Writing · Optima

The moment was not available

In November 2025 I recommended that we shut down Optima, a company I co-founded and had spent seventeen months on. This is how that decision got made: what we believed going in, what we tried, what the evidence said, and why I made the call.

Where it came from

Optima started as a sample of one. I had the problem myself, as a student: nerves in the minutes before an exam. I assumed everyone had it before everything. So in July 2024 my co-founder and I started building the thing I wished I had: an app that generated one short session for one specific high-stakes moment (an exam, a game, a pitch), built from breathwork, reframing and a pep talk. In the beta, heart-rate and HRV data from a wearable set the pacing.

We did our customer discovery after we built it. That is the wrong order, and I learned it the slow way. A product you need yourself feels like proof. The 45 conversations came later, and they reshaped everything they could reach. They could not reach the founding assumption, because by then it was already built.

What we believed

Three bets sat underneath it. Here they are, so the rest of this can be scored against them.

1. People facing a high-stakes moment would reach for a tool in that moment.
2. Personalization from wearable data would set us apart from generic meditation apps.
3. At least one segment would use it repeatedly and pay for it.

The first bet was the company, and the other two rested on it.

What we tried, in order

Positioning. We launched as a general meditation app. Discovery said to go narrower, into specific situations like exams, so we repositioned to mindset training for those moments. Usage said something harder than discovery did: people would not open an app in the minutes before the situation.

Segments. The first segment was everyone, which was naive. The second was students, and focused targeting worked on the surface. Reception improved and weekly signups more than doubled. Usage outside exam weeks stayed flat. The moments were too rare to build a habit on.

The product. We rebuilt the generated session from scratch and A/B tested the interfaces and flows around it. A better session could not beat the social reality of the moment. Before an exam, people reach for their notes or a friend, not an app.

I tracked all of it in SQL: weekly signups and weekly usage from session data, compared across each change. Signups moved. Usage did not.

The wall

The interviews and the numbers said the same thing. The minutes before a high-stakes moment already belong to ingrained habits: cramming the notes, talking to a friend. There was no empty slot waiting for a product to fill it.

The interest was real. People liked the idea and signed up. The moment was not available. Discovery running as a correction can steer a product. It cannot pick a different moment.

The call

By November 2025 the three levers a small team can pull (positioning, segment and the product core) had each been pulled and given a fair test. The pattern had not moved, and the evidence showed no path to product-market fit. The alternative to stopping was spending the remaining angel capital on hope. More runway was not going to change the answer, because runway was never the problem.

It is a strange thing to recommend for your own company: seventeen months of work, other people's money, and a co-founder who had been in it from the first day. I laid out what we had believed, what we tried, what each change did to signups and to usage, and what the interviews said, and I recommended we wind it down. We did it in an orderly way. Most decisions are cheaper than that one.

What I took from it

Sequence. Ask before you build. The same 45 conversations, held before the build, could have changed what we built. By the time we held them, they could only steer. At SKADI, the next company, 76 conversations came before anyone designed anything, and they killed the wrong idea, backcountry tracking, before it cost us anything. Same discipline, opposite order, and the order turned out to be most of the value.

Check that the moment is available. Interest and a moment someone will actually give you are different things, and Optima only ever had the first. Doubling signups felt like progress, and at the top of the funnel it was. The number that decided the company was the one that never moved.

Conviction is for positions, not for sunk costs: hold a view for clear reasons, and when the evidence moves, move with it. I would make the same call again.

Written by Nasmaan Ibrahim for nasmaan.com in October 2026, eleven months after the decision. Figures are from Optima's own data.
Back to the record · Seventy-six conversations, one page · The first fifteen minutes