ambiguity by forcing a decision. It also risked turning the first honest conversation into a test of nerve.

Chapter 8: The Cost of a Hard Line

Oliver Grant arrived for the Friday negotiation carrying a slim folder and the confidence of a man who believed the other side was always waiting for permission to take more.

“Don’t give them another workshop,” he said as Sophie reviewed the agenda. “They have the data. They have the alternative. They know our position. If you keep asking what they need, they’ll hear that we’re prepared to solve their problem at our expense.”

“We are prepared to solve some of their problems,” Sophie said. “Not all of them, and not without a return.”

“Then make the return explicit. Put down a number and make them react.”

James supported the approach. Meridian’s renewal timetable was tightening, and the board meeting was six weeks away. Northbridge had spent two sessions explaining its costs and another examining tickets. At some point, a negotiation had to reach an offer.

Sophie agreed to give Oliver’s proposal a try. She and David had modelled a reduction of £120,000 per year, but only if Meridian removed the extended manual verification work, accepted a two-year term with a review after twelve months, and provided a monthly volume forecast. The lower fee did not reflect a sudden improvement in Northbridge’s margin; it reflected work that would no longer be done and a more predictable delivery plan.

They put the offer on the table early. Sophie described the current annual fee, the proposed reduction, and the conditions. The figures were set out in a one-page schedule: £4.8 million falling to £4.68 million, before any optional weekend cover. A two-year term would begin on renewal, with a joint review after the first year. The scope change was listed in an annex, not buried in the price.

Helen read the page without comment. Rachel, beside her, marked the annex with a pencil. David watched both of them. Oliver sat back, hands folded.

“If the verification work is removed,” Helen said, “who takes responsibility for an incorrect stock feed that passes through the automated check?”

“The system owner, once the failure point is established,” Sophie replied. “The current manual review is not a contractual guarantee that every feed is correct.”

“But your offer removes a check we rely on.”

“It removes a process that has not been reliably reducing the incidents. The ticket evidence suggests that some of the same checks are being repeated on both sides.”

Rachel looked up. “It suggests that for some stores. It doesn’t show that the checks are redundant for all of them.”

Sophie heard the weakness in her own case before Oliver could speak. Priya’s analysis had identified duplication in some records, not proved that every verification task could be removed safely.

Helen set down the page. “I can take this away. I can’t recommend a two-year term in exchange for a reduction that transfers risk to our operations team.”

“The reduction is conditional on the scope change,” Sophie said. “We can’t reduce the price and keep the same work.”

“I understand the commercial logic,” Helen replied. “I’m not accepting the assumption that the work can be removed yet.”

Oliver leaned forward. “If Meridian wants the current scope at the current price, that remains available. But we cannot reserve capacity while the organisation compares us with a lower service model.”

He had spoken evenly, but the sentence altered the room. Helen’s gaze moved from him to Sophie. “Are you saying Northbridge is withdrawing the renewal offer?”

“No,” Sophie said. “The existing scope and current fee remain available. The reduced price applies only to the reduced scope and the commitments on the page.”

“Then we agree on that much,” Helen said. “But we do not have a basis for agreeing that the work can safely disappear.”

Oliver’s jaw tightened. “This is why suppliers end up giving concessions without receiving anything.”

Sophie wanted to interrupt. Instead she kept her hands still and let Helen answer.

“A lower fee is not a concession if we pay for it in operational exposure,” Helen said. “A two-year term is not free to Meridian. It limits our ability to use the alternative if the service underperforms.”

That was the minor point they won: Helen agreed that any comparison of the offers would list scope, service hours and chargeable integration work alongside headline price. She would not allow the competing quote to be presented internally as equivalent without those qualifications. It was useful. It was not an agreement on value or volume.

The meeting ended with the offer still open and a request for more detail about the removed checks. As they packed their papers, Rachel gathered the marked annex and said, “We should look at the overnight feed before we decide what can be cut.”

Sophie nodded. “That makes sense.”

Helen’s reply was cooler. “It would have been helpful to establish that before a reduction was attached to it.”

The observation stayed with Sophie after the Meridian team had left. Oliver called the exchange a successful test of resolve. They had held their price and secured a more honest comparison with the alternative. But they had also made a modest proposal sound like a verdict on Meridian’s operational judgement.

“She’ll come back,” Oliver said. “They need us.”

“They have an alternative,” Sophie said.

“A costly one.”

“Costly isn’t the same as impossible.”

James intervened before Oliver could reply. “We have the next technical review on Tuesday. Let’s not lose sight of what we got.”

On Monday, Helen sent a short email confirming that Meridian would compare the offers on a like-for-like basis. She also asked Northbridge not to circulate the proposed price reduction as though Meridian had accepted the associated scope change. Sophie replied that it would not be represented as an agreement.

The message was reasonable. Its tone was not warm. In earlier exchanges, Helen had asked questions about process and suggested who should attend. Now she sent precise instructions and copied in a member of Meridian’s legal team. The trust between them had not broken, but it had become narrower.

Sophie took the email to James. “I think we were right to separate the fee from the scope. We were wrong to use the offer as a test of whether she would commit.”

“You didn’t say it was a test,” James said.

“The structure said it for us. The reduction was contingent on a change we hadn’t validated. We made it harder for her to explore the problem without appearing to concede.”

James read Helen’s message. “Oliver will say this is what negotiating feels like.”

“Negotiating can feel uncomfortable without being needlessly defensive.”

Oliver joined them a few minutes later. Sophie did not accuse him of damaging the relationship. She described the sequence: Meridian had agreed to a like-for-like comparison, declined the proposed scope change, and requested that Northbridge keep the unaccepted offer separate from any record of agreed terms.

“We won the comparability point,” Oliver said.

“Yes. And lost room to test the scope.”

“They could have accepted the lower price.”

“They could not responsibly accept the risk attached to it. Rachel identified a genuine gap in our evidence.”

Oliver tapped the folder. “If you keep treating every objection as information, you’ll never get a decision.”

“And if we treat every objection as resistance, we won’t know which problem we’re solving.”

There was no immediate answer. James ended the discussion by asking Sophie to lead Tuesday’s technical review. It was a sensible choice: Martin Lewis, Northbridge’s technical lead, had raised the overnight integration issue before, and Sophie had not yet heard his full explanation.

That afternoon, she reviewed the ticket sample again. The data supported Northbridge’s case that response times had improved. It also exposed gaps in ownership and record-keeping that made a confident claim about removed checks impossible. The commercial point was not that Meridian should trust Northbridge. It was that both sides would need enough evidence to place risk where they could control it.

Sophie added a question to her notes for Martin: What would we have to see to know whether the verification work is actually reducing risk?

It was less forceful than Oliver’s offer. It was also more likely to produce an answer they could put into a contract.

Chapter 9: The Quiet Between Questions

Martin Lewis arrived at the technical review with a laptop under one arm and a printed diagram folded into quarters. He had the careful manner of someone who had learned that a confident answer, given too early, could become a contractual promise.

Rachel and Helen joined by video from Meridian. Priya sat beside Sophie, ready to take notes. James attended, but stayed quiet. Sophie began by acknowledging the previous proposal.

“We linked a price reduction to removing manual verification before we had agreed that the work was redundant,” she said. “That was premature. Today I’d like us to understand the process, then identify what could be tested.”

Helen’s face on the screen remained composed. “That would be useful.”

Sophie turned to Martin. “Can you take us through the overnight stock update as it currently works?”

He shared the diagram. “Meridian’s system generates a file at around half past midnight. Northbridge receives it through an integration layer, checks the format, and passes the data to the stock service. If an expected file is missing, our monitoring generates an alert. If a file arrives but contains values the system can’t interpret, it may be accepted as technically complete and still create a business problem.”

“Who owns the first response to that?” Rachel asked.

“It depends on where the error appears.”

“That sounds like no one owns it,” Helen said.

Martin accepted the point with a slight inclination of his head. “In some cases, that is effectively true. Our agreement assigns responsibility to the party responsible for the fault, but the operational process doesn’t name a single coordinator for identifying the fault. Support may wait for Meridian’s data team; Meridian may be waiting for our integration team.”

Sophie saw James make a note. It was a larger problem than the disputed tickets suggested. The contract’s allocation of liability was precise enough to sound settled, while the working arrangement left both teams unsure when responsibility actually began.

“What happens when the value is wrong but the file passes the format check?” she asked.

Martin paused. “The system may process it. A manual reviewer sometimes catches it, but the review is inconsistent. Some Meridian teams run a second check in the morning. Others do not.”

Rachel glanced away from the screen. “That matches what we see. The checks developed locally. They were never designed as one control.”

The room went quiet. Sophie did not rush to explain the commercial consequences. She waited while Martin examined the diagram. He pointed to a box labelled “exception queue”.

“This is outdated,” he said. “The queue moved eighteen months ago, but the diagram did not. Some alerts still go to the old mailbox.”

“How many?” Priya asked.

“I don’t know without checking the log. I can say that when someone notices the old route, they forward the alert manually.”

Helen spoke after a pause. “That could explain some of the cases we’ve called late response.”

“It could,” Martin said. “It wouldn’t explain all of them.”

Sophie wrote down both parts. The discovery supported Northbridge’s argument that the recorded response times did not always reflect the team’s actual work. It also exposed a Northbridge failure: the integration diagram was out of date, and an alert path had not been properly maintained. Neither side could use the finding as a clean victory.

“What would let us distinguish the two?” Sophie asked.

Martin described a joint review of a representative sample of incidents, including the timestamp when Meridian generated the file, when Northbridge received it, when an alert was raised, and when either team acted. Priya suggested adding a code for the confirmed cause, rather than relying on the first label entered by a support agent.

Rachel asked whether the review would include cases where the file was accepted but the stock value was wrong. Martin said it should. Helen wanted the sample to include both high-impact and routine incidents, so neither side could choose only the cases that favoured its position.

Sophie summarised the emerging agreement. “One hundred cases over six weeks. Both teams nominate a technical owner. We record the four timestamps, confirm the cause where we can, and mark unresolved cases as unresolved. At the end, we decide whether the manual checks are reducing risk, duplicating effort, or doing both.”

“And the outcome doesn’t automatically change the contract price,” Helen said.

“Correct. The review informs a proposal. Any change to price or scope would need a separate agreement.”

Helen looked at Rachel, who nodded. “That is acceptable as a process.”

It was not a concession, and Sophie did not call it one. Meridian was committing staff time and access to its records. Northbridge was committing technical support and a transparent review, including cases that might show its own systems at fault. The exchange created evidence; it did not predetermine who would benefit from it.

Martin folded the outdated diagram. “I should also say that we have a constraint on our side. The current connector is built around an older interface. It still works, but a full redesign cannot be delivered before renewal without pulling engineers away from another migration.”

James looked up. “That hasn’t appeared in the proposal.”

“It’s a delivery constraint, not a reason to refuse the work,” Martin said. “We can replace the alert route immediately. A deeper redesign would take about twelve weeks after access and specifications are agreed.”

Helen’s expression sharpened. “Would Meridian be charged for that?”

“Not if the review shows the defect is within our existing obligation,” Sophie said. “If the redesign is additional work or the current interface is being changed by Meridian, we would need to define that separately.”

“We can’t decide that today,” Helen said.

“No. We should not.” Sophie turned to Martin. “Can you separate the immediate correction from the longer-term redesign?”

“Yes. The old mailbox can be retired and the alert route tested within ten working days. That is ours. The redesign depends on the data specification and on whether Meridian changes its feed.”

Rachel leaned towards her camera. “Our data format is due for review, but no change has been approved. If Northbridge redesigns against a specification we later replace, we’ll both waste money.”

“Then the review should include the specification,” Priya said. “Not just the ticket counts.”

Sophie looked around the room. “Let me check I have this right. Northbridge will correct the alert route within ten working days and provide the test results. Meridian will share the current feed specification and nominate someone to confirm whether a change is planned. The joint sample will then tell us which incidents belong to the route, the data, or the hand-off. We are not agreeing now that a redesign is included or chargeable.”

