Everything Is on The Menu

The first message appeared in #warehouse-ops:

Mikael (Warehouse Manager): "Urgent. Physical inventory count shows 203 units of SKU-4782.

The system shows 500. Please advise."

Sarah Chen, CEO of Stock Dropper, was reviewing Q4 projections with George, her CFO, when the second message arrived in #customer-service:

CS Team Lead: "Getting multiple calls about canceled orders. Valentino handbag SKU-4782, flash sale. Customers said they were charged, but orders disappeared. What's happening?"

Then her phone started vibrating.

Marcus. Elena. Carlos. All are calling simultaneously.

She answered Marcus first.

"Sarah, we have a situation. In the flash sale, we oversold. Badly."

"How badly?"

"We sold 500 Valentino handbags in six hours. We only have 203."

Sarah felt her stomach drop. "How is that possible?"

"I don't know. But we have 297 customers who paid €1,200 each for bags we can't deliver."

The math hit immediately. €356,400, and €89,100 in gross profit, on orders they couldn't fulfill.

"Get everyone in the management meeting room. Now."

Sarah, Marcus, Carlos, Elena, and Deborah crowded into the conference room.

"Walk me through what happened," Sarah said.

Deborah pulled up her laptop. "The flash sale launched Thursday evening. By Friday morning, we'd sold 500 units. The website showed 500 available, so the system accepted all orders."

"But we only had 203," Carlos said, voice tight. "My warehouse team did a physical count this morning because something felt off, and updated the system."

"Why didn't they match?" Sarah asked.

"The warehouse and the website track inventory separately," Deborah said. "An automated process keeps them in sync. Somewhere, that process stopped working."

"When?"

"I don't know yet. We're investigating."

Marcus leaned forward. "What do we do about the 297 customers who already paid?"

"Refund them," Sarah said immediately. "Full refund plus a 20% discount code. Send personal apology emails. Do whatever it takes."

"That's €356,000 in refunds," George said quietly. "Plus, maybe another €50,000 in discount redemptions. Call it €400,000."

The room went silent.

"Get me answers by the end of the day," Sarah said to Deborah. "I need to know what broke, when it broke, and how we fix it."

Sarah noticed John, their CTO, wasn't in the meeting room. He was in Belgium speaking with regulators. She texted him. He replied: "Deborah has it. Keep me posted."

That was their dynamic. John hired Deborah, delegated data to her, and focused on infrastructure. Sarah trusted it without watching it

Deborah found Sarah in her office.

"I know what happened."

"Tell me."

"Three weeks ago, we began rebuilding our central data storage. Think of it as renovating the pantry while the restaurant is still open. The automated sync was supposed to continue during the rebuild, but it failed on day two of the migration. The warehouse and website have been out of sync for three weeks."

"Three weeks?" Sarah's voice went sharp. "Why didn't anyone notice?"

"The team has been focused on the migration. It's complex work. They were monitoring the new system, not the old connections."

"Deborah, we just lost €400,000 because nobody was watching the basics."

"I know." Deborah's voice was tight. "The team is underwater. We're rebuilding the data infrastructure, implementing governance, fixing legacy issues, and building new analytics for Marketing, Product, and Operations. Everyone's working 60-hour weeks, and things are slipping through the cracks."

Sarah felt frustration rising. "How many things are slipping through?"

"I don't know. We're firefighting constantly."

"That's not an answer."

"It's the only answer I have right now."

She got home at seven. She didn't remember the tram ride.

She sat on the edge of the bed still in her coat. She'd meant to change. She hadn't moved.

Her phone had seventeen unread messages. She knew without looking: Carlos, George, Marcus, someone from PR, Deborah twice. She put it face down on the duvet.

She noticed her hands first. The particular stillness of them in her lap, like they belonged to someone else. Then her chest, not pain, just pressure, the sensation of something sitting on it that wasn't there. She counted her breaths without deciding to. One. Two. She got to eleven before her phone lit up again.

She left it.

She sat there until the room went dark.

Then she got up, still in her coat, and went to the kitchen and stood at the window looking at the street below.

She didn't know how long she stood there.

Eventually she made tea she didn't drink and opened her laptop and started reading the incident report from the beginning.

She needed to understand every line of it before tomorrow.

Sarah called an emergency meeting. Everyone was already in the room before she arrived.

Carlos opened before she sat down.

"We refunded 297 orders overnight. €356,400 went out the door. My warehouse team is manually reconciling every order from the last three weeks. We have twelve people who should be receiving stock doing paperwork instead."

"Discount codes?" Sarah asked.

"€18,000 redeemed so far," George said. "Still counting. Call it €400,000 total when it settles."

"NPS?" Sarah looked at Marcus.

"Down eight points week over week. We're getting it in the reviews too. 'Charged for something that didn't exist.' That's the phrase I keep seeing." He paused. "We have another flash sale on Thursday. I cannot promote it until someone tells me the inventory numbers are clean."

The room turned to Deborah.

"The sync has been down for three weeks," Deborah said. "We know that now, which means every sale in that window needs to be audited." She added, "We haven't fixed the issue yet. We need more time."

"How long?" Elena asked.

"Three days minimum," Tina said from the corner. "Possibly longer depending on what we find."

"There's a bigger problem," Deborah continued. "This wasn't the only thing running on that connection. The warehouse forecasting model and the reorder alerts for Carlos's team were pulling from the same pipelines. We don't yet know what else was affected."

Carlos put his pen down. "What does that mean for Thursday?"

"It means I can't give you a clean number by Thursday," Deborah said. "Not one I'd stake a flash sale on."

"Then I'll have the warehouse team do physical counts before we go live," Carlos said. "Manual reconciliation. Every SKU."

He looked at Sarah. "This is the third time this quarter, Sarah. August was attribution. September was churn. Now this."

Sarah looked at Deborah. "Third time?"

"We reported them in the weekly standup," Deborah said. "They were on the slides."

Sarah thought back. She vaguely remembered slides that mentioned incidents. She'd been focused on the new capabilities they were building. not the maintenance work.

She felt something cold settle in her chest. She'd been approving new projects without realizing the team was drowning just trying to keep existing systems running.

"We need to talk about workload," Sarah said. "Set up time on Wednesday. Just you and the team."

Before the team meeting, Carlos stopped by Sarah's office.

"I need to show you something," he said, pulling up his email. "Three weeks ago, my warehouse team sent this to the data team."

Sarah read the message:

Hi @Tina, we’re seeing inventory numbers in the system that don't match our physical counts.

Not huge differences, but consistent. Can someone look into this? Not urgent, but wanted to flag. Melan

"Did they respond?" Sarah asked.

"Five days later. They said they'd look into it after they finished the data warehouse migration. That was three weeks ago. "Sarah stared at the message. The warning had been there. Nobody had acted on it.

"Why not?"

Carlos shrugged. "I'm guessing they get a hundred messages like this. Everything is 'not urgent' until it explodes."

Sarah joined the Zoom link two minutes late. Deborah, Tina, Omar, and Lisette were already on the call. Four small boxes. She could see Tina's desk behind her, two monitors, a coffee mug, and a stack of notebooks she hadn't touched.

"I'm just going to listen," Sarah said. "Pretend I'm not here."

