Reviews shape purchases.
No one trusts reviews.

16 of 21 participants described navigating directly to the photo gallery and treating user-uploaded images as their primary trust signal. They specifically favored photos with imperfect, natural backgrounds — a messy desk, an unmade bed — over staged shots, which they flagged as potentially AI-generated or brand-supplied.

Approximately 3 of 21 participants still found text-heavy reviews with 'Verified Purchase' badges credible even without photos, particularly for commodity items where visual proof mattered less than star volume.

For categories like apparel, home goods, and electronics, text-only reviews have effectively lost their persuasive power. Platforms that don't prominently surface user photos are ceding trust to any competitor that does.

5 of 21 participants explicitly filter to 1-star and 2-star reviews before reading anything else, and 13 of 21 described dismissing generic positive text outright. Users treat negative reviews as more informationally dense — a source of objective flaws and recurring defects rather than subjective praise.

Approximately 4 of 21 participants described reading top-rated reviews first, specifically for well-known brands they already trusted — using negative reviews only as a final confirmation step, not a starting point.

Burying or moderating negative reviews doesn't just fail — it actively backfires. Users who can't find honest negatives assume they are being hidden, which increases suspicion rather than reducing it.

13 of 21 participants described strong negative reactions to reviews labeled 'incentivized' or suspected of being free-product submissions. They associated promotional tags with bias by definition — regardless of the review's actual content quality. Robotic phrasing, repetitive buzzwords, and five-star unanimity triggered the same skepticism automatically.

Approximately 3 of 21 participants felt incentivized reviews could still carry useful data if they included photos and specific critiques — a pragmatic minority willing to discount the bias penalty when objective content was present.

The 'incentivized' disclosure mechanism, designed to increase transparency, has backfired — it now signals corruption rather than honesty. The disclosure itself has become the red flag.

8 of 21 participants described leaving the retail platform when the stakes were high — expensive items, unfamiliar brands, or suspiciously uniform reviews. They sought out Reddit threads, TikTok unboxings, and Google results as sources of organic, unfiltered opinion before returning to purchase.

13 of 21 participants did not mention cross-platform behavior, particularly for low-cost, repeat purchases from brands they already trusted — suggesting platform exit is triggered by stakes, not habitual skepticism.

A platform that doesn't earn trust at the review layer loses the purchase moment to a different platform or influencer. The consumer leaves, validates elsewhere, and sometimes buys elsewhere too.

6 of 21 participants explicitly cited preference for verifiable, objective data — specific dimensions, fit relative to body size, durability after extended use, or sensory descriptions like texture, scent, or weight. Specificity was interpreted as a proxy for genuine lived experience, making it credible where vague praise is not.

6 of 21 participants anchored heavily on 'Verified Purchase' badges and review volume as their primary trust filter — deprioritizing content quality in favor of quantity signals, suggesting structural trust proxies haven't fully collapsed for all users.

Review forms that only ask for a star rating and open-ended text are generating low-trust noise. The absence of structured, objective input prompts directly reduces the trust value of the entire review corpus.