I almost shut down my property listing site. Turns out buyers weren't leaving — they just couldn't find anything

I built a property classifieds site for the Gulf market about two years ago.

Not a massive play. Just a lean platform where landlords and small brokers could post apartments, villas, commercial spaces. The region made sense — high rental churn, expat population constantly moving, paper-based listings still weirdly common in smaller cities.

For the first six months, things moved. Listings were going up. Some leads were flowing.

Then it plateaued. Hard.

Traffic stayed flat. Listing count kept growing. But inquiries per listing dropped every single week. I ran every analysis I could think of — pricing, ad quality, category structure. Nothing pointed anywhere useful.

I almost blamed the market. Told myself maybe this niche was too competitive. Considered pivoting.

Then someone on my small beta user group said something that stopped me:
"I search for flats in this area and get results from the other side of the city. I just stopped searching."

That was it.

Buyers weren't leaving because the supply was bad. They were leaving because they couldn't spatially filter anything. The search was keyword-based. Type "2BHK apartment" and you'd get every 2BHK on the platform regardless of which neighbourhood, which building cluster, which side of the highway. In a market where location is everything — where being 10 minutes from a metro station changes rent by 30% — my search was functionally useless.

So I added map-based search.

Not the fancy kind. Just: plot listings on a map, let buyers draw a rough radius, filter results to that zone. Add a few proximity flags — near schools, near metro, near hospital. Basic stuff.

What happened:
→ Session time went up immediately. People were actually exploring, not bouncing after two useless results.

→ Inquiry-to-listing ratio recovered within three weeks. Same listings. Different discovery layer.

→ Brokers started posting more because their listings were finally being seen by people who actually wanted that area, not buried under city-wide noise.

The thing that frustrated me afterward was how obvious it sounds in hindsight. Property search without location filtering isn't really search. It's a sorted list. And nobody browses a sorted list when they're trying to figure out where to live.

A few things I'd push on if you're running any geo-sensitive marketplace:

1. Don't assume low inquiries mean bad supply. Sometimes supply is fine and discoverability is broken. They look identical in aggregate metrics.
2. Keyword search fails the moment your inventory gets dense. Below 200 listings, people scroll through everything anyway. Above that, without spatial filtering, you're just showing people a wall of noise.
3. The map doesn't need to be beautiful. Mine was barely styled when I shipped it. Pins on a map with a radius slider. That's all it took. Polish it later.
4. Proximity filters matter more than you'd think. "Near metro" or "near school" sounds like a nice-to-have. In practice it's the filter that converts browsers into people who actually message a landlord.

I didn't pivot. Didn't rebuild. Just changed how buyers navigate what was already there.

The platform is still running. Inquiry rates are consistently higher than they were pre-map. And that beta user who complained? He ended up renting through the platform two months later.

Sometimes the product isn't broken. The surface is.

Happy to talk shop if anyone else is building in the property or geo-sensitive listings space.