Nobody pretended.

Deborah ran the retro the way she always did. What shipped. What broke. What needed to change.

"Warehouse migration," she started. "Status?"

"Sixty percent," Tina said. "We were supposed to be at eighty."

"Why the gap?"

"The inventory incident dragged on for a few days this week. Before that, the attribution dashboard broke again. The week before, the churn model needed emergency fixes." Tina's voice was flat. Not angry. The flatness of someone who has said this before. "We get maybe two hours of focused work on the migration on a good day." "How many hours are you working?" Deborah asked.

"Fifty to sixty. Each."

Sarah opened her mouth.

"We should see if we can hire," she said. "I'll talk to George about opening a position."

Silence on the call. The kind that meant everyone was being careful.

"We had four people until last month," Omar said quietly. "Ana left. We haven't backfilled yet."

"Right," Sarah said. "A backfill first, then."

"Backfill won't fix it," Deborah said. Then, more carefully: "We'd need to know what we're hiring for. Right now, everything is high priority. A fourth person just joined the firefighting."

Sarah wrote that down. Define priorities before hiring.

"How do you decide what to work on?" she asked.

A pause.

"Whatever escalates," Omar said.

"We try to follow the sprint plan," Lisette added. "But stakeholders go around us. They message John directly. Sometimes they message us directly. By Wednesday, we're off-plan."

Sarah wrote: Bypass problem. Stakeholders are not respecting the process.

"We've discussed this in retros before," Deborah said. She was looking at the screen carefully. "It's on the last four retro boards. Same issue."

"And what happened?"

"We documented it," Tina said. "Added it to the list of things we need to fix."

"But nothing changed," Sarah said.

"No," Tina said. "Nothing changed."

Sarah looked at her notes. Two pages. Fourteen active initiatives. A team working sixty-hour weeks. Stakeholders bypassing the process. A migration four weeks behind schedule.

We need better governance, she thought. Clearer policies. Ownership. A framework for intake requests.

She didn't say it out loud. But she felt, for the first time, like she was beginning to understand the shape of the problem.

"Thank you for letting me sit in," she said. "I'll get back to you on headcount."

She closed the laptop.

Sarah sat in her home office, staring at the Slack message Carlos had flagged. Three weeks ago. The warning about the inventory data. Ignored.

Her phone buzzed. George Bergman.

She answered.

"Sarah." His voice had that tight control she'd learned to read. Not angry. Controlled. Which was worse. "I've been on two calls this afternoon, first with Christoph, then with Werner from the investors board."

She waited.

"Werner requested a written incident report by Friday. Not next week. Friday." A pause. "He also asked me something I didn't have a good answer for."

"What?"

"He asked whether this was a one-time failure or a pattern. I told him I'd look into it." Another pause. "So, I looked into it. All in the same function. All in the same year."

Sarah said nothing.

"Christoph mentioned something else," George continued. "He said there were warning signs. The analyst he set to support us saw system alerts going back six weeks. It was also flagged by John in your Tuesday sync."

The silence stretched.

"Is that true?"

A small pause. "Yes," Sarah said quietly.

"So, the system was degrading for six weeks. The warnings went to people in the coordination meeting. And nobody told Deborah."

"That's correct."

George exhaled slowly. "Sarah, I sat in that meeting room on Monday morning and thought we had a technical failure on our hands. A pipeline broke, and things fell through. I've been defending that position to Werner all day." His voice dropped. "This isn't a technical failure. This is a governance failure. And it's been building for months."

"I know."

"Who's accountable?"

"John and I are accountable," Sarah said. "I created the meeting that excluded her. I approved the projects without asking what the team could carry."

"That's not what the board wants to hear on Friday," George said. "They want to know what changes. Specifically. With a name attached to it."

"I'm working on it."

"How long has Deborah been in the role?"

"Eighteen months."

"And before her?"

Sarah closed her eyes. "Three data leaders in four years."

The silence was different this time.

"Sarah." His voice had lost the controlled edge. What replaced it was almost worse, something tired. "The board is already nervous about your operational maturity at this scale. With the recent data incidents and the €400K preventable loss, we are looking at a Series C issue. This is not a story I can dress up for Werner. It's a CEO problem. You know that."

"I know."

"Friday. Written report. What broke, why it broke, and what structural changes are needed to prevent it from happening again. Include a person's name and a structural change."

The call ended.

Sarah sat in the dark. Her hands were shaking.

She thought about what George had just said.

He'd meant it as a warning about what the board needed to hear.

But it landed differently.

She couldn't disconnect, so she opened her laptop.

Pulled up her calendar. Started scrolling.

Tuesday, 9:00 AM, Ops Sync, Recurring bi-weekly.

She'd started that meeting after the second data leader resigned. Cross-functional coordination. Operations, Finance, Product, Marketing. Keep everyone aligned.

She clicked into the meeting notes. Skimmed.

March 12, Carlos: Warehouse sync showing latency issues. Action: Data team to investigate.

March 26, Carlos: Sync delays are increasing. Action: Escalate to Deborah.

April 2, Carlos: Warehouse connection is unstable on weekends. Action: Carlos to monitor, flag if critical.

April 23, Carlos: Sync throwing intermittent errors. John: Will check with Deborah.

Sarah stopped reading.

For four months, they'd been discussing data infrastructure problems without the people who built it.

Sarah's phone rang. Dr. Eva Rousseau - Board chair.

"Sarah. I saw George's calendar invite for the emergency board meeting. What's happening?"

She told her. The incident. The board pressure. The decision she had to make.

"Are you firing Deborah?" she asked.

"I don't know what else to do."

"Walk me through the pattern."

"Three data leaders in four years. Constant firefighting. Projects that never finish. Incidents that shouldn't happen. Team working 60-hour weeks."

"And what changed between each leader?"

Sarah paused. "We hired different people. Deborah wasn't a blocker or as toxic as the others."

"What stayed the same?"

The question landed like a punch.

"Everything else," Sarah said quietly. "The projects. The stakeholders. The way we make decisions." The bad data.

"Sarah, do you know what the data team's capacity is right now?"

"They're stretched thin."

"That's not capacity. That's a feeling." She paused. "Do you know how many things they're maintaining right now, not building, that no one in the business uses?"

Sarah said nothing.

What solution do you have in mind? Do you think hiring more people will solve the issue?"

"Governance. From what I see, we need to have better accountability and processes in place."

"I recommended governance to a company two years ago," Dr. Rousseau continued. "They brought in a firm. Built a beautiful framework. Spent €180,000."

"Did it work?"

"The team is still drowning. Governance tells you how to manage what you have. It doesn't help you see what you should never have kept running." "I know someone different. His name is Arjun Mehta."

"A consultant?"

"A coach. For you, not your team. Because right now, Sarah, you and John are the bottleneck."

Sarah bristled. "I'm not."

"You're running Tuesday meetings about data infrastructure without having the knowledge or background for it. You're approving fourteen projects with three people. You're asking for strategic work while the team is drowning in operational fires. You're a brilliant CEO. But in this one area, you're flying blind. And you keep crashing the plane."

Silence.

"Can you send me his information?" Sarah asked.