“That’s right,” Helen said. Her tone had lost some of its edge. “I can work with that distinction.”

The phrase mattered. Helen had not suddenly become trusting, but she had stopped preparing for every sentence as though it concealed a price increase. When Sophie asked what evidence Meridian’s board would accept as a credible route to savings, Helen answered more fully than she had before.

“They’ll want an amount they can see, not only an improvement in process. But I can present a verified cost reduction as part of the renewal if we can show how it was calculated and what service was retained.”

“Would a reduction in manual checking count if Meridian keeps the ability to request it for high-risk periods?”

“Potentially. It would have to be a defined option, with notice and a price. We cannot leave it as an informal promise.”

Sophie repeated the point in her own words: “A lower base fee for a narrower, tested service; optional additional cover at an agreed rate; and no claim of savings until the workload has been measured.”

Helen thought for a moment. “Yes. That is closer to something I could take forward.”

It was the first clear description of a commercial structure Meridian might support. Sophie resisted the urge to move immediately to numbers. The six-week review would not fit neatly into the remaining timetable unless it began soon. The technical work could start at once, but the price proposal would need to distinguish interim arrangements from final terms.

“We can send a draft process tomorrow,” Sophie said. “Not a revised offer. Just the review, owners, data fields and timetable.”

“Please do,” Helen replied. “And keep the existing fee and service in place while we review. I don’t want a gap in cover.”

“Agreed, subject to the current contract.” Sophie glanced at David, who gave a small nod.

When the call ended, James remained by the screen. “That was more productive than Friday.”

“Because we stopped trying to price a conclusion before we had one,” Priya said.

Martin picked up his diagram. “I’ll update it. Properly this time.”

Sophie closed her notebook. The silence she had allowed had not produced a dramatic revelation, only an old mailbox, an outdated interface and a workable test. Yet those details changed the negotiation. They gave both sides a way to examine risk without pretending it belonged entirely to the other.

There was still no agreed renewal price, no commitment to a two-year term, and no guarantee that the review would reveal savings. But Helen had described what she needed to recommend a credible outcome. Sophie had learned that the route to a deal might begin not with a stronger position, but with a question precise enough for the other side to answer honestly.

Chapter 10: The Terms of Staying

The silence after Helen’s last question lasted long enough for the ventilation to become noticeable. Sophie had learnt that a pause could be a form of pressure, but this one did not feel staged. Across the table, David had stopped turning his pen. Helen’s papers were aligned in a careful stack, though one corner had begun to lift.

“If you cannot provide continuous cover for the price in your proposal,” Helen said, “I need to understand what you can provide. Not what you hope to provide after the first quarter.”

Sophie glanced at James. He had the cost model open on his laptop, its neat columns less reassuring than they had looked the previous evening. “We can offer a stronger service than the current one,” she said, “but I don’t want to describe it as twenty-four-seven support if we can’t staff that promise reliably.”

David leaned back. “That is the first sentence today I believe without qualification.”

Helen’s expression did not change, but the remark landed. Sophie felt the temptation to defend the proposal—to explain the hours already spent modelling, the margin already surrendered in the first round. Instead, she set her pen down.

“The distinction matters,” James said. “A staffed, continuous response service has a different cost base from an on-call arrangement. We priced the latter with a daytime operating team and a defined out-of-hours escalation rota. The document did not make that distinction clear enough.”

Helen looked at the schedule. “The headline says 24/7 service.”

“Then the headline is wrong,” Sophie said. “We should correct it.”

The admission seemed to unsettle Helen more than a defence would have done. She turned to David. “What do you actually need overnight?”

He took a moment. “If the ordering platform is unavailable, our stores need someone to acknowledge the incident and begin diagnosis within thirty minutes. Restoration can take longer, depending on the cause. For routine queries, next-day response is acceptable. We don’t need a senior engineer sitting beside a telephone at three in the morning.”

“But the current contract gives us a two-hour acknowledgement,” Helen said.

“And we miss it whenever the problem arrives on a Sunday.” David’s voice remained level. “That is why I want the improvement. I don’t want a promise that looks better in a tender response and fails in practice.”

Sophie wrote three headings on the whiteboard: urgency, coverage, responsibility. “Then we need to separate the service levels rather than bundle them. Critical incidents: acknowledgement within thirty minutes, round the clock, with an agreed escalation path. High-priority issues: one-hour acknowledgement during extended support hours. Everything else: next business day.”

“Extended support hours meaning?” Helen asked.

“Seven in the morning until ten at night, Monday to Saturday, with the critical rota available at all times.”

“That is not the same as full cover,” Helen said.

“No,” Sophie replied. “It is more honest, and more likely to work.”

James added, “We would also need Meridian to nominate an incident lead at all times. Our team can’t safely act on conflicting instructions from a store manager, a regional director and head office.”

David nodded once. “That is fair. We can establish a single duty contact. But I want the escalation rules tested before the holiday trading period.”

Helen drew a line beneath a note. “And the price?”

James did not answer immediately. “It depends on the term, the scope and the implementation sequence. I won’t pretend it doesn’t.”

The current proposal showed an annual fee of £1.86 million, including implementation and the expanded support commitment. James explained that implementation was a one-off £240,000, while the recurring service charge was £1.62 million a year. Revenue, he stressed, was not contribution: after staffing, hosting and subcontractor costs, the margin on the proposed service was already narrow.

Helen looked at the recurring figure. “It is still above our current spend.”

“Yes,” said Sophie. “And the current spend does not buy the service you have just described.”

“You are asking us to fund a better service when we have been told to reduce supplier costs.”

“I understand.” Sophie kept her voice even. “That makes the shape of the contract important. We may be able to change the timing of the investment, or what sits inside the service. We cannot make the same people deliver more hours for less money without increasing the risk of failure.”

David looked at Helen. “A lower number is useful only if the stores can depend on it.”

Helen’s reply came quickly. “And reliability is not an excuse for an open-ended increase.”

Neither was wrong. Sophie had thought of the meeting as a contest over a number; it was becoming a negotiation among people whose definitions of risk did not match.

On the walk back to Northbridge’s office, James shut his laptop only after they had passed reception. “We need to decide what we can walk away from,” he said. “Not as a threat. As a real alternative.”

Their best alternative was not attractive. Northbridge could retain some smaller accounts, redirect a few specialists and pursue new work, but losing Meridian would leave a gap in utilisation and weaken a reference that mattered in the retail sector. Yet accepting a loss-making contract would consume the very people those other accounts needed.

“We can’t promise full overnight cover at this price,” Sophie said.

“We also can’t treat the headline fee as the whole story,” James said. “Payment timing is part of the cost. If they insist on sixty days, we finance the work before we’re paid.”

Sophie looked at the rain moving across the office windows. “Then we offer a service we can stand behind, and make any price movement conditional on something that changes the economics.”

“Term, scope, payment,” James said. “Not a discount for being asked twice.”

She nodded. Their first useful concession would have to come with a reciprocal commitment. Otherwise it would merely disguise a weaker contract.

Chapter 11: The Other Side of the Table

At Meridian’s headquarters, Helen presented the revised service outline to David and Martin, the finance director. Priya, Northbridge’s account director, had sent it that morning with a note correcting the 24/7 wording. That correction was precise and uncomfortable. It made the original proposal look less dependable, but it also made the negotiation more credible.

Martin read the first page twice. “We have a budget ceiling of £1.7 million this year,” he said. “That includes transition. I cannot put another two hundred thousand into the plan because the supplier has discovered its own staffing costs.”

“The transition cost is one-off,” Helen said. “The annual fee is a separate question.”

“Cash is still cash.” Martin tapped the figure. “And this is not the only renewal on my desk.”

David leaned forward. “If we keep the existing service, we keep the existing failure pattern. Stores will continue to call regional managers because the support route is unreliable. That costs us too, although it does not appear on this contract.”

“I know,” Martin said. “I am not disputing the operational problem. I am disputing the assumption that this is the only way to fix it.”

Helen had expected Martin to resist the fee. What she had not expected was his focus on the one-off implementation charge. “Could we phase the change?” she asked. “Keep the essential platform work now and move the reporting enhancements into the second stage?”

David frowned. “The reporting is how we distinguish an isolated store issue from a wider failure. If you defer that, you make incidents harder to diagnose.”

“I said enhancements, not the basic incident dashboard.”

“The distinction has not been clear in the proposal.”

Martin looked from one to the other. “Then clarify it. We have room to move the sequence, not to invent money.”

Helen understood the real boundary. Martin could not easily add to the approved budget, but he could shift expenditure across quarters and consider a longer term if the total commitment was justified. David could accept staged delivery, but not a reduction in the protections that kept stores functioning.

“What if we separate must-have service from optional development?” Helen asked. “A base scope with reliable incident cover, then a priced option for the additional analytics.”

David’s expression remained doubtful. “Only if the base service gives us enough information to act.”

After the meeting, Helen stayed behind to compare the commercial schedule with the notes she had made during the earlier sessions. The quoted implementation figure of £240,000 included migration support, training, data cleansing and a reporting package that had been described as essential in one section and as an enhancement in another.

She circled the reporting package. “How much of this is genuine cost?” she asked Martin when he returned.

“You mean whether Northbridge has padded it?”

“I mean whether the price is attached to work we need now.”

Martin sat beside her. “Their annual number may include a risk allowance. It is common enough. That doesn’t make it dishonest. It does mean we should test which risks we are paying them to absorb.”

Helen remembered Sophie refusing to defend the 24/7 headline. It would have been easier to insist that the wording was a misunderstanding. “If the price was built around a service they cannot deliver, we need a clean reprice, not a token reduction.”

“Agreed,” Martin said. “But don’t confuse a lower proposal with a lower total cost. We still need a viable alternative if this falls apart.”

Meridian had one. A competing supplier had offered a transition plan with a lower first-year fee, but it would require a lengthy migration and would not provide the same familiarity with Meridian’s store systems. Their current team could also be retained temporarily, though at a higher short-term cost and with no guarantee that key staff would stay.

The alternative was credible, but it was not painless. A failure during peak trading would be costly; a transition conducted too quickly could create the same operational weakness Meridian was trying to remove.

“We should not let them use our dependence on their knowledge as a reason to pay anything,” Martin said.

“Nor should we pretend that the knowledge can be replaced overnight,” David replied. “They know which interfaces break when we change a promotional schedule. That is not a line item, but it is real.”

Helen gathered the papers. “Then our position is not ‘cut the fee by ten per cent’. It is: show us a scope that fits the budget and protects the critical service.”

Martin raised an eyebrow. “And if they say the budget cannot buy what David wants?”

“Then we decide whether the alternative is better. But we decide with the actual service and actual cost in front of us.”

As she left, Helen knew she had revised her view of the price. It might be inflated, but not necessarily in the way she had first assumed. The number concealed risk, sequence and work that the two Meridian departments valued differently.

Chapter 12: A Price with Conditions

Sophie returned to Meridian with James and Priya. She had asked Helen for a working session rather than another formal negotiation. The distinction was small, but it changed what they placed on the table: not a final offer, but a structure that could be tested.

Priya opened with the scope. “We propose three service layers. First, critical incidents: continuous access to the on-call rota, thirty-minute acknowledgement and a named escalation lead. Second, extended support for high-priority issues, seven in the morning until ten at night, Monday to Saturday. Third, routine requests handled during business hours.”

David studied the wording. “Acknowledgement is not resolution.”

“Correct,” Priya said. “We will measure acknowledgement separately from restoration. Restoration depends on the cause and on access to Meridian systems. We can commit to response actions and reporting; we cannot responsibly guarantee a fixed recovery time for every failure.”

David nodded, not satisfied but listening. “Then the contract needs to say what happens while we wait. Who updates the stores? Who approves a workaround?”

Sophie answered. “For a critical incident, Northbridge leads diagnosis and provides updates every thirty minutes. Meridian appoints one duty contact to approve operational workarounds. If that person is unavailable, the agreed deputy takes over. We will document the escalation path and test it before go-live.”

