The crisis reset

It breaks in waves. The instinct is to cut a little, hope, and cut again. Each wave restarts the fear, and by the third one nobody's working; everyone's interviewing. The survivors you were protecting with the small cut are the ones the small cut destroys.

The documented standard for doing it once is Chesky's May 5, 2020 letter, published while Airbnb was losing 80% of revenue in eight weeks. The structure, in order: principles first (every reduction maps to future strategy, do as much as possible for the impacted, communicate 1:1, no partial transparency, nothing said until all details land, because half-news makes everything worse). Then the decision: 1,900 of 7,500, once. Then the mechanics: 14 weeks of severance plus a week per year, 12 months of healthcare, the one-year equity cliff dropped so everyone left an owner, an alumni directory and a placement team worked like a real project. The leaving cohort became alumni instead of enemies. The letter circulated as a model instead of a scandal. The IPO came seven months later, on the trust that survived. You don't need Airbnb's balance sheet to copy that structure; the structure is free.

The second break is scope. A reset that only cuts people while keeping every ambition is arithmetic, not strategy. Chesky shelved transportation, Studios, and most of the hotel and luxury ambitions in the same letter, and refocused the survivor company on individual hosts. The reset kills projects and dreams too, or it's a payroll adjustment with a eulogy.

1. Declare wartime, out loud, to yourself first. Name the single objective: survive to a specific state by a specific date. "Default alive by Q3 next year" is a wartime objective. Every decision for the duration gets tested against it, and peacetime projects, peacetime process, and peacetime niceness are suspended, knowingly, with an end condition attached.

2. Size the cut to the worst credible scenario. Not the base case, and never the recovery you hope for. The worst credible one: how long does zero-to-terrible revenue plausibly last, and what cost base survives it? Cut once, to that depth. Cutting to the hopeful scenario schedules the second wave, and the second wave costs more trust than the first cut cost jobs. If the worst case proves too pessimistic, you rehire into strength, which is a good problem. The reverse never happens clean.

3. Write the principles before the decisions. On paper, before any name is on any list: what maps cuts to future strategy, what you'll do for the impacted, how people hear it (1:1, never by badge deactivation), and the rule that nothing gets announced until every detail lands. Publishing the principles makes them binding, and they double as the audit: a cut you can't map to strategy is a cut you don't understand yet.

4. Kill ambitions, not just costs. List every project and bet. Each faces the wartime question: does this serve survival and the refocused core? Shelve the rest, including, especially, the ones with your identity attached. That's the difference between a smaller company and a focused one.

5. Spend real money on the leaving cohort. Severance past comfortable, healthcare bridged, cliffs dropped, placement worked like a pipeline. This is the money currency buying back the people currency while everyone watches. Survivors calibrate their trust by how the departed were treated, and the departed carry your name forever, into every company they join and every founder they talk to.

6. Execute in 24 hours once decided. Decision to done inside a day. Every day between decision and execution leaks, and leaks turn a hard day into a cruel month. Conversations 1:1, details complete, support live, letter published the same day.

7. Run the personal currency check. You're running the reset while absorbing your own fear: your identity is shrinking with the company, and your health is paying the wartime pace. Name it, to yourself and one person outside the company. Never leak it downward; the team reads your face as forecast. And the 3am state decides nothing here either, the same as everywhere else in this system.

• Salami cuts. Three small waves cost more trust than one deep one.
• Sizing to hope. Hope is a scenario, never a cost base.
• Keeping every ambition while cutting the people who would have built them.
• Partial transparency. Rumor fills whatever you leave unlanded.
• Cheap exits to save two weeks of runway. Reputation sold at its high.
• Letting wartime become permanent. Declare the end when the state is reached, or the emergency becomes the culture.

Founder, 40-person B2B travel SaaS. Revenue down 70% in six weeks, 11 months of runway. Wartime declared: default alive by Q3 next year. Worst credible scenario: depressed revenue for 18 months, which prices a cut of 14 people and both moonshots, deeper than the 8 his gut proposed, and the gap between 8 and 14 is exactly the second wave he didn't schedule. Principles written and published first: cuts map to the refocused core (the SMB segment that's still paying), do the most possible for the 14, every conversation 1:1, nothing said until packages are final. Ambition kill: enterprise expansion and the API marketplace, shelved in writing, with a note on why the survivor company is an SMB company. Mechanics: 10 weeks of severance plus a week per year, healthcare through year-end, cliffs waived, warm intros worked like a pipeline. Execution: decision Friday, conversations Tuesday, all 14 done by 2pm, letter at 4, all-hands at 5. Personal check: he tells his wife and one founder friend that he's scared. Nobody else.

Eight months later, burn is covered by the SMB core, three of the 14 work at customer companies, and survivor engagement runs higher than pre-crisis. That's what cutting once, generously, to the honest depth, buys.