Original test photography vs stock: a 90-page product-review split

Original test photography vs stock: a 90-page product-review split

The question: does first-hand experience evidence — your own product photos, not manufacturer images — correlate with ranking after the reviews-system updates?

A frequently-cited internal experiment: a reviews publisher split ~90 comparable product pages into two cohorts. Cohort A (45 pages) was rebuilt with original unboxing photos, photos of the product in real use, and short comparison shots with serial numbers visible. Cohort B (45 pages) kept manufacturer stock imagery and otherwise-identical copy updates. Over the following two reviews-system updates, Cohort A reportedly gained an average of 4.2 positions; Cohort B gained 0.6.

This maps directly onto Google's December 2021 reviews guidance, which explicitly asks for 'evidence that you actually used the product' and lists visual proof as an example.

Counter-evidence: original photography also tends to raise dwell time and lower bounce, so any ranking effect could be mediated by engagement rather than by Google parsing the images for authenticity.

Caveat: 45 pages per arm is a small sample, and the two cohorts were matched by the operator, not randomized. Selection bias is plausible.

What we still don't know: whether Google's systems detect first-hand imagery as a signal, or whether the photos simply make pages genuinely better and users vote with behavior. The case cannot separate the two.