Helen wrote down the proposed test. “How do we know whether the service is improving?”

“Monthly reporting,” Priya said. “Acknowledgement performance, restoration times by incident category, repeat incidents, and the proportion of cases reopened. We would agree baselines during the first month because the current records classify incidents inconsistently.”

“That sounds like a convenient way to avoid committing to a target,” Helen said.

“It would be convenient,” Sophie replied, “if we left it there. We can commit now to ninety-five per cent of critical incidents acknowledged within thirty minutes. But we should not set a restoration target until we have comparable data. We can agree a target after the first eight weeks, with the evidence available to both sides.”

David considered this. “If the baseline shows the target is impossible?”

“Then we revise the scope or the target together, rather than pretending a number is operationally meaningful.”

James moved to the commercial schedule. The revised base service would cost £1.58 million a year, with a one-off implementation charge of £210,000. The reporting enhancements would be optional, at £60,000, and could be delivered in the second phase. The initial implementation would include migration, training, essential dashboards and a rehearsal of the critical incident process.

Helen checked the arithmetic. “Your original recurring fee was £1.62 million. You have reduced it by forty thousand and moved sixty thousand of work out.”

“Yes,” James said. “The lower recurring figure follows from removing the advanced analytics from the base scope and reducing the fixed support hours. It is not a discount on identical service.”

“And the £210,000?”

“That reflects the phased reporting work and a smaller training cohort in the first stage. If Meridian wants the full analytics package at launch, the original implementation cost is closer to the real delivery cost.”

Helen looked at Sophie. “You said you could move on price.”

“We can move further if the economics change,” Sophie said. “A three-year term, a named Meridian duty contact and payment at thirty days would allow us to reduce the annual fee by a further £30,000. That is conditional. We would need those commitments in the contract.”

Martin, attending by video, spoke before Helen could respond. “A three-year commitment limits our flexibility. Why should we give you that in exchange for thirty thousand a year?”

James answered without haste. “It lets us retain and train a stable team rather than staff the account on short-term assumptions. Payment at thirty days reduces the working-capital cost of delivery. The concession is linked to those changes; it is not an arbitrary reduction.”

“And if service falls below target?” David asked.

“Service credits for repeated failure against the acknowledgement measure,” Priya said. “With a corrective-action plan first, and a credit capped at a proportion of the monthly fee. The precise cap is negotiable. We do not propose credits for delays caused by Meridian access or approval, provided those dependencies are recorded.”

Helen turned to David. “The scope is narrower than the original promise, but clearer.”

“The night response is better than we have now,” he said. “I still want the analytics early. We have too many recurring incidents to wait indefinitely for the data.”

Martin said, “I can support a staged option, not an unpriced obligation. The first-year ceiling remains.”

Sophie understood the difficulty. The offer was more coherent, but the additional thirty thousand reduction depended on terms Meridian might not accept. Her concession had a price in return, and the other side had its own constraints. No one at the table could simply authorise the agreement they all preferred.

Chapter 13: The Space Between Offers

They spent the next hour testing the proposal against the first year’s budget rather than arguing over a single headline. With the base service and implementation, Meridian’s first-year cost would be £1.79 million. That exceeded Martin’s ceiling by £90,000. Without the analytics option, the second-year recurring cost would be £1.58 million, or £1.55 million if Meridian accepted the three-year term, thirty-day payment and named duty contact.

“I cannot approve £1.79 million this year,” Martin said. “Moving the analytics into a later phase does not solve the implementation figure.”

James opened the cost schedule. “We could split the transition charge across two financial periods, but the work still has to be funded. We would need a signed commitment to the second instalment, not a hope that the budget survives.”

“The second instalment could fall after the new financial year,” Martin said. “That is possible, subject to delivery milestones.”

Helen looked at David. “Would a later analytics phase create an operational gap?”

“If the essential dashboard is ready at launch, no,” he said. “But I want a firm date for the expanded reporting. We cannot leave it as an option we may never exercise.”

Sophie marked a line on the draft. “We can make the second phase a scheduled deliverable, not an optional purchase. The scope would be fixed now, with the work starting after eight weeks of live-service data. The fee would be £60,000, payable on acceptance. We would agree the acceptance criteria today.”

Helen said, “That helps with timing, not total cost.”

“It helps with the decision,” Martin replied. “We can reserve for the second instalment if the first phase is successful.”

David was cautious. “I don’t want the analytics delayed because a dashboard takes longer than expected. Nor do I want Northbridge to claim that poor source data excuses a missed milestone.”

Priya proposed a responsibility schedule. Northbridge would configure the dashboard and define the incident categories. Meridian would provide named data owners, access to the relevant systems and decisions on disputed classifications within five working days. If a dependency was late, the delivery date would move by the length of the delay, but Northbridge would have to notify Meridian promptly and show the impact.

“That is reasonable,” Helen said. “Provided the notification is contemporaneous, not added to a report after the deadline.”

The conversation moved to payment. Meridian’s standard terms were sixty days. James explained that the lower annual price assumed payment within thirty. If Meridian required sixty days, Northbridge could retain the £1.58 million fee but not the additional £30,000 reduction. Helen objected that the supplier was asking Meridian to finance its staffing model.

“Partly,” James said. “And we are asking not to finance the whole service for two months before receiving payment. It affects our cost of delivery. You can choose the longer term and lower rate, or retain the standard payment period and the standard rate.”

Sophie added, “We are not asking for a concession without a return. Likewise, if you ask us to lower the price, we need a change in term, scope, or cash timing.”

Martin’s face on the screen remained unreadable. “Our finance team will resist thirty days.”

“Then the £30,000 reduction is unavailable,” James said. “That is a commercial choice, not a penalty.”

Helen heard no threat in it. The clarity made the trade-off harder to dismiss.

After a short recess, she spoke privately with David in the corridor. “Your position has changed,” she said.

“My position is the same. Reliable service.”

“You were asking for full cover at the beginning.”

“I was asking for confidence that someone would act when a store went dark. I don’t need a fully staffed night shift if the rota works and we can see the evidence.”

“Martin is concerned the price was inflated.”

David gave a small shrug. “Some of it was scope we may not need on day one. That is not the same as padding. But we should not pay for a service description that doesn’t match the staffing.”

Helen understood that she had treated Meridian as one buyer, as if its departments shared a single measure of value. David valued operational resilience; Martin valued budget certainty. Neither was a disguise for the other. “We need an agreement that lets both of you defend it,” she said.

“And lets the stores use it,” David replied.

Back in the room, Helen proposed a three-year term with an annual review of service measures, thirty-day payment subject to finance approval, and the £30,000 reduction only if both conditions were confirmed. The implementation charge would be split: £150,000 at signature and £60,000 after the new financial year, following acceptance of the initial migration and dashboard. The analytics phase would be scheduled for week nine, priced at £60,000, with the data responsibilities and delay rules written into the contract.

Martin objected to tying the second instalment to acceptance without a defined test. David insisted that the essential dashboard be available before go-live. James wanted the payment date for the later instalment guaranteed. Priya began drafting acceptance criteria, while Sophie watched the shape of the possible agreement hold and shift beneath their questions.

“We have a structure,” Helen said at last. “We do not yet have an agreed contract.”

“No,” Sophie replied. “But we now know which decisions are commercial, which are operational and which need evidence.”

Martin requested a revised cash-flow schedule. David asked for the incident rehearsal to be brought forward. James agreed to recalculate the cost of the staged transition, but not to absorb the later analytics work without payment. Helen collected the marked-up pages.

The meeting ended without a handshake. Sophie did not mistake that for failure. Both sides still had alternatives, each imperfect; both had named what they could trade and what they could not. The next draft would show whether those distinctions could survive the language of a contract.

Chapter 14: The Small Print of Night

The meeting room at Northbridge Solutions had been booked until six, though nobody had said what they expected to have achieved by then. Sophie arrived with two versions of the renewal schedule: the clean copy, which showed the latest commercial offer, and the marked-up copy, where the wording had begun to resemble a map of disputed territory. She set both beside the water jug.

Across the table, Rachel from Meridian Retail Group was reading the service appendix. Martin, her operations director, had brought no papers at all. He had come straight from a call and still wore his headset around his neck. Helen sat beside Sophie, calm and exact, with Priya opposite her. James had joined by video. Oliver was absent; he was meant to be preparing the revised pricing table.

‘Before we go back to the annual figure,’ Rachel said, ‘I want to settle the incident language. Your draft says that an out-of-hours incident begins when Meridian reports a material failure. Our current agreement says when the service is materially affected. Those are not the same test.’

Sophie glanced at the clause. She had read it twice, but not with that distinction in mind. ‘Can you give me an example?’

Martin folded his arms. ‘A store cannot complete a stock transfer at ten past ten on a Saturday evening. The system still responds, technically. Your team says that is not an incident until the fault is confirmed. Our people are already calling shops and moving stock by hand. From our side, the service is affected.’

Priya leaned forward. ‘If we treat every degraded response as a major incident, we need to distinguish a genuine platform fault from a local network problem or a third-party device. We cannot promise a twenty-minute restoration when we have not yet identified what is broken.’

‘Nobody is asking you to promise restoration,’ Martin said. ‘We need you to start the clock when the business impact is real.’

The difference sounded small in a sentence and large in the risk schedule. Sophie looked at Helen. ‘What would the support model need to do?’

‘First, define the trigger,’ Helen said. ‘Then separate acknowledgement, triage and restoration. At present those are being bundled together in the draft. If the clock starts on a report, regardless of impact or access to evidence, the exposure is open-ended. If it starts only after our engineer confirms a fault, Meridian carries the waiting time.’

Rachel closed the appendix. ‘And this is why the headline price is not the whole agreement.’

They had spent weeks discussing the renewal as if it were a contest over a single number. The price mattered: Meridian wanted a reduction against the proposed annual fee, while Northbridge had been resisting any cut that would make the service unprofitable. But the clause now on the table could turn a modest fee into a much larger liability, or make the renewal look cheap while leaving Meridian to absorb the very disruption it was buying protection against.

James spoke from the screen. ‘We have priced the existing support hours and the current severity definitions. We have not priced an unlimited out-of-hours commitment.’

‘Nor should you,’ Martin said. ‘We’re not asking for unlimited cover. We’re asking you not to define an incident out of existence.’

Sophie heard the heat in his voice and did not answer it. ‘Let’s separate the interests. Meridian needs a dependable response when stores are affected. Northbridge needs an observable trigger and a commitment it can staff. Is that fair?’

Martin gave a brief nod. Rachel said, ‘Yes, provided the definition can’t be narrowed later by an internal interpretation.’

Priya opened her notebook. She proposed three levels: a confirmed platform outage, a material degradation affecting a defined group of stores, and a single-site issue. Each could trigger a different response target. Meridian would provide a named incident contact and basic diagnostic information; Northbridge would acknowledge within twenty minutes for the first two levels, begin triage immediately, and give a restoration estimate once the fault was understood. A single-site incident would receive a different target, unless the evidence showed a wider pattern.

‘That is closer,’ Martin said. ‘But what does “material degradation” mean?’

‘A sustained inability to complete a listed critical transaction,’ Priya replied. ‘We can define the transactions in an operating schedule. If a stock transfer is on that list, a workaround does not make the service healthy.’

Rachel looked at Sophie. ‘Would you accept that?’

Sophie considered the service data. Northbridge could support the proposed model with a rota and a limited increase in weekend cover, but not at the present renewal price if every degraded transaction carried the same remedy as a total outage. ‘I would accept the operational definition, subject to a joint schedule of critical transactions and a review of the service credits. The schedule must be fixed before signature, not left to a later workshop.’

‘A review can become a way of postponing,’ Rachel said.

‘Then let’s make completion of the schedule a condition of the renewal, with a deadline and named owners.’ Sophie turned to Martin. ‘You know which transactions stop a store functioning. Priya knows which signals we can measure. Neither side can define this alone.’

Martin’s expression shifted, not into agreement but into attention. ‘That would be useful. My team has been calling too many things critical because the current categories are vague. We need the definition to be credible in a store at eleven at night, not only in a contract.’

