Why we publish our methodology in public (and why nobody else does)
Every local services marketplace has a ranking algorithm. Almost none of them are willing to publish it. Here is why we do, and what we think it costs us.
There is a short answer and a long answer to why HiredLocalPros publishes its full ranking methodology. The short answer: because we think a directory whose formula is secret is a directory whose formula can be bought. The long answer is worth five minutes.
A ranking formula is a contract between a directory and the people who read it. If you search 'best pool cleaner in Chino Hills' on Yelp and the first result is labeled 'Sponsored,' that contract is honest — you know exactly what you are looking at. The problem is the second, third, and fourth results. Those are also shaped by ad spend, by review-gating (are negative reviews filtered?), by the pro's Yelp advertising relationship, and by a dozen other commercial signals that nobody outside Yelp's ranking team ever sees. The contract is secret, so the contract is violable.
We decided to write our contract down. The methodology page lists the five scoring dimensions, the exact weights, the signals that feed each dimension, the confidence thresholds that determine which badge a pro earns, and the five commercial rules (N1 through M4) that govern how money can and cannot affect ranking. Rule N1 is the most important: no pro can pay to rank higher, ever. The other rules govern things like subscription tiers that add value (faster response SLAs, API access) without affecting rank.
Publishing this has costs. Competitors can copy the formula — fine, the formula is not the moat. Some pros will be unhappy when they see exactly why they are number six instead of number three — also fine, we would rather have an accurate ranking than a happy pro. Homeowners and regulators can audit us — that is the whole point.
The reason nobody else publishes is that most directories make their money from pros paying for higher placement. If the formula is public, the arbitrage is dead. We chose a different revenue model — CPA commission only when a job closes — so that publishing the formula does not break our business. It took longer to build and it is harder to scale, but it means our ranking can be read by a skeptical homeowner and still hold up.