Spiky growth
When I joined Jitty, the product was promising but growth was manual. Users came when we ran growth activities and stopped when we didn't. Every spike in the traffic chart was something I'd pushed; every trough was me working on something else. We needed a channel that kept working when nobody was pushing it.
The problem with SEO
SEO is the lifeblood of property portals, and Rightmove and Zoopla dominate it. They've spent two decades building authority on every term that matters. We had a domain rating of 30 and no realistic chance of ranking for "houses for sale in Manchester" or any of the thousands of head terms the incumbents own. This is why early-stage property startups usually write SEO off entirely.
Finding our edge
Jitty's core innovation is using LLM vision to read property photos and turn what's in them into structured data: features that appear in the images but never in the listing description. We'd also launched vision search, which let users search for almost anything: "warehouse conversion", "south-facing garden", "art deco house".
That data was the opening. Rightmove and Zoopla dominate the short tail, but there are hundreds of thousands of long-tail terms they rank for without owning: "art deco homes for sale", "land with planning permission", "houses with a downstairs toilet". They can't build good pages for these queries because their data doesn't contain the answers. Ours did. I did the keyword research and identified hundreds of terms where we could genuinely compete.
Proving it out
There was reasonable internal scepticism, both about how long SEO takes to show results and whether the prize was big enough to justify engineering time. So I set out to prove the channel as cheaply as possible.
I did everything I could without engineers: the keyword research, the page mapping, the content plan. I built traffic projections to make the case for the engineering work I couldn't do myself, and kept that ask to the 80/20 — the smallest set of changes that would tell us whether the channel worked.
We launched in May 2025. By June, users were arriving from search. By August, it was our largest traffic source.
Scaling it up
Early-stage companies need generalists, but a generalist's job is to validate a channel, not to scale it. Once it's validated, the goal is to hand it to the best-placed person as fast as possible.
I found and brought in an agency with domain expertise in exactly this: programmatic SEO for marketplaces. With real traffic numbers behind the channel, I also got buy-in for proper engineering resource, and we overhauled the site. The biggest changes:
Fixed rendering and crawling. Google couldn't reliably render a chunk of our pages, and was wasting crawl budget on low-value ones. We fixed the rendering issue and cut the crawl bloat so the rest of the work could actually be indexed.
Built out pages for every place and property type worth ranking for. We created pages for 1,058 locations where we had 10+ listings, organised from country down to neighbourhood and generated automatically from our live inventory. Keyword research also found 40 more property types people were searching for — barn conversions, eco homes, houses with annexes — and we built pages for those too.
Let the inventory decide what gets indexed. Previously we were manually choosing which pages Google should index. We replaced that with a simple rule: any page with more than three homes on it gets indexed automatically. A page with enough homes has real value to a searcher, and therefore a chance of ranking. Combined with the location and property-type set, this meant hundreds of long-tail pages could switch themselves on and off as our inventory changed, with no one curating the list.
Automated internal linking. Every page now links to related locations and property types, and those links update weekly based on where we have the best chance of ranking — scored on search volume, competition and how many listings we have. The site keeps promoting its own best opportunities without anyone touching it.
Outcome & impact
Average monthly visits grew 25x, and the channel has driven millions of unique visitors with zero acquisition spend. Jitty went from growth that stopped whenever I did to a channel that compounds on its own.