When you walk into a company with negative EBITDA and a runway measured in weeks, the instinct is to build a plan. The right move is to build a map first.

"In a distressed situation, the most dangerous thing is false clarity. A bad plan executed confidently is worse than no plan at all."

Hour 0–4: Cash first

The single most important question is "how many days of cash do we have." Not the number from the briefing document — the real number, with actual bank balance, committed outflows for the next 30 days, and receivables that are genuinely collectible.

• Current bank balance across all accounts
• Payroll obligations hitting in the next 14 days
• Supplier payments due and negotiable
• Receivables genuinely collectable in 30 days

Hour 4–12: Walk the floor, not the org chart

In every distressed company there are three or four people who know exactly what is broken and have been trying to fix it for months without authority. Finding them in the first half-day is worth more than any consultant's diagnostic.

Hour 12–24: Map the P&L by unit

The restructuring of Gorillas France only became actionable when we broke the P&L down by warehouse — 25 individual units. Some were profitable. Some were burning €80k per month. The aggregate had been masking the spread for over a year.

"The aggregate P&L is a political document. The unit-level P&L is an operational one."

Hour 24–48: Three priorities, not ten

• A cash preservation action
• A revenue protection action — a customer or contract at risk
• An organisation action — a decision bottleneck slowing everything else

AI can now handle forecasting, anomaly detection, and reporting with speed no human team can match. The conclusion that you need fewer senior operators is dangerously wrong.

"AI makes good operations faster. It does not make bad operations good."

What AI has genuinely changed

• Real-time P&L visibility at unit level
• Demand forecasting and inventory optimisation
• Customer churn prediction and early warning signals
• Automated first-draft investor reporting

What it cannot replace

AI cannot walk into a room where the team has lost confidence and make a decision people will follow. It cannot negotiate with a supplier threatening to cut credit terms while managing the founder's ego and the investor's panic simultaneously.

The trust problem

Operational transformation requires people to trust the person making the decisions. No AI system has accountability. It has no career, no reputation, no skin in the game.

"More data with less experience is just faster confusion."

The months after a Series B close are the most dangerous. More cash, more expectations, same operational infrastructure.

"The fundraise is the end of one story and the beginning of a much harder one."

The three structural failures

1. Hiring ahead of process

A sales team of 15 without a repeatable process is not 15 times more effective than 3. It is 15 times more expensive and approximately as effective.

2. KPIs that lag instead of lead

Pipeline velocity, activation rates, unit-level margins — the metrics that actually predict performance are rarely tracked with the same rigour as revenue.

3. The founder still in execution mode

At Series B, the founder needs to transition from doing to deciding. The result: a founder blocking every decision, unable to see the strategic picture.

How to fix it in 30 days

• Day 1–5: Build the leading indicator dashboard
• Day 5–10: Audit the hiring plan against process infrastructure
• Day 10–20: Decision meetings only — not reporting meetings
• Day 20–30: Transfer decisions to the management team explicitly

The patterns of failure in SaaS operations are almost identical to what I saw running one of Europe's largest last-mile delivery networks — 800 people, 25 warehouses, a burn rate that could kill the company in 90 days.

"In physical ops, a broken process surfaces in hours. In SaaS, it surfaces in your next board meeting."

The three failures that cross every sector

1. Processes designed for the company you were

The company scaled headcount without scaling processes. The onboarding flow that worked for 10 customers does not work for 100.

2. Metrics that measure activity instead of output

At Gorillas, what mattered was contribution margin per order per warehouse — not orders processed. The same logic applies everywhere.

3. Decision-making that doesn't scale

At 80 people, founders in every decision is the single biggest operational bottleneck. Build the infrastructure that allows decisions to be made at the right level.

"The aggregate always lies. The unit always tells the truth."

Every board deck has a unit economics slide. The numbers look defensible. Six months later, nobody can explain why the model isn't working. The unit economics were not wrong. They were hiding.

"Aggregate unit economics are a board-level document. Unit economics by segment and cohort are an operational document."

What aggregation hides

At Gorillas, one view across 25 warehouses showed improving contribution margins. When we broke it down: eight near breakeven, seventeen deeply negative. The average was masking the structure entirely.

Where it hides in SaaS

CAC in aggregate

A blended CAC of €8,000 with LTV of €40,000 looks like a healthy 5x. Break it by channel: inbound at €3k, outbound at €22k — and outbound is growing. The aggregate is fine. The trend is catastrophic.

NRR masking cohort deterioration

Strong early cohorts and deteriorating recent ones can produce excellent NRR for 18 to 24 months before the dynamics show up in the headline number.

The five questions for every financial review

• What does this metric look like by segment, channel, or cohort?
• Is the mix shifting in a way that affects the aggregate in 6–12 months?
• What would this look like without our top 10% of customers?
• Where is the variance highest, and what is driving it?

Get the COO hire wrong and you have a senior executive misaligned with stage, draining capital, slowing decisions, and eroding team confidence.

"The most common mistake is hiring for the company you want to be in two years, not the company you are today."

Why most COO hires fail

Hiring seniority instead of fit

A COO from Series D brings instincts calibrated for a different organisation. They create reporting structures for 300 people in a company of 60. They move at a speed the company cannot afford.

Hiring to solve the founder's discomfort

If the processes, metrics, and decision-making frameworks don't exist, hiring a COO doesn't create them. It creates a senior person trying to build without authority.

"A COO without a mandate is an expensive shadow."

The alternative most founders overlook

Between "founder manages everything" and "hire a full-time COO" there is a third option: an embedded operator who builds the infrastructure and transfers ownership — without the permanence and cost of a C-level hire.