The first 90 days of any new venture set the trajectory. Get them right and the next two years are downstream of good decisions; get them wrong and you’re paying compounding fees on a wobbly foundation. Across five companies, the same checklist runs every time.

Days 1–14: thesis

Before legal, before hiring, before product:

  • Write a one-page thesis: who is the customer, what is the painful job, what is your unique angle
  • Talk to 20 customers. Not friends. Real prospects. Listen for words you didn’t expect.
  • Write down the kill criteria. “If by day 90 we haven’t [X], we shut this down.”

Days 15–30: legal and structural

  • Incorporate. Pick the right jurisdiction for your tax and tax-residency case
  • Open the bank account. Capitalise the operating account, not the holding entity
  • Sign founders’ agreement (if there are co-founders)
  • Lock the trademark and the domain

Days 31–60: build the smallest possible offering

  • Write the smallest version of the product that solves the painful job for one customer
  • Charge that customer money. Real money. Not a friend price.
  • Get a written testimonial
  • Iterate on what you learned in week 8

Days 61–80: get to repeat

  • Sell the second customer. Slightly different than the first; close enough to learn the pattern
  • Sell the third. Now you’re seeing what’s repeatable
  • Document the sales process. The version you ran for #1, #2, #3

Days 81–90: go/no-go

Honest review against day-zero kill criteria:

  • Did three customers pay real money?
  • Did one of them refer the next?
  • Is the unit economics defensible at scale?

Yes to all three: keep going, hire your first operator, prepare to fundraise or self-fund the next stage. No to any: shut down or pivot. Most founders’ first instinct is to extend; the better instinct is to respect the kill criteria you set when you were thinking clearly.

What I always wish I’d done sooner

Hire the operator on day 75 instead of day 200. The right first hire compounds for years; delaying that hire delays everything downstream.

What I always wish I’d skipped

Branding work in the first 90 days. The brand reveals itself through customer conversations; pre-emptive brand work always gets thrown out. Pick a workmanlike name, ship, refine the brand once you actually know what you’re selling.

The principle

The first 90 days are about discovery, not building. Most failed ventures spent those 90 days building. The successful ones spent them learning what to build. Talk to people, sell something small, prove repeatability, then scale.

In your first 90? I’ll review your thesis page.