The meeting moved to a practical test: who could raise the incident, what evidence was required, and how the parties would handle a report made before a Northbridge engineer was available. They agreed that a report from Meridian’s on-call manager would start acknowledgement immediately. Confirmation of a fault would not be a precondition to triage. The precise service-credit formula remained open.

Sophie wrote the distinction in the margin: response clock; diagnosis; restoration. She had spent much of the negotiation hearing positions—lower price, broader cover, fixed credits—and treating them as if each were a complete account of what the other party wanted. Martin did not want a penalty for its own sake. He wanted a response that did not wait for a supplier to agree that the problem existed.

At the end, Rachel asked for the latest pricing table. Sophie glanced at the empty chair where Oliver should have been. ‘He’s preparing it. I’ll send it as soon as it’s checked against the service assumptions.’

‘The assumptions are changing now,’ James said through the speaker. ‘I can’t sign off a revised margin until we know the cost of the rota.’

‘Nor am I asking you to,’ Rachel said. ‘But we need progress we can take to our people.’

Sophie promised a date rather than a number. By the time the Meridian team left, the technical wording was less obscure and the commercial position less comfortable. The out-of-hours definition had not been discussed in any of the earlier rounds because both sides had assumed it was settled. It was not. Sophie understood that the contract’s risk had been hiding inside a phrase small enough to pass unnoticed.

In the corridor, Helen said, ‘We can staff a defined response. We cannot staff a blank cheque.’

‘I know,’ Sophie said. ‘The question is whether we can show them the difference without making it sound like an excuse.’

Chapter 15: The Message That Travelled

At eight the next morning, Sophie found an email from Oliver marked URGENT. It had been sent to a distribution list that included Northbridge’s finance team and, through an autofill error, Rachel at Meridian. The subject line read: ‘Concession framework—hold the line on credits’. The first sentence said, ‘We can absorb the support uplift if it buys us a clean price close; Meridian will take the credit cap.’

Sophie read it twice. Then she read Rachel’s reply, sent at 7.42. ‘I am surprised to see our position described as something to be “taken”. We have not agreed to a credit cap. Please clarify whether Northbridge has already decided the service terms are negotiable only in one direction.’

The words were polite. The wound beneath them was not.

Sophie called Oliver. He answered after three rings. ‘I’ve seen it,’ she said.

‘I meant it as an internal working note. I was trying to get James to price a package rather than keep reopening every line.’

‘You sent it to Rachel.’

There was a pause. ‘I know. I’ve apologised.’

‘You also wrote that Meridian would take the cap.’

‘That was shorthand. I meant we should offer the cap in return for the support commitment.’

‘That isn’t what the sentence says.’

Oliver exhaled. ‘I understand. I can explain it at the next meeting.’

Sophie looked at the marked-up appendix. The email had made the negotiation sound like a private exercise in assigning concessions, with Meridian cast as a party to be managed rather than a party whose approval mattered. An explanation might correct the wording. It would not erase the impression that Northbridge had been making decisions about Meridian’s commitments without asking.

‘Don’t send another message yet,’ she said. ‘I’ll speak to Rachel.’

She called Rachel, who answered from a station platform. Behind her voice came the indistinct announcements of a train service. ‘I’m not going to pretend Oliver’s wording was harmless,’ Sophie said. ‘He was preparing an internal package, and he sent it to you by mistake. That explains how it happened, not why it reads as if your agreement is assumed.’

Rachel was quiet for a moment. ‘I appreciate the distinction. But the distinction is also the problem. Your internal note treats the credit cap as something you can give or withhold, and our service risk as a price lever.’

‘You’re right that it reads that way. The intention was to trade a cap for a defined support commitment. We have not agreed either term.’

‘Then why did you not say that?’

Sophie could have said that Oliver had made the mistake, that she had not seen the email, that the distribution list was at fault. Each was true and none was enough. ‘Because I should have set clearer rules for what went into the working drafts and who received them. I’ll own that. I can’t make an offer on the phone, but I can make sure the next one shows what Northbridge is asking for and what Meridian would receive in return.’

Rachel’s voice softened only slightly. ‘I need something I can take to Martin. The service issue is not a bargaining ornament.’

‘I understand. I’ll keep it tied to the operating model, not describe it as a favour.’

They ended without a reconciliation. Sophie sat with the phone in her hand until James’s call arrived. He wanted to know whether the pricing table should assume the credit cap. ‘Not yet,’ she said. ‘We’re going to cost scenarios, not smuggle an assumption into the base case.’

Oliver came into her office ten minutes later. He looked tired and defensive. ‘I didn’t mean to undermine you.’

‘I believe that,’ Sophie said. ‘But intent doesn’t fix the effect. Why did you write “Meridian will take the credit cap”?’

‘I was trying to get us out of the circular debate. Every time we move on price, the support language changes.’

‘Because support changes our cost and their risk. Those aren’t separate.’

‘I know.’ He sat down. ‘I thought a clean number would give them a reason to move.’

Sophie kept her voice level. ‘A clean number that hides a condition is not clean. It is a delayed argument.’

Oliver looked at the table rather than at her. ‘I should have sent it to you first.’

‘Yes. And I should have made it explicit that no one sends a client-facing package until we have checked the reciprocal commitments. I let the drafts travel too freely because we were all trying to move quickly.’

It mattered to say it. Not because his error was hers, but because a negotiation team could not ask the other side to trust its process while pretending that process belonged to no one. Sophie had been tempted, in the first moment, to let Oliver carry the whole blame. It would have been efficient and unfair.

She asked him to prepare three costed scenarios by noon: current support with the proposed price, enhanced out-of-hours response with a credit cap, and enhanced response without a cap but with narrower, measurable remedies. Each had to state its assumptions and the commitment required from Meridian. Oliver agreed without protest.

At eleven, Sophie and Rachel spoke again with Martin present. Sophie began by acknowledging the email, then gave the facts: it was an internal draft sent in error; no credit cap had been agreed; the figures were not approved. She did not ask them to disregard it.

‘What are you changing?’ Rachel asked.

‘The process. You’ll receive a single comparison showing the price, the service obligation, the credit exposure and the evidence required for each option. No option will be described as a concession unless it is matched by a commitment or a value Meridian recognises.’

Martin said, ‘Good. But a comparison won’t solve the incident definition.’

‘No,’ Sophie replied. ‘It will show what each definition costs and what protection it provides. Then we can decide whether the cost is justified.’

After the call, James told her the additional rota could be staffed for an estimated £68,000 a year, subject to recruiting two people and adjusting the on-call arrangement. The enhanced response was feasible; unlimited restoration guarantees were not. The distinction was becoming visible in the numbers, though not yet acceptable to either side.

Sophie opened the renewal schedule and changed the heading from ‘Concessions’ to ‘Options and reciprocal commitments’. It was a small edit. It did not undo the email. It did, however, make the next conversation harder to misrepresent.

Chapter 16: A Saturday at Half Past Ten

The incident arrived on Saturday evening, at 10.32, while Sophie was helping her daughter revise for an examination. Her phone rang once, stopped, then rang again. The message from Northbridge’s service desk said that Meridian had reported failures in stock transfers at three stores. The system was responding, but transactions were timing out. The on-call engineer had acknowledged the report at 10.41.

Sophie moved to the kitchen and called Helen. ‘Do we know if it’s our platform?’

‘Not yet. Priya is checking the transaction logs. It may be a regional network issue. The stores can still sell; they can’t move stock between locations.’

‘How many?’

‘Three confirmed, one more being checked.’

Sophie opened the draft clause on her laptop. Under the current agreement, the incident might be classified as a limited service degradation, or as a local connectivity problem. Under the proposed wording, it might qualify as an out-of-hours material incident. The negotiation had become a live test of the words they had not yet settled.

She called Martin. He answered quickly. There was a loud, hollow sound behind him, as if he were in a car park. ‘The stores are using paper records,’ he said. ‘We’ve got a late delivery at one of them and no reliable view of what can be transferred.’

‘We have the report. The engineer has acknowledged it and is working through the logs. I don’t yet know the cause.’

‘I’m not asking you to guess.’ Martin’s voice was clipped. ‘I’m asking what happens next.’

Sophie gave him the service desk’s next update time, not a promise about restoration. ‘Priya is checking whether the failures share a pattern. We’ll call you at eleven, even if the answer is that we’re still diagnosing.’

‘And if it’s not your platform?’

‘We’ll tell you what we can establish, and what we need from your network team.’

They agreed that Martin would provide the store identifiers and local contact details. Sophie did not invoke the draft clause as if it were already binding; nor did she hide behind the fact that it was not. She said Northbridge would handle the report under the existing agreement while using the proposed escalation process as a practical guide.

At 11.06, Priya joined the call. Two stores were using the same third-party network gateway; the third had a separate route but showed a similar delay in the transaction queue. Northbridge had found a backlog in a connector service, but not yet the cause of the backlog. The available evidence did not support a simple claim that Meridian’s network was responsible or that Northbridge’s platform had failed.

‘Can you clear the queue?’ Martin asked.

‘We can restart the connector for the affected stores,’ Priya said. ‘That should restore transaction flow. It will not tell us why the backlog formed, so we’ll keep the logs and investigate overnight.’

‘Do it.’

The restart began at 11.14. Sophie stayed on the call, saying little. She could feel the pressure to turn each update into reassurance. Instead, she asked what the team knew, what remained uncertain, and when the next observation would be available. The questions slowed the conversation in a useful way. No one had to defend a guess as if it were a fact.

By 11.38, the first store had completed a stock transfer. The second followed six minutes later. The third had cleared most of its queue by midnight. Martin confirmed that the stores could return to normal procedures, though he wanted a written incident record the next morning.

‘You’ll have it by ten,’ Sophie said. ‘It will separate the timeline, the confirmed cause, the likely contributing factors and what we don’t yet know.’

‘That’s what I need,’ Martin replied. He sounded less angry, but more tired. ‘The operational damage isn’t only a system being down. It’s the time people spend deciding whether they can trust it.’

At 12.17, Sophie ended the call. James had sent a message asking whether the event should alter the service-credit model. She did not answer immediately. There was not enough evidence yet to classify the event, and the parties had not agreed the proposed definition. To award or deny a credit on instinct would only repeat the same mistake in a more public form.

At home, her daughter had gone to bed. Sophie sat at the kitchen table with the incident timeline and checked each time against the service desk record. The report had been acknowledged nine minutes after the first alert. Triage began four minutes later. The system was operating again within an hour, though full confidence in the cause would take longer.

The facts were not dramatic. Their value lay in what they could prevent: exaggeration by Meridian, minimisation by Northbridge, and the convenient rewriting of events once the commercial consequences were known.

On Sunday morning, Priya’s preliminary report identified a queue-management defect following a routine configuration change. The defect had affected a narrow set of transactions; the network gateway had made the symptoms more visible at two sites but was not the root cause. The configuration had been rolled back, and a code fix was in testing. There was no evidence of data loss.

Sophie read the report with Helen. ‘We should tell Martin before the full root-cause review,’ Helen said. ‘He needs the operational facts, not a polished conclusion.’

They sent the timeline and preliminary finding, clearly marked as subject to further verification. Martin replied that Meridian would preserve its store records and provide the transaction logs from its side. He did not request a credit. He did ask that the event be included in the renewal discussion.

On Monday, the teams met by video. Martin described the effect: four stores had used manual stock controls, and one late delivery had needed a supervisor’s attention. No sales had been lost, but the disruption had consumed staff time and reduced confidence in the inventory view.

James asked, carefully, whether Meridian had calculated a financial loss. Martin said it had not. ‘I’m not going to invent a number because the contract has a credit table. The cost was real, but we haven’t measured it well enough to claim a figure.’

Sophie appreciated the restraint, and said so. ‘We’ll include the incident in the service record. We’ll also distinguish the direct operational effect from any contractual remedy. Those are related, but they aren’t interchangeable.’

Rachel joined the conversation. ‘This is precisely why the definition matters. Your team responded well. The contract should not depend on whether the relationship is friendly on the night.’

