Yad2 is where Israel buys, sells, and rents almost everything — and its Real Estate vertical is one of the busiest markets in the country. The brief: pick an opportunity worth pursuing, prioritize it, and take it from a landscape scan all the way to a working prototype.
Most marketplace thinking optimizes for the advertiser — more leads, more listings, more contacts. I deliberately flipped the frame to the seeker: the person scrolling forty near-identical listings at 11pm, trying to decide which apartment is actually the better deal. Yad2 is excellent at helping people find listings. It does almost nothing to help them choose between them — and choosing is where the search actually stalls.
Finding listings is solved. Deciding between them isn't.
A serious apartment hunt isn't one search — it's a shortlist of five to ten listings that all look plausible. To compare them today, seekers open a tab per listing, scroll each one for price, size, floor, parking, safe-room (ממ"ד), balcony, elevator, and quietly build a spreadsheet in their head. It's slow, error-prone, and the details that matter most are buried differently in every listing.
- No shared frame — every listing presents the same facts in a different order, so nothing lines up
- Price without context — ₪/m² is the number that actually compares apartments, and nobody computes it by hand
- Decision fatigue — the more listings on the shortlist, the harder it gets, so seekers stall or settle
- The drop-off is invisible — the marketplace sees the search, not the choice that never gets made
Scan eight opportunities, then bet on one.
Rather than jump to a feature, I mapped the seeker's journey and pulled out eight opportunity areas across it — from search and filtering to evaluation, listing quality, and AI assistance. Three clustered around the same under-served moment: Evaluate & decide. I then ran a lightweight RICE-style pass over the strongest three candidates.
The winner
Reach, effort, and confidence all pointed the same way: highest impact on the decision moment, buildable on data Yad2 already has. It won the prioritization outright.
Higher ceiling, higher risk
A conversational guide with real upside, but heavier to build and harder to make trustworthy on a first release. A strong next bet, not the first one.
Useful, but adjacent
Surfacing demand and urgency helps, but it nudges the seeker rather than solving the core “which one” question. Parked.
Help people choose
One focused feature that turns a shortlist into a decision — the moment the whole search has been building toward.
Pick up to four listings, see them line up.
Side-by-Side Comparison lets a seeker mark listings straight from the results page (up to four), gather them in a tray, and open a single clean table where every apartment is described by the same rows in the same order. Fourteen parameters — price, size, price per m², rooms, floor, floors-in-building, elevator, parking, balcony, safe-room, AC, solar, storage, and accessibility — laid out so the trade-offs are obvious at a glance.
The table doesn't just list facts, it takes a gentle position: for each measurable row it flags the best value across the set — the lowest ₪/m², the larger apartment, the higher floor — while honest ties get no badge at all. The seeker still decides; the feature just does the arithmetic they were doing badly in their head.
A working prototype on real Tel Aviv listings.
To make the pitch concrete, I built a Yad2-lookalike results page in its actual visual language — the nav, the pill filter bar, a live map, and row-style listing cards — populated with real for-sale listings from Tel Aviv. You browse, toggle listings into the comparison tray, and open the full RTL comparison overlay with best-value highlights across all fourteen rows.
The map is real and locked to Israel, price pins track selection, and the whole flow works end-to-end. It's not a mockup of a screen — it's the feature, running, on data a seeker would actually recognize.
Yad2 already helps you find every apartment. The hard part is choosing one.