Why I Killed My First Product (And Why Zero Sales Was the Best Outcome)

Total revenue from my first software product: zero pounds. Zero pence. Not a single sale.

And honestly? That might be the best thing that could have happened.

The backstory nobody asked for

For about six years, I ran a public speaking coaching business in London. Toastmasters division champion, weekly workshops, one-on-one coaching, the lot. At its peak, the business was doing north of £70,000 a year. I was good at it and people got real results.

But I was also working full time as a product manager at a finance/research company. Then my son was born in late 2024, and something had to give.

By May 2025, I told my business partner I was done. Not dramatically — I just stopped pushing. Dropped from two groups a week to one. Stopped doing sales calls. By October, it had quietly flatlined. I felt relief more than anything.

For a few months, I was just... a dad with a job who started learning piano as a hobby in the evenings. It was nice. But it wasn't enough.

The itch

By November 2025, the absence of a side project started to gnaw at me. I've always had this idea of building a one-person software business. Own the product, own the upside, build something that works while I sleep.

I'd started a JavaScript course in October. Halfway through, I heard about Claude Code — Anthropic's AI coding tool. I installed it, asked it to build a simple to-do app, and my jaw hit the floor.

It just... built the thing. Working code. In minutes.

I remember thinking: why the hell am I learning how to code when I can just build?

The build

The idea was obvious. I had six years of coaching expertise and an email list of 4,500 people. What if I built an AI public speaking coach — one designed by an actual coach, not a team of engineers who'd never stood in front of a room with shaking hands?

Claude Code built me a working prototype in a few nights. Not a mockup. A working app with auth, an AI coaching engine, and Stripe payments. Claude also helped me segment my email list, draft campaign sequences, and put together a marketing plan.

Within two weeks of starting, I had a product and a launch plan. Two weeks. That would have taken months a year ago.

The mess in the middle

Then life happened. My son's birthday party. A six-week trip to China — his first time meeting my family there, me working from our Shanghai office for two weeks. By the time January rolled around, I'd lost all momentum.

I want to be honest about what "lost momentum" actually looked like. It looked like doom scrolling at 9pm. Watching YouTube videos about building instead of building. Playing Zelda until midnight. Telling myself I'd get back to it next week.

This went on for about two months. I'm not proud of it, but I'm not going to pretend it didn't happen. If you're building something in the margins of a full life — a job, a family, a kid who wakes up at 6.30am — autopilot mode is a real risk. And it's hard to snap out of because nobody's holding you accountable except yourself.

The wake-up

In early March, I sat down for a structured coaching session with myself. (I spun up a Claude Cowork session, and use it as an accountability partner/coach with full context on my goals and projects. It sounds ridiculous. It works.)

That session hit different. I looked at the gap between what I said I wanted and what I was actually doing with my evenings, and something clicked. Not in a motivational-poster way. More like: that's enough, mate. Get back to work.

What came out of that session was clarity. Not about the product — about the identity. I'm a builder. Not a coach, not a product manager, not an aspiring entrepreneur. A builder. The products are just what comes out of that.

I picked up the email campaigns, finished the sequences, and launched.

The data

Warm list — about 300 former coaching members who knew me personally:

• Reconnection email: 45%+ open rate. Three people replied saying it was great to hear from me.
• Offer email (£199 lifetime access — they used to pay £299/month for a lifetime deal): three clicks. Zero sales.

Cold list — the rest of the 4,500 subscribers, offered a 7-day free trial at £29/month:

• 0.3% click-through rate. Zero sales.

Total revenue: £0.00.

The decision

Here's where most people would say "I need to tweak the landing page" or "maybe the price is wrong" or "let me try a different angle."

I didn't do that. Instead, I ran a structured decision session against the kill criteria I'd set before I built the product.

The signal was clear. My warmest possible audience — people who already knew me, trusted me, and had paid me hundreds of pounds for coaching — didn't even click on the offer. This isn't a copywriting problem. It's not a pricing problem. It's a demand problem with this specific audience.

They paid for the human. They don't want the AI replacement. And that's a perfectly reasonable position for them to hold.

Decision: keep the app running passively (it costs almost nothing), and move on to the next idea.

What I'm taking forward

The total time I invested was probably a few hours a night for about three weeks, spread across a few months. That's it. The cost of failure with AI-assisted building is genuinely tiny. A year ago, this would have been a six-month project and the sunk cost would have made it much harder to walk away.

The real shift is this: the bottleneck has moved. It's no longer "can I build it?" — with Claude Code and tools like it, a non-engineer can ship a real product in days. The bottleneck is now "should I build it?" And the only way to answer that is to build it, ship it, and let the market tell you.

I set my kill criteria in advance. I honoured them when the data came in. That's not failure. That's the system working.

What's next

This was product number one. The machine doesn't stop.

I now have something I didn't have before: a repeatable process. Ideate, build fast, test with real people, read the data, decide. The whole cycle took weeks of actual effort, not months. And every cycle teaches me something the last one didn't.

Zero sales on my first product isn't the end of the story. It's the first data point.