The Post-Mortem

This is Part 4 of a series on MonitorIntent. In Part 3, we dealt with churn, the Clay defection, and watched our moat disappear. This part is the post-mortem - what killed us and what I’d do differently.

We shut it down. Here’s what I learned.

After 10 months of building various products, 7 paying customers, and 3,000+ qualified leads delivered - MonitorIntent is done. This is the honest breakdown of what went wrong, what I’d salvage, and what I’d do differently.

The 5 things that killed us

1. Unsustainable economics. We were unit profitable. That sounds good until you realize the fixed costs to keep the business running were brutal. Lots of our scraping infrastructure - think API subscriptions, cold outreach infrastructure, tools that help building SaaS products, research, etc. - all had fixed costs. Revenue couldn’t outpace expenses fast enough, and the gap wasn’t closing.

2. Our technical moat eroded. We’d built a technically sophisticated intelligent scraper to source leads from LinkedIn and other platforms. Then companies like BrightData started selling that same data to anyone with a credit card. Our technically sophisticated intelligent scraper went from competitive advantage to commodity overnight.

3. The data sources got sabotaged. Social media platforms started suing data providers and clamping down hard on scraping. The very foundation our product sat on was actively being dismantled. Not by competitors - by the platforms themselves.

4. Bad incentives for our customers. This one stings the most. We were asking customers to work hard with our curated, high-quality leads. But it was cheaper for them to just spam 10x more low-quality leads and play the numbers game.

5. A brutally saturated market. Well-funded players like Websets and Extruct AI were commoditizing the space - and doing it cheaper. We figured our edge would be the ability to qualify leads reliably. But there are already startups doing that. Building new tech and benchmarking it against them would be a whole other startup. Not a pivot - a restart.

What I’d salvage

Not everything was a loss.

We wrote and rewrote core components 2-3x over the course of development. A lot of that code - mostly around LLM tooling - is genuinely reusable. I plan on open-sourcing some of it.

I also think there’s a real opportunity in using techniques like GRPO fine-tuning on existing LLMs specifically for lead qualification. The caveat is that GRPO requires objective data to train on, and that kind of data was simply very hard to get in this space. But if someone cracks that data problem, the approach would be powerful.

On a separate note, I’d look at inbound leads if I were to re-enter this space. People are risk-averse. A bird in hand is worth two in the bush for most, and they’ll pay much more to retain what seems like (a bird in hand) than to try casting a net at those in the bush. That makes cold outbound software a fundamentally harder sell.

The hard-earned lessons

Speed matters. Not “move fast and break things” platitude speed. I mean: the window between your idea being novel and it being commoditized is shrinking every month. Especially in AI.

Sell while you build. Or before. We spent too long perfecting before validating willingness to pay at scale. The 7 customers we had were almost all startup founders. That’s a narrow, price-sensitive segment. We should have been selling from day 1 to a broader market.

But I’m not exactly sure what we should have been selling. Perhaps a waitlist, or some sort of commitment before the product was ready.

I am aware of people who run consulting as a way to develop IP over time - selling semi-automated solutions before they actually build out the final product. That always seemed like an interesting approach to me.

Never build without a co-founder. I don’t care how smart you are. Don’t do it solo. My partnership with my co-founder was one of the most rewarding experiences of my life. I got to build something exceptionally interesting with one of my best friends. That part I wouldn’t trade.

But the weight of a startup - the technical decisions, the sales calls, the strategic pivots, the 2am debugging sessions - you need someone in the trenches with you.

The numbers, honestly

• Bootstrapped. No outside funding.
• Burn rate: a few thousand a month. Lean, but still bleeding.
• 7 paying customers. Almost all startup founders.
• Churn was too high. Users consistently perceived more value in inbound leads than what we were delivering outbound.
• 10 months from first line of code to shutdown, after various pivots, and ~7 months from our latest product.

What’s next

I’m moving on to industry research and consulting. The partnership was incredible, and I fully expect to build something with my co-founder again - after a bit of a break.

If you’re interested in working together on research projects or have specialist consulting work, I’d love to chat. You can learn more at my Work with me page.