"Sarah, don't hide behind Deborah's performance. If the data strategy was flawed from the top, firing her won't fix your reputation with the shareholders. Let's be clear about what we're doing here."

The call ended.

John called.

"I heard from George. I want to push back."

"Go ahead."

"She fixed things. When I hired her, the team was in open revolt. She reduced that. The stakeholder relationships aren't hostile anymore. Those are real changes, Sarah. They took real work."

"I know."

"So we sacrifice someone who's actually making progress to tell a better story."

"We change direction to have a chance at the round." A pause. "And John, are you confident she can get us where we need to go? Not where we were. Where do we need to go?"

Silence.

"No," he said finally. "What we need now is someone who thinks about data as a product, not a service. That's not Deborah."

"That's not Deborah," Sarah agreed.

"Tina. For now. She's been running governance, she knows the ownership model. But sixty days, Sarah. Not more. She's an engineer, not a leader."

"I will keep it in mind."

"And the governance structure is right. The accountability behind it isn't. That still has to happen."

"I know."

"Then do it kindly. She didn't fail. The system did."

The call ended.

Sarah sat for a long time before she opened Deborah's email.

Sarah opened a new email.

To: deborah.rodriguez@stockdropper.com

Subject: Meeting Tomorrow

Deborah,

I need to meet with you tomorrow at 9:00 AM.

I'm sorry for the last minute.

Sarah

She read it a second time. Hovered over Send.

Then she sat in the dark

I did this.

At 11:17 PM, she finally went to bed.

She didn't sleep.

Deborah had been up since 5:00 AM running diagnostics on the warehouse sync. She pulled up the system logs. Scrolled back. One week. Two weeks. Three.

Then she saw it.

Warehouse sync warnings. Dozens of them, going back six weeks. Intermittent failures. Data quality alerts. Capacity warnings. All flagged in the monitoring system.

All sent to operations@stockdropper.com.

She pulled up her email. Searched for anything from Operations about sync issues.

Nothing.

She walked to Antoine's desk. "Did Operations receive warnings about the warehouse sync? Going back six weeks?"

Christoph looked uncomfortable. "We mentioned it in the Ops sync meeting. John said he'd take care of it."

"What Ops sync meeting?"

"The Tuesday morning stakeholder sync. Sarah hosts it."

Deborah felt something cold settle in her chest. "I'm not invited to that meeting."

Christoph looked away. "You should talk to Sarah."

They'd been meeting about her team's systems for four months without her.

She opened her laptop and started typing her resignation letter. Got three sentences in before the calendar alert popped:

Sarah, 1:1, 9:00 AM, Today.

Not their regular slot. A special meeting.

She knew.

Deborah sat down and set her bag beside the chair.

Not behind her. Beside her. Like she wasn't planning to stay long.

"I was going to resign this morning," she said. "I had the letter open when your calendar invite came."

"I think we both knew this was coming."

"Yes."

A beat. Not hostile. Something more honest than that.

"Then let me say something first," Sarah said. "Before the paperwork."

Deborah waited.

"In four years, you're the first leader I've had in this function who wasn't a blocker. The two before you, every conversation was about why something couldn't be done. You never did that. You always found a way."

Deborah looked at her for a moment. Something moved across her face that Sarah couldn't read.

"I know," Deborah said quietly. "That's why the team broke."

Sarah opened her mouth. Closed it.

"The leaders who say no get labeled difficult," Deborah continued. "Toxic. Not aligned with the business. They get fired in six months for being blockers." A pause. "The leaders who say yes get fired in eighteen months when the team collapses. The only difference is how long it takes."

She looked at Sarah steadily.

"You just told me my biggest flaw like it was a compliment."

Sarah didn't have an answer for that.

"The board wants a leadership change to prepare for Series C. So, the question is whether we do this honestly or we perform it."

"Honestly."

"Then I need to hear what I've been missing. Before you go."

Deborah thought for a moment.

"The ownership model is broken," she said. "When I came in, data was entirely centralized. Every department treated us like a vending machine. So, I pushed in the other direction, gave the data producers ownership of the data they produce."

"And it's not working."

"The idea is right. But the execution lacks accountability. The operations team receives sync alerts and filters them as noise. The product team changes a schema without telling us, and three pipelines break. You can't give people ownership without giving them the reason to care about it."

"So, we need governance on top of the ownership model."

"We need governance, yes. Clear accountability. Clear consequences when alerts get ignored. Clear escalation paths." She stopped. "I was trying to build that. But every time we tried to enforce it, every time we sent an alert back to the team that owned the broken pipeline, we got pushback. And without leadership behind us, we couldn't hold the line."

"The Tuesday sync," Sarah said. "That was part of it."

"The Tuesday sync is a symptom," Deborah said. "The disease is that this function has never had a seat at the table with real authority. John sees us as infrastructure. He cares about uptime, tools, and architecture, but he doesn't ask whether the marketing attribution model is actually driving better decisions."

"You needed a different kind of sponsor."

"I needed a CDO who understood that data is a strategic function. Not an IT function. Not a reporting service." "I don't say that to blame you. You didn't know what you didn't know. Neither did I, when I took this job."

A silence.

"Tina," Sarah said. "Interim. Thirty days while we recruit."

"Tina can hold the systems together," Deborah said. "She won't want it. She'll say she's an engineer. She is. But she's also the only one who knows where the bodies are buried."

"Will she do it if you ask?"

"She'll do it if I ask and if she sees you've actually changed something. Not just the name in the org chart." Deborah looked at her directly. "Have you?"

"I don't know yet," Sarah said honestly. "I'm trying to figure out what needs to change."

"That's more honest than I expected."

They moved to the paperwork. Three months' severance. A reference Sarah meant. A transition document that was realistic about what two weeks could cover.

At the door, Deborah stopped.

"What you're looking for next," she said. "Someone with product experience, yes. Someone who can look at our data assets the way a product manager looks at a feature roadmap, what's creating value, what's dead weight, what needs to be retired." "But that person will fail too if the system doesn't change."

"I understand."

"Do you?" Not aggressive. Genuinely asking.

"I'm starting to."

Deborah nodded once. Not satisfied exactly. But done.

"The data team doesn't need a hero," she said. "They've had heroes. They need a system." She looked at Sarah steadily. "Not a governance framework. A system."

The door closed.

Sarah sat alone.

She opened her notebook. Wrote:

Governance, enforce the ownership model

Tina, 30-day interim

Tuesday sync, fix the structure

She underlined governance twice.

She thought it was the answer.

Sarah closed her laptop and walked to the data corner.

She didn't call Tina to her office. She went to Tina's desk. That was important. You don't summon someone when you're begging.

Tina, a senior data engineer, was staring at her monitor, typing SQL aggressively. She didn't look up.

"Tina."

The typing stopped. Tina turned. Her face was a mask of betrayal. "She was the best manager we've had."

"I know."

"Who's next? Who's the next lucky winner?"

"No one," Sarah said. "Not yet. I need you to step in. Just for thirty days. Interim Head."

Tina laughed. Short, harsh. "Absolutely not. I saw what this job did to Deborah. I'm an engineer. I build pipelines. I don't do the politics."

Sarah pulled up a chair and sat. She was lower than Tina now.