‘Agreed,’ Sophie said. ‘The event shows we need a clear start to the response clock, a measurable definition of material impact, and a separate method for assessing remedies. It also shows that a fast response can limit harm without erasing it.’

Nobody turned the incident into a victory. Northbridge had found and addressed a defect; Meridian had supplied information and managed its stores. The service had recovered, but trust would depend on the accuracy of the account and whether the contract made the next response dependable.

After the call, Sophie amended the options paper. She added the Saturday event as a worked example, not as a threat or a precedent. The figures did not decide the negotiation. They made it harder to pretend that the clause was merely legal language.

Chapter 17: The Value of the Boundary

By Tuesday morning, Northbridge’s executive team wanted a recommendation. The renewal was now close enough to affect the quarterly forecast, and the existing contract’s notice period was approaching. Sophie’s managing director had not intervened in the negotiation; he had simply asked for a decision he could defend to the board. James wanted the margin floor stated plainly. Helen wanted the support rota funded before any service promise was made. Oliver wanted to know whether the client would move on price.

‘They are not one question,’ Sophie said.

James tapped the scenario sheet. ‘The current fee is £1.24 million a year. Our proposed renewal is £1.31 million, reflecting the additional reporting and the price adjustment in the original term. Meridian’s last position was £1.17 million. At that level, with expanded support, we fall below the margin floor.’

‘Revenue is not margin,’ Sophie said. ‘And their price position is not necessarily their spending limit. What do we know about the cost of each service option?’

James turned the sheet towards her. The existing support model, at the current service scope, produced a contribution margin of 24 per cent at the proposed fee. Adding the weekend rota reduced it to 19 per cent before any service credits. A capped credit model could keep the margin above 17 per cent if the cap and triggers were measurable. An uncapped model with broad incident language could not be responsibly priced.

Helen added that recruitment would take at least six weeks. ‘If we sign the service schedule now and start the rota later, we need a clear transition date. We can’t be contractually in breach on day one because we’ve promised a team we haven’t hired.’

Oliver, sitting at the far end, said, ‘Meridian may prefer the existing hours if we give them a larger price reduction.’

‘Possibly,’ Sophie replied. ‘But we need to find out what they value, not guess from their opening position.’

She looked again at the options. One path protected margin but left Meridian paying for reassurance it might not receive. Another gave broader cover at a fee Northbridge could not sustain. The third—defined response, limited enhancement, agreed incident categories, evidence-based credits—was commercially possible, but only if Meridian accepted the operational boundaries and Northbridge delivered them.

‘I’m recommending the third,’ she said. ‘Not as a halfway point. As a different package. We should show them what it buys and what it excludes.’

That afternoon, Rachel and Martin came to Northbridge’s office. Rachel had brought a revised mandate from Meridian’s finance team. The renewal had to fit within a fixed annual budget unless the additional service could be funded from a separate resilience allocation. Martin’s interest was different: he needed a dependable escalation path for store-critical failures, and he was willing to accept a narrower scope if it was clear at the point of use.

Sophie laid out the three scenarios. She started with the interests rather than the numbers. ‘You have a budget constraint, Rachel, and an operational need that cannot be met by a vague promise. We have a staffing constraint and a margin threshold, but we can support a defined increase in response. I want to test whether the resilience allocation can pay for the part of the service that creates the additional cost.’

Rachel studied the first page. ‘Our budget is not a puzzle for you to solve by moving charges between headings.’

‘No,’ Sophie said. ‘It is a constraint for us to respect. The proposal is transparent: base service at £1.24 million, plus £68,000 for the enhanced out-of-hours rota, if that funding is available. If it is not, we can discuss a narrower support window at a lower price. We won’t label the second option equivalent.’

James explained that the rota cost covered staffing and supervision, not a guaranteed restoration time. The response commitment would be an acknowledgement within twenty minutes for a defined critical incident, immediate triage and updates at agreed intervals. Restoration estimates would follow diagnosis. A service credit would apply if the response target was missed, with a capped annual amount and a separate, joint review for repeated failures.

Martin asked, ‘Would a store unable to complete a listed stock transfer qualify?’

‘Yes, if it is sustained and reported through the agreed channel,’ Priya said. ‘A temporary workaround would not remove the incident. But the schedule must say which transactions are critical, and the evidence must let us distinguish a platform failure from an isolated local device fault.’

‘That is workable,’ Martin said. Rachel did not agree. She was looking at the total fee. ‘The enhanced option is still above the figure we put forward.’

Sophie let the silence remain. She had learnt that filling every pause with explanation could make a position sound less considered, not more. Rachel turned the page herself.

‘What is the lower-service option?’ she asked.

Sophie showed it: £1.17 million for the current hours and existing scope, with a modest improvement to reporting but no dedicated weekend rota. The response rules would still be clarified; the larger credit exposure would not be included. The choice was not between a good service and a bad one. It was between two levels of availability with different costs and risks.

‘If we select the lower option,’ Martin said, ‘we still need the incident definition fixed.’

‘Yes,’ Sophie said. ‘Clarification of what an incident is should not be sold as an enhancement. The additional cost is for staffed coverage, not for being clear about the service we already owe.’

Rachel looked at her. ‘That is a useful distinction. It also means Northbridge cannot claim the rota is a concession.’

‘Agreed. It is a paid service option. The reciprocal commitment from Meridian would be the named on-call contact, access to the relevant logs and a joint schedule of critical transactions by the agreed date. The credit cap is part of the risk allocation, not something you receive in exchange for answering the phone.’

Martin gave a small nod. ‘And if we agree the schedule, you will not later say that the transaction is outside scope because the system technically responded.’

‘Not if the agreed test says the transaction could not be completed. We’ll use the Saturday incident as a test case for the wording, while keeping its actual contractual treatment under the current agreement.’

Rachel asked for the proposed cap. James gave the figure, then explained how it had been calculated against the annual service fee and the maximum likely credit exposure. Rachel did not accept it, but she did not reject it as arbitrary. She asked for a second calculation at a higher cap with a corresponding change to the fee. James agreed to provide one.

The meeting ended without a signed renewal. It did end with three clear tasks: Meridian would confirm whether the resilience allocation could fund the rota; Northbridge would cost the higher-cap scenario; and Martin and Priya would draft the critical-transaction schedule by Friday. Rachel would take both service options to Meridian’s finance committee, rather than a single headline price stripped of its conditions.

As they stood, Martin said to Sophie, ‘You’ve stopped describing our request as more service for less money.’

‘Because that wasn’t the whole request,’ she said.

He gave a brief, tired smile. ‘No. It wasn’t.’

After they had gone, James asked whether Sophie thought the deal was close. She considered the question before answering. ‘Close enough to make a careless promise expensive. Not close enough to call agreed.’

Oliver gathered the papers. ‘The lower option is within their number. We could probably close it if we pushed.’

‘We could offer it clearly,’ Sophie said. ‘We can’t push them into pretending it meets an operational need they have already described.’

Oliver paused. ‘And the enhanced option?’

‘It is viable if they have the budget and if we recruit to the plan. It is not viable if we discount the fee and keep the same exposure. That is the boundary.’

Helen closed the staffing sheet. ‘I can support either scope. I can’t support both promises at once.’

Sophie gathered the two options, no longer embarrassed by their difference. A negotiation did not become fair because both parties moved by the same amount. It became workable when each side knew what it was buying, what it was accepting, and what it would have to do in return.

The next meeting was set for Friday. There would be no grand conclusion, only a revised credit calculation, a completed incident schedule and a decision from Meridian’s finance committee. Sophie knew that a viable agreement might still fail if either side mistook a boundary for a threat, or a movement for an obligation. She also knew they had stopped arguing about one number as if it could carry the entire relationship.

That evening, she sent Rachel a short note confirming the three tasks and the assumptions behind each price. She did not call the proposals concessions. She called them choices. Then she checked the recipients twice before pressing send.

Chapter 18: The Trade

On Monday morning, Sophie found the revised proposal waiting in her inbox beside a message from Helen: “Before we reopen the price, be clear what we are buying with any movement.” The sentence was brisk, but not unkind. Helen had spent the weekend answering questions from Northbridge’s board about a renewal whose revenue looked reassuring until somebody asked what it would cost to deliver.

Sophie printed the proposal and carried it to the small meeting room. James was already there, tapping a pen against the margin of a spreadsheet. Priya joined by video from the implementation floor, where two teams were testing a release that had nothing to do with Meridian and nevertheless made every spare engineer feel precious.

“Oliver’s latest position is still two-point-one million a year,” James said. “And he wants the full service from the first day.”

“That is what he is asking for,” Sophie replied. “It is not necessarily what he needs.”

James looked up. “We have been making that distinction for weeks.”

“Then we should finally use it.” She turned the proposal towards him. “His interest is a predictable operating cost and a credible route to national coverage. The position is a lower annual number and immediate delivery. Those are not the same thing.”

Priya’s image shifted as she leaned closer to her camera. “The route matters. If we take the regions in sequence, I can keep the existing teams intact and train Meridian’s people alongside each stage. If they insist on launching everywhere together, we need a temporary second shift and contractors. That is not a saving, whatever the headline fee says.”

James drew a line through a figure. “A phased launch means lower revenue in year one.”

“And lower delivery cost,” Sophie said. “Revenue is not margin. We have been treating every pound we do not invoice as a pound lost, while ignoring the cost of earning it.”

There was a short silence. James capped his pen. “All right. What are you proposing?”

Sophie wrote three headings on the whiteboard: TERM, SEQUENCE, PAYMENT. “A three-year commitment. National coverage delivered in phases, with agreed dates and acceptance criteria. A payment schedule that gives Meridian a lower charge while fewer regions are live, then the full run-rate once the final stage is accepted.”

“Longer term in exchange for a softer first year,” James said. “And earlier cash in exchange for…”

“Not earlier than we can reasonably ask,” Sophie said. “Forty-five days from a valid invoice, rather than another argument about what the invoice means.”

David, who had been listening from the doorway, came in with a cup of tea and an expression that suggested he had heard the word cash. “Forty-five days helps, but it doesn’t fund a careless scope.”

“It isn’t meant to,” Sophie said. “The scope is explicit. The sequence is explicit. The fee changes as the service comes on line.”

Helen arrived a few minutes later, carrying no papers. She listened while Sophie walked through the structure. At the three-year full-service rate, Northbridge would earn less in year one than under the original proposal. But the phased implementation would reduce subcontracting and overtime; the longer commitment would make it reasonable to reserve a dedicated team. On the model David had prepared, the margin was modest, not brilliant, and better than the version in which Northbridge accepted Meridian’s price while retaining all its delivery obligations.

“What do we give up?” Helen asked.

“Immediate national coverage,” Priya said. “And the illusion that everything can be done at once without consequence.”

“What do we get?”

“A committed term, a sequence we can deliver, and a customer team trained before each handover,” Sophie said. “We still carry implementation risk. They carry the risk that a later region does not go live until its agreed readiness conditions are met.”

Helen considered the board for a moment. “Do not describe a longer term as a concession if it gives them a better price.”

“I won’t. It is an exchange.”

That afternoon, Oliver joined by video. Rachel, Meridian’s legal director, sat beside him with a marked-up draft. Martin, the executive sponsor, was absent; he had told Oliver he wanted the commercial team to settle the mechanics before asking him to judge the result. Sophie thought that sensible, though it left nobody in the room with the authority to declare a victory.

Oliver listened to the phased proposal and made notes. “You’re asking us to accept less coverage in the first year and pay more once the service is complete.”

“We’re asking you to pay for what is live,” Sophie said. “And to commit to the route to full coverage rather than hold us to a date that depends on readiness at your end as well as ours.”

“Our stores do not get to phase their customers.”

“No,” Priya said. “But we can phase the work behind them without changing what a store experiences once its region is live.”

Rachel looked up. “What happens if one of your stages is late?”

“We agree a recovery plan, an accountable owner, and a review date,” Sophie said. “Any service credit follows the measured impact, not a general dissatisfaction clause.”

