The 0 → 1 diagnostic: what I look at in the first two weeks

Every engagement starts with a brief, and the brief is almost always wrong. Not dishonest, just wrong in the way a map drawn from inside the building is wrong. The founder says the problem is engineering velocity. Two weeks later we agree the problem is that nobody has said no to a customer in eighteen months.

So I don't start with the roadmap. I run the same diagnostic every time, and I protect the first two weeks for it even when everyone wants to jump straight to fixing things.

Week one: evidence, not opinions

I read the last hundred support tickets before I read the strategy doc. Tickets don't perform for an audience. If the same confusion shows up forty times, that's the product telling you where it hurts.

I use the product like a paying customer, from signup to the moment of value, on my own laptop, without anyone guiding me. It's uncomfortable how often the team hasn't done this in months. Then I sit in on two or three sales or onboarding calls and listen for the gap between what's promised and what I just experienced.

On the technology side I look at the deploy pipeline before I look at the code. How long from merge to production? Who can press the button? A team that ships daily with a rough codebase will beat a team with beautiful architecture that releases monthly. The code review can wait; the release cadence can't.

Week two: decisions and money

I trace the last ten meaningful product decisions. Not what was decided, but how: who was in the room, what evidence was on the table, how long it took. This is the fastest read on a company I know. Slow, crowded, evidence-free decisions upstream explain most of what looks like an execution problem downstream.

Then I follow the money and the calendar. Where does the payroll actually go, and does that match what the strategy claims matters? If the deck says AI is the future and one engineer works on it part time, the deck is fiction.

The memo

The output is a one-page memo, not a fifty-slide readout. It names the real constraint, proposes three bets with owners and dates, and lists at least one thing to kill. The kill list matters. A diagnosis that only adds work isn't a diagnosis, it's a wish list.

Two weeks feels slow when the house is on fire. But I've never regretted the diagnostic, and I've often regretted skipping it.