Methodology & Data
Everything needed to reproduce or dispute these rankings.
How comparisons are made
Every comparison starts from the same source photo of a real empty room. Each vendor's tool stages that identical photo. Voters see two staged results side by side, with the original room for reference, and pick the better one. Tool names are hidden until after the vote, and left/right position is randomised on every serve.
Scoring
Ratings use standard Elo (K=32, starting rating 1500). A win counts 1, a tie ½. Votes marked "both are bad" are recorded and published but do not move ratings. Vendors under 20 rated games are flagged provisional.
Vote hygiene
- One vote per comparison per browser session; sessions are anonymous random IDs.
- Votes cast faster than 800ms are recorded but excluded from ratings.
- A minimum interval between votes is enforced per session.
- IP addresses are stored only as salted hashes, for abuse analysis. No trackers, no analytics scripts, no ads.
- Ratings can be deterministically recomputed from the public vote log at any time.
Independence
No vendor pays for inclusion, placement, or removal. We hold paid accounts with the tools we test, purchased at retail like any customer. If this project ever takes any form of vendor money, that fact will be disclosed on this page and those vendors will be excluded from rankings.
Source photos
Source rooms are photographs we own or that are licensed for commercial reuse. Each room's provenance is recorded. Staged outputs are the product of running the identical source photo through each vendor's standard workflow with default settings and a comparable furnishing style request.
Open data
/api/leaderboard.json— current ratings./api/export/votes.csv— full anonymised vote log (pair, vendors, choice, timing).
Contact / disputes
Vendors: if you believe a comparison misrepresents your tool (wrong plan, outdated version, non-default settings), open a dispute and we will re-run the room and document the change publicly.