"I'm not asking because I want to elevate you," Sarah said. "I'm asking because I'm terrified. If you don't take the helm, Operations goes blind tomorrow."

Tina was quiet.

"Give me thirty days," Sarah said. "Help me keep the lights on while I figure out what to fix."

Tina looked at her for a long moment. Then she leaned back.

"Can I ask you something?"

"Yes."

"What does the data team work on this week?"

Sarah opened the folder she was holding. "I have the fourteen initiatives."

"I'm not asking what's on the list," Tina said. "I'm asking what we actually work on. This week. Today."

Sarah looked at her.

"Right now," Tina said, "Omar is fixing the revenue reconciliation pipeline after George messaged him directly at seven this morning. Lisette is rebuilding a campaign report after Marcus told her it was blocking a board presentation. I was supposed to be working on the warehouse migration, which is four weeks behind, but I've spent this morning on the inventory sync because you and Carlos need a status update by noon for the Friday report."

She waited.

"None of that is on your list," she said. "None of it was planned. All of it came from a senior deciding their problem was the most urgent in the company and going directly to whoever they thought could fix it fastest."

Sarah said nothing.

"When you ask me what the team works on," Tina continued, "the honest answer is: whatever the loudest person asked for last. That's not a backlog. That's not a sprint plan. That's just whoever escalated most recently."

"That's why we need a proper intake process," Sarah said. "Clearer governance around requests."

A silence.

"Did you know that?" Tina asked.

Sarah looked at the folder in her hands.

"Not in those words," she said.

"How do you decide which of the fourteen is most important?" Tina asked.

"They're all." Sarah stopped herself.

Tina waited.

"I don't have a system for that," Sarah said quietly. "I approved them because they all had good reasons."

"Right," Tina said. "And we don't have a system either. So we run on whoever escalates. That means we run on emotion. That means we never finish anything because there's always a newer, louder emergency."

She looked at Sarah directly.

"That's what you're asking me to manage for thirty days."

"Yes," Sarah said. "I know how that sounds."

"Then why would I say yes?"

"Because if you don't," Sarah said, "Omar and Lisette are going to spend the next thirty days getting direct messages from every C-level in the building. With no one to absorb it. With no one to at least try to hold the line." "You can't fix it. I know that. But you can stop it from getting worse while I figure out what to fix."

A long silence.

Tina looked over at Omar's desk. He had his headphones on, the posture of someone who had stopped expecting help.

"Thirty days," she said finally. "I'm not doing the budget. Or the board slides. I keep systems running."

"Deal," Sarah breathed. "Thank you."

"One more thing," Tina said. "That intake process Deborah built."

"If someone bypasses it and goes directly to my team, I'm going to send them back to the form."

"For big requests," Sarah said. "We can't send Marcus to fill out a ticket every time he has a quick question."

"Fine. What's a big request?"

Sarah opened her mouth. Closed it.

"Something that takes more than a day," she said.

"Omar spent three hours this morning on George's pipeline. Is that big?"

Sarah didn't answer.

"Everyone says it's urgent," Tina said. "That's the whole problem. Without someone owning the decision, we just run on whoever escalates loudest."

"That's why we need the governance framework finished," Sarah said.

Tina nodded slowly. "That's what Deborah was trying to build. I've been leading the project."

"Then that's your priority," Sarah said. "Thirty days. Keep the systems running and finish the governance work."

It wasn't a full answer. Both of them knew it.

But it was something to point at.

"Thirty days," Tina said.

Sarah spent the afternoon in a state of desperation.

Call 1: Executive recruiter. "Data leader, Series B scale-up? Eight to ten weeks minimum, probably twelve. Budget?"

€100–150K base. The board had already questioned her burn rate.

Call 2: Fractional executive firm. "Three days a week. €12K per month, six-month minimum."

€72K for someone who wasn't even full-time.

She put the phone down. Opened her laptop and searched: data governance consultants Amsterdam.

The first result had a clean website. Framework implementation, fourteen to eighteen weeks. "Investment in a governance project €60–75K range, depending on scope."

She called the intake line. The conversation lasted thirty minutes. She heard herself say: "Can you send a proposal by Friday?"

She thanked them and hung up.

Then she thought about what Deborah had said the morning she arrived eighteen months earlier. They already had parts of the governance. They'd had product for two years. The team was still drowning.

A framework for managing what they had wasn't the problem. The problem was that she didn't even know what they had.

She put the phone down and called Eva.

Eva Rousseau picked up on the first ring.

That was always how she was. Sarah had noticed it the first time they spoke, three years ago, during Series B due diligence. Every other board member made you wait. Eva answered on the first ring and got straight to the point. Sarah had thought: this is someone who actually wants to solve the problem.

"Walk me through where you are," Eva said. "Structurally, not emotionally."

Sarah almost smiled. The men on the board asked how she was doing. Eva asked where she was structurally.

She looked at her notes and started.

The governance model. The decentralized ownership structure Deborah had built. The logic behind it and why it had collapsed. The accountability gap. The Tuesday sync. Her own role in creating the conditions for the failure.

Eva listened without interrupting.

"Good," she said when Sarah finished. "That's honest and structural. The board can work with that on Friday. Don't lead with the incident. Lead with the diagnosis."

"The board wants to know what changes."

"The board wants to know whether the Series C is still on track," Eva said. "The incident is €400K. That's recoverable. What isn't recoverable is walking into due diligence in six months with a data function that can't answer basic investor questions. That's what Friday needs to address."

"Which is why I need the right hire."

"Tell me what you're thinking."

"CDO level, maybe. Someone with a product background thinks of data as a roadmap, not a pipeline. Technical enough to earn the team's respect and senior enough to hold the line with stakeholders. Someone who can finish the governance work Tina started and then build on it."

"The CDO argument," Eva said. There was something careful in her voice.

"You don't like it."

"I've seen it several times. A company at your stage brings in a CDO. Big name, strong resume, genuine capability. Twelve months later, they're gone. Not because they were wrong for the job, but because the job was wrong for the company."

"Head of Data."

"Head of Data with the profile you're describing is the right hire," Eva said. "Technical enough to be credible with the team. Product-minded enough to think about data assets rather than just data infrastructure. Senior enough to push back on stakeholders without being labeled difficult." "Eight to twelve weeks to find. But yes. That's the right profile."

Sarah walked her through the governance thesis, the Tina gap, and the Timeline.

Eva listened and confirmed each piece.

Then: "There's something else. I want to be honest with you; I can't give you a clean argument for it. It's more pattern recognition than data."

Sarah waited.

"I've served on ten boards over the past eight years. Seven of them had data team problems that looked exactly like yours. Two of those companies are still in that shape, and two aren't."

"What changed?"

"The CEO changed," Eva said simply. "Not the team. Not the tools. Not the data leader. The CEO speaks about data in business terms. What the data function costs, what it returns, what it enables. When that happened, the whole function repositioned. It stopped being a technical expense and became a growth asset."

"I can learn to speak about data. That's not a consultant problem."

"You can learn the language," Eva agreed. "But right now you don't know what you don't know. You're about to write a job description based on what you think the data function needs. And I'm not sure that picture is complete yet."

She paused.

"Coming back to Arjun and his work with three CEOs who were exactly where you are, who told me before they met him what they needed. They changed their minds after two weeks of working with him."

