For fifteen years, every decision I made had a number attached to it. CFO at Typeform, where I could tell you the unit economics of every growth lever. CFO at TravelPerk, where I built scenario plans that let the leadership team adapt in real time when COVID wiped our revenue to zero. I thought in spreadsheets. I modelled risk. I made decisions defensible. That was my job, and I loved it.
Then I became CEO. And I discovered that the decisions that keep me up at night are the ones no model can answer.
When we decided to build readywhen.ai, every financial signal said don't. The burn rate was uncomfortable. The TAM was unproven. We had decent ARR at Oliva Health in a business that worked. I could build you a beautiful model showing exactly why walking away from proven revenue to build an unproven AI product was a terrible allocation of capital. I've built that model for other people's companies. I know what it looks like. I know what the recommendation would be.
We did it anyway.
Not because we ignored the numbers. Because the numbers couldn't answer the question that actually mattered. The team and I want to create something meaningful. Something at scale. Something we're proud of. Coaching worked, but it doesn't scale. Mental health benefits worked, but the market commoditised beneath us. An AI-native product that helps managers act on the commitments that matter... that could reach millions. That could change how work actually works.
You can't model "meaningful" in a DCF. You can't scenario-plan "proud of."
Here's what I didn't understand when I was a CFO, and it took becoming a CEO to see it. No matter how strategic I was, no matter how close to the business, I was always a partner to the CEO. The scenario plans, the burn multiples, the runway models... all of them were tools to inform the decision. But the mission wasn't mine. The vision wasn't mine. I could tell you whether we could afford to do something. I could never tell you whether we should exist to do it.
Those are different questions. And the second one doesn't have a cell in the spreadsheet.
A CFO asks: what does the data say? A CEO asks: what do we believe? The CFO's job is to make the decision defensible. The CEO's job is to make the decision, full stop, including the ones that can't be defended with data. Including the ones where the conviction has to come from somewhere the spreadsheet can't reach.
I still build the models. I still run the scenarios. I still hold the burn multiple in my head every single day. The financial discipline is muscle memory and I'm grateful for it. But I've stopped expecting the numbers to tell me whether we're building the right thing.
The hardest adjustment wasn't learning new skills. It was unlearning the instinct that every decision should be backed by data before it's made. Some decisions can only be backed by data after they've played out. The decision to build readywhen.ai will either be vindicated by the numbers in two years, or it won't. But waiting for the numbers to tell me it was safe to start would have meant never starting at all.
I'll leave you with something Jony Ive once said: "The most important decisions are the ones that are most difficult to quantify."
The numbers got me here. The answer isn't in them.