Oliver’s mouth tightened. “That sounds like less protection.”

“It is more precise protection,” Sophie replied. “The current language makes both sides argue about what counts as failure. We can improve the remedy, but it has to attach to something we can measure and control.”

Oliver glanced at Rachel. Neither of them answered at once. Sophie let the silence remain. The first useful movement, she knew, would not sound like agreement. It would sound like a question about how the exchange worked.

“If we take the three-year term,” Oliver said, “how much of the price is fixed?”

“The contracted service fee is fixed for the first twelve months. After that, any annual adjustment is capped at the lower of the agreed index or three per cent. No adjustment applies to a stage before its acceptance date.”

“You have not offered that before.”

“Because we had not yet agreed what you would commit to in return.”

Rachel turned a page in her draft. “And the payment schedule?”

“Forty-five days from receipt of a valid invoice. We will agree the invoice format and the evidence required for each stage. No disputed item holds up an undisputed one.”

Oliver leaned back. “That is a considerable change from the sixty days in our standard terms.”

“It is,” Sophie said. “The three-year commitment and the phased fee make the cash flow workable for us. The schedule gives you an invoice you can verify against delivery.”

James, beside her, kept still. Sophie had expected him to emphasise the value of the commitment. He did not. That restraint mattered: they were offering a trade, not presenting an advantage as though Meridian should thank them for it.

Oliver asked for a break. When he returned, his first question was not about the price. He wanted to know whether Meridian’s regional leads could attend Northbridge’s training sessions before their stages went live. Priya explained that they could, provided Meridian nominated people with authority to confirm local readiness. Oliver wrote that down.

They did not settle the fee that day. They did, however, agree to model a three-year term, a phased roll-out, and a forty-five-day payment cycle together rather than as separate demands. Sophie left the call with a list of figures to test and the distinct impression that the negotiation had finally acquired a shape.

In the lift, James said, “You make it sound as though we have given them something.”

“We have.”

“And we’ve received something.”

“That is why it may hold.”

He looked at her for a second, then nodded. “I’ll redo the margin model.”

“With the actual staffing assumptions,” she said. “Not the version where every person is available twice.”

“I know how to build a model.”

“I know. I’m asking you to build this one.”

His expression softened, though not quite into a smile. The distance between them had been made of small things: late messages, unspoken blame, and the assumption that the other person would understand. It would not disappear because they had had one competent conversation. But Sophie could at least stop adding to it.

That evening, she read Priya’s implementation notes. They set out three regional waves, each dependent on data checks, named Meridian contacts, and a sign-off meeting. The timing was tight but possible. Priya had added a warning in the margin: “A date without a customer owner is a hope, not a plan.”

Sophie copied it into the draft as a principle, not a slogan. The proposed agreement was beginning to contain the work as well as the promise. That made it less elegant and more useful.

She sent the revised commercial structure to Oliver and Rachel with a note: “This is not a discount for a longer term. It is a lower first-year charge for a phased service, in exchange for a defined commitment and a payment schedule both teams can administer. Please challenge the assumptions, not just the total.”

Oliver replied shortly before nine. “We will. I can see the logic. I cannot yet say the number works.”

Sophie read the message twice. It was not approval. It was better than another position.

Chapter 19: The Measure of Service

The next meeting took place in Meridian’s offices, in a room with a long window facing a wet terrace. Oliver arrived with Rachel and Martin. The executive sponsor had brought a narrow notebook and no prepared speech. He listened as Sophie opened with the operating model, then asked for the proposed service measures.

Priya had travelled down to present them. She placed a single-page schedule on the table. “We have separated availability, response, resolution, and implementation. At present, the draft treats them as if they were interchangeable. They aren’t.”

Martin examined the page. “Our concern is the time a store has to wait before it can trade normally. The dashboard can look healthy while a store is still making workarounds.”

“That is fair,” Priya said. “A response within an hour is not a resolution. We can report both, and we can define which incidents have a direct operational impact.”

Oliver tapped the schedule. “We need a stronger service promise if we are taking a longer commitment.”

“You should,” Sophie said. “But a stronger promise has to be deliverable and measurable.”

The draft proposed monthly availability of 99.7 per cent for the live service, excluding agreed maintenance windows and failures caused by Meridian’s own systems or third-party connections outside Northbridge’s control. Priority-one incidents would receive an initial response within thirty minutes and a named incident lead within an hour. The teams would agree a recovery objective for each service component during implementation rather than pretend one number fitted every failure.

“Why not promise resolution within four hours?” Martin asked.

“Because we cannot honestly promise that for every cause,” Priya replied. “We can commit to continuous work, updates every thirty minutes, and a recovery plan within two hours for a priority-one incident. For a fault in our platform, we can set a stronger resolution target. For a failure in a store’s network, a four-hour promise from us would be theatre.”

Martin studied her. “You are saying our obligation is to distinguish the cause before asking for a remedy.”

“I’m saying both teams need to cooperate in finding it. If our platform is at fault, we own the fix. If the evidence is unclear, we investigate together. Neither side benefits from arguing about attribution while stores wait.”

Rachel made a note. “We will need a mechanism for disputed attribution.”

“A technical review within five working days,” Sophie said. “Escalation to the service leads, then the commercial sponsors if it remains unresolved. No party loses the right to dispute a credit, but the operational response continues.”

Oliver looked at the figures. “What do we get if availability falls below the target?”

Sophie placed the proposed credit table beside the service schedule. Credits would rise with the severity and duration of a measured failure, calculated against the monthly service fee for the affected service. The total service-credit exposure would be capped at eight per cent of that month’s fee. Repeated failure would trigger a formal review and a right to terminate for material, uncured breach under the contract—not an unlimited credit that could exceed the value of the service.

“Eight per cent is not a strong incentive,” Oliver said.

“It is meaningful without turning the credit into a substitute for performance,” Sophie replied. “You also get transparent reporting, a recovery plan, and a review point. We are not asking you to accept a weak service. We are refusing to pretend money alone repairs it.”

Martin turned to Rachel. “Could we make the cap apply to the affected service rather than the whole monthly fee?”

Rachel checked the draft. “It already does.”

“Then we should make that plain in the schedule.”

The room’s tone changed by a degree. Not agreement, exactly, but attention. Sophie noticed Martin had not asked for a larger number simply to demonstrate authority. He had been looking for a remedy that mattered when stores were affected.

Oliver raised the broader liability clause. Meridian wanted Northbridge to accept exposure up to three times annual fees for any service failure that caused business interruption. Northbridge’s proposed limit was one year of fees in aggregate, with the usual exceptions for fraud, wilful misconduct, and other liabilities that could not lawfully be limited.

“We cannot accept three times the annual fee,” James said. He had joined the meeting to cover the financial and contractual boundary. “It would make the agreement commercially impossible to insure.”

“The operational consequence of a failure could be greater than the contract value,” Oliver said.

“It could,” James replied. “But a cap is not a claim that the loss cannot be larger. It is an allocation of risk that reflects price, control, and insurance. We can’t take an uncapped commercial exposure for an event caused by systems we do not own.”

Sophie saw Oliver prepare to answer and gave him space. He had a legitimate concern. The stores depended on the service; their losses would not be neatly limited to a monthly credit. She also knew that accepting a number Northbridge could not carry would make every other promise less credible.

Martin spoke first. “What can you improve?”

“Accountability,” Sophie said. “A named service director, an incident lead for serious events, a written root-cause report within ten working days, and quarterly reviews with both operational teams. For the liability cap, we can discuss a separate higher sub-cap for defined data-protection and confidentiality claims, subject to insurance and the exclusions in the draft. We cannot accept three times annual fees for every service failure.”

Rachel’s pen paused. “You have authority to offer a higher sub-cap?”

“Subject to Helen and our insurer confirming the limit.”

“That is not a number,” Oliver said.

“No,” Sophie replied. “It is the boundary of my authority. I won’t invent a number that Northbridge cannot honour.”

Martin closed his notebook. “That is preferable to a number you withdraw later.”

Oliver looked at him. “We still need an answer to the risk.”

“We need an answer we can buy,” Martin said. “A three-times cap on everything is not necessarily an answer.”

The discussion turned to responsibility. Meridian would provide clean reference data, name a product owner for each regional wave, and give Northbridge access to the relevant test environments. Northbridge would supply the implementation lead, migration tools, training materials, and a weekly readiness report. Each side would appoint a decision-maker capable of resolving ordinary scope questions without sending them to the executive level.

Priya proposed training two Meridian teams for each wave: the operational leads and a small group of local champions. “If we train everyone at the end, we create a large audience and no one who remembers the decisions. Training before go-live gives the people in each region a chance to test the process.”

“That adds work for our regional managers,” Oliver said.

“It does,” Priya said. “It also gives you a trained team at each acceptance point. If you cannot release those people, we can move the date, but not absorb the delay and still pretend the original timetable stands.”

“You are making our readiness a condition.”

“It is a responsibility,” Sophie said. “Just as our staffing and technical readiness are yours to rely on.”

There was no comfort in saying it. The sentence exposed the reciprocal nature of the deal: Meridian could not transfer its own implementation risk to Northbridge simply by calling it a service expectation. Northbridge could not claim that a reliable launch depended only on customer cooperation.

At the end of the meeting, Rachel agreed to revise the schedules. Oliver did not accept the proposed eight-per-cent cap, but he agreed to test it against the credit model. Martin asked for the named-accountability structure in writing and wanted the quarterly review to include trend data, not merely a report on the previous month.

On the train back, Priya sat opposite Sophie and watched the rain drag the city into grey streaks. “I think Martin understood the problem.”

“He understood the explanation,” Sophie said. “That is not the same as accepting the solution.”

“Do you ever let yourself enjoy progress?”

“After it has survived the redlines.”

Priya laughed once. “You are learning to be pleasant.”

“I’ll try not to overdo it.”

They reviewed the implementation responsibilities on a laptop balanced between them. Priya had marked several points where the commercial draft implied work without assigning an owner. One clause promised that “the parties will ensure readiness” but said nothing about who checked the data, who approved the test result, or what happened if a contact failed to attend.

Sophie replaced the sentence with a short table. Each task had an owner, an evidence requirement, and an escalation route. It made the agreement longer. It also made it harder for either side to hide behind the plural word “parties”.

By the time they reached Northbridge, James had sent a note: the revised model showed a viable, though thinner than target, margin if the three-year term and phased staffing assumptions held. He had added: “No spare capacity for an unpriced expansion.”

Sophie forwarded that line to Priya. “We need a change-control mechanism that works before somebody asks for a favour.”

“And a clear definition of a favour,” Priya replied.

Sophie saved the draft. Service improvement, she thought, did not begin with a dramatic promise. It began with a sentence somebody could act on at half past two in the morning, while a store waited and two teams tried to find the cause.

Chapter 20: The Draft Between Them

The first integrated draft arrived on Wednesday with more comments than white space. Rachel had reorganised the schedules, Oliver had challenged the price table, and Northbridge’s solicitor had returned the liability clause with a note in the margin: “Do not trade this without confirming insurance.” Sophie read the document at her desk, then took it to a meeting room before the track changes began to look like a quarrel.

Helen joined her with a folder. “David’s final view is that the deal is viable if the implementation assumptions are protected. He dislikes the payment concession.”

“Forty-five days?”

“He would prefer sixty. He accepts that the phased billing makes forty-five a reciprocal term rather than an unsecured loan.”

“And the margin?”

“Below our preferred threshold in year one. Within range over the term, provided we do not add scope without price or recoverable effort.”

Sophie set down the draft. “So we are accepting a lower first-year margin in exchange for a committed route to a stable account. Not a promise of future growth.”

“Exactly. Helen approved the commercial boundary, not every possible future.”

The clarification steadied Sophie. She had been hearing the phrase “strategic customer” used as if it were a financial instrument. It was not. Meridian could become valuable, but only if the work remained valuable to Northbridge.