"What does he do?"

"He comes from data but thinks like an operations person. He looks at a data function the way I used to look at an IT function at Staples. Not: are these people technically capable? But: is this function financially accountable? Is it producing a measurable return? Is the CEO able to evaluate it the way they evaluate any other business unit?"

"We have George for the financial evaluation."

"George evaluates what he can see. Headcount, licenses, incidents. What Arjun makes visible is what nobody's measuring, including the cost of maintaining things that don't create value, the cost of reactive work versus strategic work, and the cost of a CEO who can't tell investors why their data function is an asset rather than an overhead." "At eBay, I watched technical teams build endlessly. Nobody asked whether what they built was returning value. It took an operations person asking the obvious question, what this actually costs us to run and what we would lose if we turned it off, to change the conversation. That's what Arjun does for data."

"Did it work? With the CEOs?"

Eva was quiet for a moment.

"Two of them, yes. Genuinely. The CEOs started talking about their data function in terms investors understood. Not storage size and dashboard count. Return on data investment. FTE time freed. Capacity shifted from reactive to strategic." "The third one didn't work. The CEO understood the framework intellectually but never actually changed how they operated. Arjun called it after two months. He doesn't stay where it's not working."

"It works if I commit to it."

"It works if you actually change how you work. Not just what you say." A beat. "Which is why I'm not telling you to hire him before the Head of Data. I'm telling you to talk to him before you finalize the job description."

"Or he might not change anything."

"Or he might not," Eva agreed. "One conversation. That's all I'm asking."

Sarah looked at her notes.

Governance. Head of Data with a product background. Finish what Deborah started. Series C in six months.

"Can you do that right now?" Eva asked. "Answer every investor question about your data function without calling a technical lead for backup?"

She couldn't.

"One conversation," Sarah said. "If he has the capacity."

"I'll call him tonight," Eva said. "Now, Friday report. Walk me through the structure."

They worked for another thirty minutes.

When the call ended, Sarah sat with her notes.

But Eva's question was still in the room.

Can you answer every question without calling a technical lead for backup?

She couldn't. She'd never been able to.

She wasn't sure whether it was Arjun who fixed it.

John answered on the second ring. She could hear the city behind him. Brussels, still.

"I read your message," he said. "I'm sorry, Sarah. I know how hard that was."

"I need to think through next steps. Do you have twenty minutes?"

"Yes."

She'd worked with John for three years. He'd come to Stock Dropper from Zalando, not as a founder or as someone building from scratch, but as someone who knew what scale looked like and how to build the infrastructure to survive it. In Zalando's early days, the engineering teams had one job: build the platform that let the business move fast. That was the model. That was what good looked like.

She'd trusted that model without examining it. For three years, she'd trusted it.

"Deborah wasn't the wrong hire," Sarah said first. "I want to be clear about that."

"She wasn't," John agreed. "She was exactly right for where we were eighteen months ago. She came in and stabilized it. That was real work."

"But we're in a different place now."

"That's the right profile," John said about the Head of Data. "The governance piece especially." "Hold on."

A pause.

"I'm pulling up the Looker admin panel," he said. "I've been meaning to look at this for weeks."

Sarah waited.

"Okay." His voice shifted into the register he used when reading data. Focused. Slightly distant. "We have three hundred and twelve dashboards on the server. Three hundred and twelve." A pause. "Connections. I can see the status flags here. Green, yellow, red. I'd estimate about 120, maybe 140, showing yellow or red. That's broken or degraded connections. Pipelines feeding bad data or no data."

"That's almost half."

"That's almost half," he confirmed. "And I'll be honest, I don't even recognize some of these. There are dashboards in here from campaigns that ended two years ago. Attribution models for products we no longer sell." He closed something. "This is a governance failure. Nobody owns these things. Nobody decided to turn them off. They just accumulated."

"Governance closes the loop."

"Governance closes the loop," John said. "We have the tools, including Looker, Snowflake, and dbt, the whole stack. Best in class for a company at our stage. The infrastructure is solid. What we don't have is accountability behind it." "Someone who understands that data is an enabling function. It serves the business. The business moves fast; data needs to move with it. The governance layer makes that possible without everything breaking."

Sarah wrote it down.

Data serves the business. Governance enables speed.

It sounded right. Both of them believe in it.

"Eva mentioned Arjun Mehta," Sarah said. "She thinks I should talk to him before I write the job description."

"I know the name," John said. "Not well. I've seen him post in a few data leadership forums. He has an interesting angle, more operational than technical, more financial than most data people are." A pause. "Honestly, that's not my world. I'm not a data person, Sarah. I'm an infrastructure person. I can tell you whether the stack is right. I can't tell you whether the operating model is right."

"You think the call is worth having."

"I think you should talk to anyone who might help us write a better job description. But I want to be direct. We need someone technical in the seat. Not just a framework. Not just a coach. Someone who can sit with Tina and Omar and actually work the problem. Arjun might be useful alongside that. He doesn't replace it."

"Agreed."

John confirmed he'd send George a note before Friday to share the accountability. The Tuesday sync was on both of them.

"The CDO question," Sarah said.

"Eva's right to push back," John said. "A CDO is a signal hire; it tells investors you're serious. But honestly? Bringing a CDO to a team of three is overkill. You don't put a VP of Engineering over two engineers." A pause. "Head of Data with the right profile, someone who can grow into a CDO role as the team scales, is the smarter move. “Eight to twelve weeks."

"Which means Tina for longer than thirty days. Have that conversation with her early."

She thanked him and ended the call.

She wrote: governance closes the loop, John and Eva both.

She felt, for the first time that week, like she had a thesis she could walk into the board room with.

It was the right answer to the wrong question.

Sarah sat in her office. Everyone else had gone home.

She pulled up Eva's message. Checked Arjun's LinkedIn. 56 mutual connections.

She still didn't want a coach. She wanted a technical leader. But Eva had already been right about two things this week, and Sarah had learned, slowly, to track that pattern.

She started typing.

To: arjun.mehta@gmail.com

Subject: Referral from Dr. Eva Rousseau

Mr. Mehta,

Dr. Eva Rousseau suggested I reach out. I'm the CEO of Stock Dropper.com, a Series B flash-sales e-commerce company with 120 employees in Amsterdam.

We had a €400K data incident last week. I've made a leadership change within the data function. It's the third such change in four years.

For context on where we are technically, we run on GCP, Snowflake, dbt, and Looker.

The team has been running agile sprints for eighteen months. We have a backlog, a governance project in progress, roughly 300 dashboards, and somewhere north of 200 active pipelines.

The stack is solid. The process exists. Something else is broken, and I haven't been able to name it.

I'm opening a search for a new data leader at the CDO level, product-minded and technically hands-on. The thinking is that we need someone senior enough to bring real strategy to the function and drive clearer priorities. Right now, data exists to enable the business, and it keeps failing to do so.

Eva said you help CEOs understand how they talk with the data team. I'm not sure that's what I need. But I'd appreciate a conversation before I finalize the job description.

Available Monday or Tuesday.

Sarah Chen

CEO, Stock Dropper.com

She read it back. Direct. Maybe too direct about the uncertainty. She left it.

