Background
About Perves.
Perves spent a decade inside consumer marketplaces building the analytics that decided which sellers got promoted and which listings got suppressed. He worked on trust-signal scoring, seller-quality models, and the ranking systems that quietly move billions in commerce every quarter. He knows how platforms actually judge trust signals from the inside - not from a Search Central blog post.
He joined BGR Review's Thornhill analytics hub in 2023 to bring that discipline to reputation management. His removal eligibility scoring system runs at 94% accuracy against actual case outcomes across multiple jurisdictions - meaning when Perves says a review is likely to come off, it does. When he flags a case as low-probability, we tell the client up front rather than burning their retainer.
His revenue-lift model is the piece of work most clients quote back to their board. It turns a rating improvement (say 4.2 to 4.6 across a multi-location retail group) into a defensible incremental-revenue number, complete with confidence intervals and the assumptions behind them. It's what turns a reputation retainer from a marketing cost into a P&L line the CFO will sign off on.
Every promise in reputation management should be defensible in a spreadsheet. If a claim can't be modelled, it shouldn't be sold. That's the standard I hold every forecast, every eligibility score and every ROI number to.
Ranking systems don't care about your press release. They care about signals. My job is to help clients read those signals honestly - the good, the bad, and the ones we can move.
