Spotting revenue in a support channel
Nobody asked for this. It started as a pattern in a Slack channel, became a simple product, I built alone over four days, and grew into a business unit generating £3.5M a year across 12 manufacturers.
The signal
Auto Trader's Customer Success Slack channel was a goldmine. One dealership listed a car straight from the Geneva Motor Show, before launch. The ad was pulled — but in three days it took 40,000 views and 2,000 leads.
Everyone read that as a compliance problem. I read it as demand arriving months before we had anything to sell it. When a manufacturer announces a model, search spikes immediately and we had no product pointed at that moment.
I was convinced this was a product, not a one-off. A single-model landing page would give users what they were already searching for, strengthen our SEO, and open an OEM revenue stream we didn't have.
I couldn't get it funded on a hunch, so I stopped asking. I facilitated our next hackathon across design and product and pitched it alongside a few other ideas. It picked up traction with peers and stakeholders — but traction isn't a build plan, and there was no engineering capacity.
Almost all of the work that followed happened outside my day job. Not as heroics: it was the cheapest way to test a commercial thesis I believed in without asking the business to bet on it first.
The conviction
The Monday after the hackathon we had concepts, prototypes and a user journey, and no plan to build any of it. That afternoon the rough prototype reached Vauxhall's UK Marketing Director, who wanted to buy it immediately for the Corsa-E launch.
The deal was done. Now it had to exist.
The deal came before the build
Four days, one person and no AI
Learn what's needed, make it work, ship it with minimal resources so you learn fast. I used Squarespace as a CMS and recoded the template to mirror Auto Trader's navigation and footer, integrated our Tailwind design system, and used HTML, CSS, JS and Python to close the gaps.
Under four days later the Corsa-E page was live with Hotjar and Snowplow tracking. The build and deployment brought in £150,000, including end-to-end user monitoring through an integrated Snowplow API.
What I take from this: commercial opportunities usually show up as noise before they show up as a business case. The deleted advert was a support ticket to everyone who saw it. Reading it as demand, and then being willing to prove it rather than pitch it, is the part I'd repeat.
Measuring what matters
I built a Team of 4: 2 front-end engineers, 2 product designers.
56 landing pages in the first 12 months.
36M unique searches.
175s average dwell time.
12 vehicle manufacturers and their agencies.
14% conversion from site-wide traffic drivers to hubs.
£195k annual business unit cost.
£3.5M revenue by year two — an 18x return on run cost.
£850k Biggest single deal, Ford EV range.