Hit Send.

The response came back two hours later.

Sarah,

Thanks for the context.

One thing I noticed: you described your data function as existing to enable the business. That's a reasonable position. It's also the position every company I've worked with held until something broke badly enough to question it.

You had three data leaders in four years and a serious technical investment that isn't delivering. That combination is specific. It's not a people problem. It's probably not a tool’s problem either.

I'm not sure my work is what you need. What I do is help CEOs see the operating costs of what they've built, what it's actually returning, and whether what they're about to hire solves the right problem.

Whether that's relevant, I genuinely don't know yet.

I work in a specific way, and I set the terms of any engagement. That's not arrogance. The work only holds if the conditions are right.

Monday, 10 AM CET?

Arjun

Sarah read it twice.

He hadn't offered to fix anything. Hadn't listed credentials. Hadn't asked about budget or timeline.

He'd read her email and reflected something back that she hadn't said out loud.

The position every company held before something broke badly enough that they had to question it.

She forwarded it to Eva with one line: He's either very good or very annoying. Possibly both.

Eva's reply came before Sarah had finished her coffee.

Both, usually. That's why it works.

Monday is confirmed; He moved something to make it happen. Go in with your real questions. He'll know the difference.

One more thing: read the part about enabling the business again. That's where he starts.

Sarah read the reply again.

She still wasn't sure she needed what he was selling.

She wasn't sure he was selling anything yet.

But she'd have the call.

Deborah stopped at Sarah's doorway. Her jacket was on.

"I need you to hear this one thing, Sarah."

"This is not my fault, and it's not yours or John's either; you didn't care enough to stop it. You handed the entire strategic function to the CTO, who sees only technology and code. He saw us as an infrastructure problem, not a business solution."

Deborah stepped closer.

"We have Agile. We have planning meetings, a backlog, and a Jira board full of tickets. We execute sprints. But two days into every sprint, five different heads and C-levels are yelling about five different emergencies, and something new always takes precedence. We are firefighters, Sarah, not a data team."

She took a sharp breath.

"And when we tried to fix it? When we tried to enforce boundaries and protect our capacity? You intervened. You told us to drop what we were doing because that stakeholder was a priority or that deadline was critical. You forced us back into the old pattern every time, because it was easier than having a difficult conversation."

She turned and was gone.

The corridor was silent.

Sarah sat with the echo of it.

We are firefighters, Sarah, not a data team.

She didn't know what a data team was supposed to be instead. But the question was different now.

Deborah called in sick for the rest of the separation period.

Sarah couldn't sleep. At 3:47 AM. She opened her laptop and searched: " How to manage data teams.

Fifteen articles. Governance frameworks. Boxes and arrows. She almost opened a tab to email a consultancy that sold the diagram as a product. Then stopped.

Governance would give her 300 dashboard owners and policies. It wouldn't answer the question she couldn't ask yet. Why were they still running? Who decided they should keep running?

She pulled up the last eighteen months of approvals instead. Every initiative she'd ever signed off on.

Not one of them replaced anything.

Every single one was an addition. New capability on top of everything already running. Nothing ever traded off. Just more, and more, and more.

She opened her notebook.

Questions I can't answer:

How do we decide what the team works on when everything is urgent?

How do we know if the team is overloaded before things break?

How do we say no to stakeholders without it feeling arbitrary?

What do we do with the things we've built that nobody uses?

She stared at the last one.

These weren't technical questions.

She thought about the investors. Eva. The board. The €8–10M that was supposed to close in six months. She thought about what it would mean to reach due diligence and have someone ask her, calmly, across the table: "Walk us through your data function." She thought about what she would say.

She had no idea what she would say.

That fear she couldn't share with anyone. Not John. Not George. Not the team. You couldn't hand your investors' money to the people who were supposed to help you protect it and say: I don't know if I can do this.

So, she sat with it at 3:47 AM and wrote questions in a notebook.

Tomorrow, maybe, she'd find someone who could help her answer them.

The Zoom link opened at exactly 10:00 AM.

Arjun appeared on her screen. Home office, a bookshelf behind him. Oddly enough, the book "Kitchen Confidential" by Anthony Bourdain was clearly visible to the camera, and a child's drawing was pinned to the wall. Not a consultant's staged backdrop. Just a room where someone worked.

He looked younger than she expected. And less certain. That surprised her.

"Sarah. Thanks for making time."

"Thank you for the email," she said. "It was direct."

"I find it's faster." He opened a notebook. "Can I ask you a few questions before we get into anything else? I want to understand the problem before I decide whether I have anything useful to add."

"Go ahead."

"Tell me how the data team operates. From your perspective, what you see, what gets reported to you, what lands on your desk."

From your perspective. She noticed that.

She told him what she knew. The weekly status updates from Deborah, then Tina. The incidents that arrived as messages from Marcus or Carlos before the data team knew about them. The board questions she couldn't answer without calling someone. The sprint reviews she'd glanced at once but never attended. She believed in ownership. Deborah owned the function; That was the model.

"When something broke," Arjun said, "how did you find out?"

"Usually from a stakeholder. Operations, product, and finance. They'd come to me because they couldn't get traction with the team."

"And when you asked the team about it?"

"They had context I didn't. A dependency issue, a pipeline already in the backlog, a migration behind schedule. There was always a reason." "I stopped asking detailed questions because I didn't have the technical background to evaluate the answers."

He wrote something down.

"The three data leaders. What did each of them tell you was the main problem when they left?"

Sarah thought about it. "The first one said the team was too junior and the stakeholders too demanding. The second said the infrastructure wasn't ready for what we were asking. Deborah said the problem was structural. That individual ownership without accountability infrastructure couldn't hold."

"And you agreed with her."

"I agree with all three of them," Sarah said. "That's the problem."

Arjun looked up from his notebook. Almost smiled.

"Tell me about the stack," he said.

She told him. GCP, Snowflake, dbt, Looker. The agile implementation eighteen months ago. The governance project. The estimated 300 dashboards, with over 200 active pipelines.

He didn't react to the tools the way technical people usually did, no nods of recognition, no questions about architecture decisions. He listened like he was counting something she couldn't see.

"How many of those dashboards does the business actually use?" he asked.

"I don't have that number exactly. Roughly a third, maybe. Maybe less. John, our CTO, pulled up the Looker admin panel last week and estimated that about a hundred and twenty were showing degraded or broken connections."

"So close to half the infrastructure is running, being maintained, alerting when something breaks, for users who may not exist anymore."

"That's what governance solves," Sarah said. "Clear ownership. Clear maintenance responsibilities. That's exactly why the next hire is focused on finishing the governance work."

"I'm not arguing with that," Arjun said. And he meant it. She could hear that he meant it. "Governance is right. The technical hire is right. The product mindset is right. Agile is right for what it does." "Can I try something?"

"Go ahead."

"Do you know the survival rate for restaurants in the first year?"

"Low. Far lower than most people can handle."

"About a third don't survive year one," he said. "And within five years, nearly sixty percent are gone. The ones that do make it through almost always have two things: a strong chef who knew exactly what they were building from day one and a menu that was planned, not accumulated."

"Okay."