James arrived carrying a spreadsheet. He had changed the staffing model to show a named implementation team, a controlled allocation of specialists, and a reserve that could be released only by approval. “I’ve accounted for the training and the review meetings,” he said. “The margin still holds, narrowly. But if Meridian adds a fourth regional wave, that is a change request.”

“Put it in writing,” Sophie said.

He glanced at her. “I have.”

The exchange was almost easy. Then he added, “You could have brought me into the price discussion earlier.”

Sophie set her pen down. “I could have. I was trying to keep the negotiation controlled.”

“It was controlled without me.”

“I thought you wanted the numbers to settle first.”

“I wanted to know when my assumptions were being used to make an offer.”

His voice was calm, which made the point harder to dismiss. Sophie could have explained that the previous weeks had not offered much room for trust. Instead, she recognised that the explanation would not change the effect.

“You’re right,” she said. “I should have brought you in when the structure changed, not after the shape was already decided.”

James waited, perhaps expecting a qualification. When none came, he nodded. “Next time, ask me before the proposal becomes a fact.”

“I will.”

He opened his spreadsheet. “Then let’s make sure this one is a fact we can deliver.”

The call with Meridian began at noon. Rachel had accepted the basic price architecture but wanted the annual fee stated in a way that could not be read as covering work outside the defined scope. Oliver still wanted the first year closer to his original target. Martin had asked both teams to present their remaining issues, not repeat their opening positions.

Sophie shared the fee table. The full annual run-rate would be £2.28 million once all agreed regions were live. During the first year, charges would follow the phased schedule: £2.08 million for the first twelve months, assuming the three waves reached their planned acceptance dates. If a Meridian readiness dependency delayed a stage, its launch date and related fee would move by the same agreed period. If Northbridge caused the delay, the teams would follow the recovery process and the corresponding charge would not begin before acceptance.

Oliver read the table. “The first-year figure is still above our ceiling.”

“It is above the number you asked us to meet,” Sophie said. “It is below the cost of delivering the full service from day one, because we are not delivering the full service from day one.”

“And the three-year total is higher than the annual number we began with.”

“Because it includes the support, training, reporting, and implementation capacity in the agreed scope. We can reduce the total by removing work. We cannot keep the work and pretend it costs nothing.”

Martin intervened before the conversation hardened. “Is there any scope we can defer without creating operational risk?”

Priya explained that the enhanced dashboard could be delivered after the first regional wave, but only if Meridian accepted a more basic monthly report during that period. The dashboard was useful; it was not essential to the initial service. Deferring it would reduce Northbridge’s year-one workload and bring the initial charge down by a modest amount, but the later release would be scheduled and priced within the existing scope.

Oliver asked whether that meant paying twice.

“No,” Priya said. “It means the dashboard is delivered in the second wave rather than the first. The total scope and fee do not change. The timing does.”

“Would that help your margin?” David asked from the Northbridge side.

“It makes the delivery plan less congested. It does not turn a poor deal into a good one.”

Sophie appreciated the honesty. A small operational adjustment could not be sold as a financial miracle. Oliver and Martin conferred quietly. Rachel wrote a note.

“We can accept the dashboard after the first wave,” Oliver said. “Provided the report includes the service measures and the regional exceptions.”

“Agreed,” Sophie said. “That is a specification, not an additional feature.”

They recorded it. The change improved the sequence and reduced a pressure point without changing the headline price. For a moment, Sophie felt the negotiation take on a quieter rhythm. Then Rachel raised the disputed clause on liability.

Helen had authorised a higher sub-cap for specified confidentiality and data-protection breaches, but only after confirming the limit with Northbridge’s insurer. The approved figure was twice the annual fee, subject to a separate aggregate cap and the statutory exclusions. It did not extend to ordinary service interruption. Sophie explained the distinction and sent Rachel the insurer-approved wording.

Oliver did not accept it immediately. “Our concern is still that service failure can cause losses far beyond the credits.”

“We understand,” Sophie said. “The credits are a prompt and a remedy for measured service performance. The liability cap allocates broader exposure. Those are related, but they are not interchangeable.”

Martin asked whether the service schedule could include a repeat-failure trigger, not just a monthly credit. Sophie proposed two consecutive months below target or three in any rolling six-month period would require a formal service recovery plan, executive review, and a cure period. If the failure remained material and uncured, Meridian would retain its contractual termination rights.

“That gives us a path to act before the year is over,” Martin said.

“Yes,” Sophie replied. “And it gives us a chance to fix the cause before the relationship becomes a calculation of credits.”

Rachel asked for the wording to be tied to the same measurement definitions used in the monthly report. Sophie agreed. It was a small point, but one that prevented the remedy from resting on a different definition of failure than the service promise.

By late afternoon, the open items fit on one page. The annual fee and phased schedule were agreed in principle. The three-year term was accepted, with termination rights for material breach and a mutual review before any optional extension. Payment at forty-five days was accepted, provided invoices followed the agreed evidence template and undisputed amounts remained payable while a specific item was investigated.

The eight-per-cent monthly service-credit cap remained under discussion. Meridian wanted credits to roll forward when a serious incident occurred near a month-end. Northbridge did not want one incident to generate an unbounded liability through a calculation repeated over several periods.

Sophie suggested a single credit assessment for the incident, measured by its duration and impact, with the credit applied to the next invoice if the month’s invoice had already been issued. The monthly cap would remain eight per cent of the affected service fee; a separate recovery and termination process would address repeated failures.

Oliver said, “That sounds tidy for you.”

“It is tidy for both sides,” Sophie said. “It means the same incident is not counted twice because it crossed a billing date. It also means a failure that continues into the next month is measured in both months.”

Rachel checked the example. “That is fair, if the end time is recorded consistently.”

Priya offered an incident log visible to both service teams. The log would record detection, acknowledgement, updates, workarounds, restoration, and the best-supported cause. Where the cause remained uncertain, it would say so rather than assign blame by default.

Martin nodded. “Put that in the service schedule.”

At six, they ended the call without signing. Sophie remained at the table while James collected his papers. “It’s nearly there,” he said.

“The important parts are nearly there. The wording still has to survive people who were not on the call.”

“That is what legal review is for.”

“And what commercial review is for.”

He gave her a sideways look. “You don’t have to make every sensible sentence sound like a warning.”

“I am working on it.”

The next morning, Priya sent Sophie a note with a list of unresolved implementation assumptions: Meridian’s test data was not yet clean; one regional lead had not been nominated; the training dates conflicted with a major retail period. The contract could not correct those facts, but it could name them and set a route to resolve them.

Sophie added a readiness gate before each wave. At the gate, both teams would confirm data quality, access, named contacts, training attendance, and a passed end-to-end test. Either side could raise a failed condition with evidence. The launch date would then be revised by agreement, and the impact on fees and staffing would follow the cause of the delay.

James read the change and said, “That clause could make us look as though we expect Meridian to fail.”

“It makes us look as though we expect reality.”

“We need to phrase it better.”

Together they did. The revised language described readiness as a joint control, with separate responsibilities and an escalation meeting within two working days of a failed gate.

Sophie sent the new draft to Oliver with the remaining issues highlighted. She did not write “final offer”. She did not call the document a compromise. She wrote: “This is the structure we can deliver on the assumptions shown. If any assumption is wrong, please identify it. The next discussion should be about the risk each clause leaves with each team.”

His reply arrived later that afternoon: “Agreed. We will review against our operating requirements and come back with the points we cannot accept.”

Sophie closed the email. They had stopped asking who could move furthest. They were beginning to ask who would carry what if the plan went wrong. It was a less triumphant question, and a more serious one.

Chapter 21: The Obligations

The review meeting was held in a smaller room at Northbridge, with the revised contract projected on the wall. Helen sat beside Sophie. James and David took the other side, while Priya joined in person, carrying a notebook crowded with implementation dates. Meridian had Oliver, Rachel, and Martin. No one brought a presentation.

Rachel began with the issues Meridian could accept. The three-year term, the staged rollout, the forty-five-day payment schedule, and the readiness gates were all agreed, subject to final wording. Oliver accepted the annual run-rate of £2.28 million at full coverage and the £2.08 million first-year schedule, provided the stage fees followed the acceptance dates rather than a calendar assumption.

“That is fair,” Sophie said. “The calendar is a plan. Acceptance is the event that starts the charge.”

“And a delay caused by our missing data does not create a free extension of Northbridge’s capacity,” Oliver said.

“Correct. The teams will assess the cost and timetable through change control if the delay requires work beyond the agreed plan. A simple rescheduling does not automatically create a new fee.”

David looked at Sophie, then at Oliver. “We should be precise about that. We will not charge for the same work twice. But if the delay requires a second migration exercise or additional deployment support, that is additional work.”

Oliver nodded. “That is defensible.”

For the first time in several weeks, nobody tried to turn “defensible” into “generous”.

They moved to scope. The contract covered the Northbridge platform for the agreed Meridian regions, the defined data feeds, standard support, the implementation work, the training programme, monthly service reporting, and the enhanced dashboard from the second regional wave. It did not cover additional acquisitions, new data sources, bespoke store hardware, or substantial redesign of Meridian’s internal processes.

Sophie explained the change-control process: a written request, an impact assessment covering price, timetable, risk, and capacity, and approval by named commercial owners before work began. For urgent operational changes, the service leads could authorise a limited temporary response, but only within a stated cap of hours and with a written record.

“I want to make sure the process does not become a wall,” Martin said. “If a store issue cannot wait for a committee, the team has to act.”

“They can act to contain an incident,” Priya said. “They cannot build a new permanent feature on the assumption that the contract will catch up later.”

Martin accepted that distinction. The agreement would allow immediate containment and investigation, while permanent scope changes required approval.

The service measures took longer. Meridian wanted availability measured across the whole service, including its own network connections. Priya showed how that would make Northbridge responsible for outages it could neither observe nor repair. Instead, the schedule divided the measure into the Northbridge platform, agreed interfaces, and customer-controlled elements. A shared dashboard would show each separately, with a combined operational view.

“That gives us three places to point fingers,” Oliver said.

“Only if the teams use it that way,” Priya replied. “The incident lead remains responsible for coordinating the response, even when the fault is outside our platform. Ownership of the repair is not the same as ownership of the conversation.”

Sophie added, “A Meridian-controlled fault will not count against Northbridge’s availability target. But we will not leave your team to manage it alone if it disrupts stores. We will support diagnosis and coordinate with the relevant supplier within the contracted scope.”

Rachel accepted the division, subject to clear evidence and a joint technical review where attribution was disputed. The service schedule now had an owner for every response, even where it did not assign blame.

The liability clause remained less comfortable. Meridian accepted the general cap at one year of fees in aggregate and the separate two-times annual-fee sub-cap for specified data-protection and confidentiality claims, subject to the stated exclusions. It would not receive a three-times cap for all service failures. Northbridge would not accept one.

Martin asked, “If the same serious failure happens repeatedly, what protects Meridian beyond the credits?”

“The recovery plan and termination trigger,” Sophie said. “Two consecutive months below target, or three in a rolling six-month period, requires executive review and a documented cure plan. If a material failure remains uncured after the agreed period, your termination rights apply.”

“And if it is fixed, then happens again?”

“The rolling test continues. We cannot make the service-credit cap disappear by calling the same failure a new event, but we can make repeated failure matter contractually.”

Oliver took a moment. “I can live with that if the reports are complete and the review includes the incident history.”

“They will,” Priya said.

The credit table was accepted at eight per cent of the affected monthly service fee, applied once to each measured incident, with continuing failures assessed across the periods in which they occurred. No credit would replace a recovery obligation or remove a right to terminate for material breach. No credit would be paid for an event outside Northbridge’s control, but the incident would remain in the record.

David asked for a break before confirming the final figure. In the corridor, he told Sophie that the margin was still narrow. “If the first wave runs late because our internal team is under-resourced, the model will tighten.”

“We have a named team and a reserve,” Sophie said.

“The reserve is not infinite.”

“Nor is our ability to absorb customer delays. The gates and change control are there for that.”

David’s face softened. “I believe the mechanism. I am less certain both organisations will use it early enough.”