"The restaurants that drop dead in month two or three always follow the same script: they panic. The door count dips, so the owner gets schizophrenic. They start cluttering the menu. A new Tuesday night special to lure in the locals. A cheap, low-rent pasta dish to compete on price. A vegan stir-fry because one loud-mouthed customer complained to the hostess. They run desperate, cut-rate promotions. They try to cheat the margins by buying in bulk, throwing everything in the walk-in freezer, and praying the microwave can resurrect it to order without the customers noticing. They’re moving at a frantic, breathless clip, trying everything, hoarding forty different dishes on a single laminated page because they’re terrified of saying no to anyone."

By month four, the chickens come home to roost. You’ve got a miserable, exhausted kitchen staff trying to execute thirty-eight disparate dishes on a four-burner line. The prep cooks are burned out. The signature dish that drew people through the door in week one goes completely sideways because the kitchen's focus has been entirely cannibalized by the chaos. Yelp reviews drop, the word gets out, and quality plummets to match the panic.

Sarah was listening.

"Data teams do the same thing," Arjun said. "Every new business request is a new dish. We can't say no. The business needs it, the stakeholder is urgent, and the request is reasonable. So it goes on the menu. And nothing ever comes off. Every dish was added for a reason. Every pipeline has a stakeholder. Every dashboard was built because someone needed it once."

"So, we need fewer dashboards."

"You need a system for deciding which dishes stay on the menu," he said. "Not just a system for making the kitchen run better. Governance makes the kitchen run better. That's real, and it matters. But it doesn't answer the question of whether you should still be cooking thirty-eight dishes."

Sarah looked at her notebook. Sitting open on the desk from Saturday night.

What do we do with the things we've already built that nobody uses?

She didn't say it out loud.

"The other piece," Arjun continued, "is the frozen food problem. When teams get overwhelmed, they start collecting everything and making it available, including all the data, tables, and raw events. Just in case. A reasonable instinct. But now people build dashboards on top of raw data that nobody has validated. They discover bad data. They lose trust in the function entirely. And the team spends more time defending the data than using it."

"Which is what happened last week," Sarah said quietly.

"Usually something like that triggers the governance conversation," Arjun said. "Which is the right response. I'm not here to tell you governance is wrong. I'm here to tell you there's a layer underneath it that governance alone doesn't reach."

"What layer?"

"The question is what the data function is accountable for," he said. "Right now, the data team exists to enable the business. They build what's requested, maintain what's been built, and respond when things break. The business decides what matters, and the data team executes."

"The enabler model puts accountability in the wrong place," Arjun said. "The business requests. The data team delivers. But the business lacks the context to judge whether what it's asking for is worth building. So when something breaks, it suffers, without ever having owned the decision that caused it."

She looked directly at him, confused, or was it frustration?

"The shift I work on is moving from data as an enabler to data as an impact. The business co-owns. It knows what a product actually costs to keep alive, not just the infrastructure bill, but every incident, every workaround, and every sprint consumed maintaining something nobody is improving. It knows what has to stop to make room. And it can judge whether what gets built is returning value."

"How do you measure that?"

"Three numbers. What it costs the data team to keep the portfolio running, the real number, not the cloud invoice. What it costs the business to work around the portfolio when it doesn't fit, the hours your people spend doing manually what the data should be doing for them. And whether the team has any capacity left to fix either problem, or whether they're fully consumed keeping the current state alive."

"So this lives in Looker? That's what we use."

"Not with Looker specifically. Not with any tool. The framework operates at the operating model level. It doesn't matter whether you're on GCP or AWS, or using Tableau or Looker. The system is about how decisions are made, not what runs underneath them."

Arjun closed the laptop.

Rona drawing was still on the wall behind him. He'd forgotten to take it down before the call.

He opened his notebook. Read back what he'd written during the call.

He also knew he'd described the kitchen and hoped she'd see the dishes. He'd given her the philosophy but hadn't shown her the week-by-week details of what it actually looked like in practice. The honest answer was that he was still building that part. The capacity framework was solid. The portfolio management logic was on the right track. The rest was still being field-tested.

He put the notebook down and opened LinkedIn.

Three new job postings in the last twenty-four hours. VP of Analytics, Amsterdam. Head of Data, Berlin. Senior Data Product Owner at a logistics company in Stuttgart.

He read the Amsterdam one twice.

Strong technical background required. Python, Spark, experience building data platforms at scale.

Not him. Never him. He'd built on platforms. He'd never been the one building them.

He closed the tab.

If he took a full-time role, he'd be back inside the problem, building the thirty-eight-dish menu one dish at a time, watching the kitchen slow down, knowing exactly why but unable to say it out loud because he was an employee now, not a practitioner.

He'd done that before. Three times.

He picked up his phone and sent Eva a message.

The call happened. She's thinking. I'm not sure I made the case well enough. Can we talk tomorrow?

Eva's reply came back in four minutes.

You never make the case well enough. That's not where you close. I'll call you at 9.

He read it twice. Put the phone face down on the desk.

Opened the Stock Dropper notes document and kept writing.

"Tell me how it went," Eva said.

"You described the problem well," she said when he finished. "You always describe the problem well. That's not where you lose people."

"Where do I lose them?"

"When you stop describing the problem and start describing what you do about it. It gets abstract. You say 'operating model' and 'portfolio management' and 'financial accountability' and the CEO hears consulting language. They stop seeing their own company and start wondering if you're selling something."

"I'm not selling something."

"I know that," Eva said. "You know that. They don't know that yet." A pause. "The kitchen metaphor is good. The frozen food is good. The thirty-eight dishes, that lands. Stay in the concrete longer than feels comfortable. You move to the framework too fast."

Arjun wrote that down.

"She went quiet at the end," he said. "I don't know if that's good or not."

"It's good," Eva said. "Sarah doesn't go quiet when she's dismissing something. She goes quiet when she's working something out."

"She had her counter-arguments ready. I could see it. She just didn't use them."

Eva was quiet for a moment. "I've known Sarah for three years. She writes everything down when she doesn't have an answer yet. She's not dismissing you, Arjun. She's stress-testing."

"She'll talk to John her CTO. And probably George the CFO."

"George will appreciate the way the financials are presented," Eva said. "That's not the real issue. The challenge is the budget and the argument that you're not the technical expert. She needs both, and she knows the board will question why she's paying for two services when she can hardly justify one.

"That's a fair argument."

"It is," Eva agreed. "It's also the wrong frame." "Can I tell you what I saw in the other portfolio companies? Not the abstract version. The specific thing."

"Yes."

"In every case where it worked, it started with the CEO changing how they talked about data in board meetings. Not the technical lead. Not the governance framework. The CEO. They stopped saying 'our data team is working on it' and started saying 'our data function costs us X, returns Y, and here's what we're retiring to make room for Z.' Investors noticed, not because it was impressive, but because it was unusual." "Sarah can't do that right now. Arjun, you can teach her to do that. That's the Series C argument. The CEO's ability to speak about data as a capital allocation decision."

Arjun sat with that.

"I didn't say it that way to her," he said.

No. You mentioned 'impact and operation' and 'portfolio management.' A pause. Next time, say: I will teach you to discuss your data function as confidently as you talk about any other business investment, to your board, your investors, and yourself.

"That's cleaner."

"That's what you do," Eva said simply. "The frameworks are how you do it. That's not the pitch."

Arjun was quiet for a moment.

"I looked at the LinkedIn job board this morning," he said.

Eva didn't answer immediately.

"And?"

"VP Analytics in Amsterdam. Data strategy in Berlin."

"Were they right for you?"

"No."

"Then why are you telling me?"

He pondered the situation. "I want you to understand that I'm still questioning this. I'm uncertain if I've created something genuine yet. I don't know if what I have is transferable, if Sarah is the right test case, or if I'll walk in and discover the framework collapses under real-world conditions."

"All of that might be true," Eva said.

"That's not reassuring."

"I'm not here to reassure you," she said. "I'm here to tell you what I saw. I watched three companies change. I watched the before and the after. The before looks exactly like Stock Dropper. The after looks like something I didn't know was possible until I saw it." A pause. "Whether you can reproduce it consistently, yes, that's still the question. But that's the question you need to be inside a company to answer. Not on the LinkedIn job board."

Arjun closed the tab he'd reopened without noticing.

"If she says yes," he said, "I'll need your help with the board conversation."

"I will be there," Eva said. "I've already started."

John had twenty minutes between calls.

She gave him the version she'd written in her notebook. The kitchen metaphor. The thirty-eight dishes. The missing system for deciding what to stop.

John was quiet for a moment.

"That's not a data argument," he said. "That's an operations argument."

"I know."

"At Zalando, we had the same problem in engineering. New features kept getting added. The platform got slower. We kept hiring to compensate. It took two years before someone asked: what if we deprecated half the feature set instead of scaling the infrastructure? Nobody wanted to have that conversation because everything had a stakeholder."

"So, you've seen this."

"In engineering. Not in data." A pause. "The difference is that engineering has technical debt as a concept. Everyone understands that old code costs you. Data doesn't have that language yet. What he's describing is the cost of maintaining what you've built versus what it returns; that's technical debt for data. That's the missing concept."

"He doesn't call it that."

"He should." A pause. "Sarah, I can't validate what he does technically. That's outside my lane. But the operating logic makes sense to me. If he can quantify what the portfolio costs to run, you have a completely different conversation with the board about headcount and investment."

"You support the conversation."

"I support you having more information before you finalize the hire," John said. "Which you've already decided to do."

She hadn't said that yet.

She didn't correct him.

She didn't use the kitchen metaphor. She used the numbers.

There are 300 dashboards, of which around 140 are degraded or broken. They incur maintenance, alerting, and infrastructure costs, even for users who may no longer exist. Currently, there is no system to identify which dashboards should be retired, nor a financial model to evaluate the cost of maintaining the portfolio against its benefits.George was quiet for longer than she expected.

"Isn't this a governance problem? What does he measure?" George said, finally.

"FTE cost of reactive work versus strategic work. Cost of maintaining what's been built versus what it returns. Whether decisions made with data are better than decisions made without it."

"Financially modeled?"

"That's the claim."

Another pause.

George said, "Every quarter, I ask John about the cost of the data function this year." He usually mentions headcount and licenses, but I've never been able to ask if those expenses actually yielded any outcomes. "I've never had a model to ask that."

"Arjun builds that model."

"Does he." It wasn't a question; it was a surprise. He was thinking. "I'd like to see the formula. What is his cost?"

"That's part of what I need to work out."

"Work it out," George said. "And Sarah, frame it as a capital allocation question when you bring it to the board. Not a coaching engagement. A capital allocation decision for the data function." A pause. "That's a conversation I can support."

She hadn't expected that.

She thanked him and hung up before he could add conditions.

The apartment was quiet.

She sat at the kitchen table with her notebook open. The same notebook from the call. The same page she'd been looking at when Arjun described the thirty-eight-dish kitchen.

She read back through her notes.

Three leaders, three correct diagnoses. All of them walked into the same structure.

The business requests. The data team delivers. Nobody is accountable for whether it returns anything.

The system for deciding what to stop, it doesn't exist.

Technical hire is right. Governance is right. Agile is right. Is this the missing link?

She turned to the page from Saturday. The 3 AM questions.

She read them slowly.

He hadn't answered them. He'd described the conditions that produced them. And she'd sat through the call with three pages of counterarguments, covering budget, timeline, technical gap, and Series C pressure, and looked down at this page without using any of them.

Not because he'd convinced her.

Because she couldn't answer the question herself.

Not about her own company. Not after three years.

Show me one dashboard that helped you decide to stop something.

She couldn't.

That was the missing component. Not a person. Not a framework. The absence of a system to make the stop decision. Everything at Stock Dropper was built to accelerate. Nothing was built to evaluate.

She wasn't certain about Arjun. She wasn't certain his approach would work, that the frameworks would hold under real conditions, that the board would support the cost of two hires when the budget was already tight.

She was certain that hiring a fourth technical leader into the same structure was not the answer.

She wrote one line at the bottom of the page.

Against best judgment. Try.

Then she closed the notebook and went to bed.

The call had eight people on it. She'd asked for thirty minutes.

She took twenty-two.

She led with the structural diagnosis, not the incident, not the €400K, not Deborah. The accountability gap. The portfolio problem John had helped her name. The cost of what they'd built versus what it was returning. The absence of a stop system.

She used George's framing. Capital allocation decision for the data function.

Christoph asked the first hard question. "We're approving two hires? In this environment?"

"A technical lead and an operating model engagement," Sarah said. "The technical lead builds the function, and the engagement teaches the CEO and the board to evaluate it as a business investment rather than an infrastructure cost."

"What does the engagement cost?"

She gave the number.

Silence.

"That's not cheap," Christoph said.

"It's cheaper than a fourth failed data leader," George said quietly. It was the first thing he'd said.

Another silence.

"I've seen this approach work," Eva said. "In two other portfolio companies at this stage. I wouldn't recommend it if I hadn't."

"The Series C argument," Christoph said. "Walk me through it."

Sarah guided him through it, focusing not on dashboards and KPIs or storage scale and event tracking.

Christoph was quiet for a moment. Then: "Fine. Thirty-day review."

Eva caught Sarah's eye on the screen.

Not a smile. Just a small nod.

Approved.

She sent the email to Arjun Thursday night.

Board approved. Ready to discuss terms when you are.

His reply came on Friday morning.

Sarah,

Good. Before we schedule, here are my working conditions. These aren't negotiable.

Direct access to you. Weekly. Not a project manager, not a delegate. You.

Presence in the rooms where real decisions happen. Not the communications meetings. The actual decision meetings.

A ninety-day engagement minimum. The first thirty days are assessment only; I won't recommend structural changes until I understand what's actually happening. If you act on impulse before we have the picture, I will stop the engagement.

Full transparency from the team. They need to know why I'm there and what I'm measuring.

We can disagree. When a decision is made, we commit and move on.

I'll share a full scope document next week. If the conditions work for you, we can start the week of the fourteenth.

One more thing: I work alongside your technical hire, not in place of them. When you open that search, I'd like to be part of the profile conversation.

Arjun

Her first reaction: diva.

Her second: he's describing exactly the conditions Deborah never had.

She typed one line back.

The fourteenth works. Send the scope document.

She hit send before she could think about it again.