It was a sensible doubt. A clause could tell the teams how to raise a problem; it could not make them feel comfortable doing so. Sophie said, “Then the first quarterly review should be about what is getting hidden, not only what has gone wrong.”

David gave a short nod. “Put that in the agenda, not the contract.”

Back in the room, Helen confirmed that Northbridge could proceed on the agreed figures, provided the scope remained as drafted. She did not call the agreement a win. She said it was investable if the delivery discipline matched the commercial discipline.

Martin confirmed Meridian could recommend the package to its finance committee. The lower opening price they had originally requested was not achieved. In return, they had secured a phased service, defined accountability, measured improvements, capped credits, training, and review points. They would pay more as coverage expanded, and they would accept that their own readiness affected the dates.

“There is one more point,” Oliver said. “Our internal review expects a response commitment beyond the initial incident lead. We need to know when a senior manager becomes involved.”

“For a priority-one incident, the service director joins if there is no viable workaround within two hours,” Sophie said. “For a continuing issue, the executive sponsors are notified at the recovery threshold. We should not escalate every fault to people who cannot fix it.”

“Agreed,” Martin said. “But the notification threshold should be objective.”

Priya proposed a combined trigger: a priority-one incident still without a workaround after two hours, or a confirmed risk to multiple regions. Either would prompt the service director and Meridian’s operational sponsor to join the incident call. A further executive escalation would occur if the issue remained unresolved at the agreed recovery review.

Oliver accepted it.

The contract was now sufficiently complete to pass through legal review without pretending every uncertainty had vanished. Rachel and Northbridge’s solicitor would check definitions, cross-references, and consistency. The commercial teams would confirm the schedules. A final signature date was pencilled in for the following Friday.

When the Meridian group left, Sophie stayed to collect the printouts. James picked up the service schedule. “You held the liability line.”

“We offered accountability instead.”

“Which is easier to measure and harder to deliver.”

“Yes.”

He paused. “I was wrong to think the boundary was the whole negotiation.”

Sophie did not answer at once. “And I was wrong to treat your caution as a refusal to move.”

His expression suggested he had not expected her to say it. “We both make poor use of assumptions.”

“Apparently.”

They worked together on the final internal note. James listed the financial conditions: the agreed term, the scheduled fees, the forty-five-day payment cycle, no unapproved scope, and a review of margin after each implementation wave. Sophie set out the operational commitments. Priya added the capacity plan and the named training leads. David asked them to include the cost of keeping a reserve team available.

“That cost was in the model,” James said.

“Then make it visible,” David replied. “If we hide the work, we will later call it unexpected.”

Sophie watched James add the line. It was a small act of repair, though not an apology. Neither needed to make a scene of it.

That afternoon, she called Oliver to confirm the signature process. He was between meetings and sounded distracted. “Rachel says legal can finish the review by Thursday. Martin has the committee slot Friday morning.”

“If the wording holds,” Sophie said.

“It should. I’m not going to pretend I love the price.”

“You shouldn’t.”

There was a faint laugh at his end. “You may be the first person who hasn’t asked me to say I do.”

“I’d rather you knew what you had agreed to.”

“So would I.”

They discussed the first readiness meeting, the name of Meridian’s regional lead, and the information required before Northbridge could validate the test data. Oliver admitted that one region’s records were likely to arrive late. Sophie did not use the information as leverage. She asked for an estimate and a named owner so Priya could assess the impact.

“I’ll get you an answer tomorrow,” he said.

“Thank you.”

After the call, Sophie opened the project workspace and created a list of obligations by owner. It was not the satisfying final document she had imagined when the negotiation began. It was a work plan, a risk register, and a set of promises that could be tested.

Helen stopped by her desk. “Are you pleased?”

“I’m relieved that the numbers and the operating plan agree.”

“That is not quite the same thing.”

“No. We will find out what we have agreed when the first gate fails.”

Helen considered her. “Then make sure there is a way to say so before it fails.”

Sophie looked at the list of names, dates, and evidence requirements. “I’ll try.”

Chapter 22: What We Can Honour

The agreement was signed on Friday afternoon, after one last disagreement about the meaning of “valid invoice”. Rachel wanted a definition that prevented Northbridge from delaying the clock with administrative defects. David wanted enough evidence to verify that each stage had been accepted. They settled on a short list: the relevant stage, the agreed fee, the acceptance record, and the purchase-order reference. A disputed item would be identified within ten working days; undisputed amounts would remain payable on time.

It was not the sort of clause anyone would mention in a speech. It was, Sophie thought, the kind of clause that saved a relationship from becoming a sequence of resentments.

Helen signed for Northbridge. Martin signed for Meridian. The renewal covered three years, the agreed regions, and a phased implementation. It included the £2.08 million first-year schedule and the £2.28 million full-service annual run-rate once all regions were live; payment at forty-five days; the defined service measures; training and reporting; the eight-per-cent monthly credit cap; the liability limits; the responsibilities and readiness gates; the change-control process; and the escalation route for incidents and disputes.

No one applauded. Martin said, “We have a workable plan.” Helen replied, “Now we have to work it.”

Oliver gathered his copy. He looked at Sophie across the table. “I still wanted the lower price.”

“I know.”

“You were right not to sell me a number the service couldn’t support.”

Sophie heard the qualification he did not add. He had not said she was right about everything, and she did not need him to. “And you were right to keep asking what happens when the service fails.”

“I expect we’ll have to find out.”

“I expect we will.”

The first implementation meeting was booked for Monday. Before she left, Sophie watched Priya and Oliver compare calendars. They were already arguing about whether a regional lead could attend training during a stocktaking week. Their voices were practical, mildly impatient, and more familiar than they had been a month earlier.

On Monday morning, Northbridge’s internal launch call began with a spreadsheet rather than a celebration. Priya assigned the implementation lead. James confirmed the dedicated team and the reserve allocation. David reviewed the first-year cash-flow profile and reminded them that the agreed margin depended on controlling unplanned work. Helen asked who would report a scope concern before it became an invoice dispute.

“I will,” Sophie said.

“And if you think the customer will be unhappy?”

“I’ll still raise it.”

“Good.” Helen moved to the next item.

The first readiness meeting with Meridian was less smooth. Their test data had arrived, but a field used to identify store regions was inconsistent. Meridian’s local lead, Martin’s nominee, admitted he had not yet received approval to change the source file. Priya showed the examples and explained why the mismatch would contaminate the migration.

Oliver’s voice was tight. “Can Northbridge clean it?”

“We can map a limited set of known variations,” Priya said. “We cannot safely infer which of these records are wrong. That requires Meridian to confirm the source.”

“How long?”

“If we get the corrected file by Thursday, the test can proceed next week. If not, we move the gate and review the effect on the first wave.”

Nobody liked the answer. Sophie could feel the room preparing to search for a softer one. The old version of herself would have said Northbridge would “see what it could do”, hoping the promise bought time. This time she asked, “Who owns the correction, and when can they confirm?”

Oliver named the data manager and promised an answer by the end of the day. He did not blame Priya for the delay. Priya, in return, offered a short working session with the data team to explain the mapping rules. The work was not free of tension, but it was visible.

When the call ended, Priya looked at Sophie. “You didn’t jump in.”

“You had it.”

“I know. It’s useful to notice.”

“I’m practising.”

The first week produced no dramatic success. Meridian delivered a corrected file late on Thursday. Priya’s team worked through it on Friday, and the test moved by three days. The teams recorded the cause, revised the readiness calendar, and agreed that the first wave’s acceptance date would move accordingly. The fee schedule remained linked to acceptance, so the later start did not create a dispute over a calendar charge. The reserve team was used for two days, a cost James had included in the model but would rather not have spent.

By the next review, the migration test had passed. The service report showed one minor incident and no breach of the agreed availability target. Oliver thanked Priya for the data session. Priya replied that the corrected file had made the difference. Sophie noticed how carefully each of them gave credit: not as politeness, but as a record of what had changed.

The relationship was not repaired by signing. It was repaired in small exchanges where people could admit a dependency without surrendering their position.

Three weeks later, a new issue arrived from Meridian’s side. A regional operations manager wanted an additional field in the dashboard before the second wave. The request was reasonable, and the field would help stores understand an exception. It was also outside the agreed specification.

The message came to Sophie with a note from Oliver: “Can we include this as a small adjustment? It should be straightforward.”

James read it over her shoulder. “That is how unpriced scope starts.”

“It may be straightforward.”

“It may also require a data change, testing, and a new report definition.”

Priya had already assessed it. The field was feasible, but Meridian’s source system did not provide it consistently. Northbridge could add it after the second wave, once the source data was reliable. Doing it now would require manual exceptions and extra testing. The cost was not enormous, but neither was the requested change. The teams had agreed to follow change control for exactly this sort of question.

Sophie called Oliver. “We can do it. We cannot include it in the current release without changing the test plan and taking on a data-quality risk.”

“It’s one field.”

“One field in the display. Several assumptions behind it.”

“What are you proposing?”

“We can assess it as a change request for the second wave, with the cost and timetable set out. Or we can include it in the later dashboard release, once the source is stable, with no change to the fee if it remains within the agreed reporting scope.”

“You are asking us to wait.”

“Yes. And we are offering to validate the requirement now so that waiting does not mean forgetting it.”

Oliver was quiet. “The regional manager will say you are making the contract more important than the customer.”

“Then let’s show him the risk and the choices. If he still wants it earlier, he can approve the change.”

The following day, Sophie joined a call with Oliver and the regional manager. Priya described the inconsistent source values without making Meridian’s team sound careless. She showed how the report could misclassify stores if Northbridge guessed at the missing data. The regional manager listened, then asked whether a temporary manual flag could be used for the first month.

“It could,” Priya said, “if Meridian owns the manual entries and confirms them weekly. We would need to record them separately from the automated report.”

Oliver looked at Sophie. “Would that be within scope?”

“A limited temporary workaround is within the change process we agreed. We can authorise it up to the stated effort cap. If it requires more, we pause and agree the next step.”

The regional manager accepted the condition. Priya documented the work; Oliver nominated the Meridian contact; James approved the hours. The requested field appeared in the report the following month, marked as provisional until the source system was corrected.

It was not a grand demonstration of the contract. It was a demonstration of what the contract was for.

That evening, Sophie found an old note in her notebook, written during the hardest stretch of the negotiation: “Don’t let them move you.” She crossed it out and wrote beneath it: “Know what you can move, and what must move in return.”

The change did not make negotiation gentle. It made movement legible. Price could move when term, scope, timing, or risk moved with it. A service promise could improve when the measure and the remedy were clear. A boundary could remain firm without becoming a threat. And a customer could be heard without being granted every request.

At the end of the quarter, the teams held their first joint review. The dashboard showed steady availability, one missed response target, and a migration delay caused by a late data correction. The service director presented the incident record. Oliver questioned the delay; Priya explained the dependency. The two of them agreed on a revised data hand-off date and an earlier check with Meridian’s source team.

Martin asked what had surprised them. Oliver said he had expected the service schedule to be a list of reasons for refusing responsibility. Priya said she had expected the customer to treat every delay as Northbridge’s fault. Both statements were delivered without accusation, which made them more honest.

Sophie listened. She had no neat conclusion to offer. The agreement had not removed disappointment or risk. It had given the teams a way to meet those things without inventing a new bargain every time they appeared.

After the review, Oliver stopped beside her at the lift. “I still dislike the liability cap.”

“I know.”

“I also know what the service director does when a store has a problem. That matters.”

“Then we should keep doing it.”

He smiled, briefly. “That is a very commercial answer.”

“It is the answer I can honour.”

The lift arrived. They stepped inside with their papers and their separate calendars, neither side having won the argument in the way they had once imagined. The work ahead remained considerable. Sophie knew Northbridge would have to guard its capacity, Meridian would have to meet its own readiness commitments, and both teams would have to speak early when an assumption failed.

But the agreement had been made to bear weight. That, she thought, was the art of the deal: not the triumph of one position over another, but the patient design of terms that people could honour when the room was empty and the work had begun.