Reviews touch the journey at four distinct points: discovery, comparison, hesitation, and post-purchase rationalisation. Each stage produced a different signal in the cohort. The frequencies are higher than the standard surveys report because session replay captures behaviors people forget to mention in self-reported polls.
1\. Reading reviews before clicking the listing
72.4% of shoppers in the e-commerce cohort read at least one review snippet on the SERP before clicking through. The average shopper read 1.8 snippets. The behavior was less common in home services (51%) and more common in hospitality (78%) where SERP review snippets are richer.
2\. Sorting by lowest rating first
28% of shoppers actively sorted reviews by lowest first. The percentage rose to 41% for purchases above $200. Shoppers do not avoid negative reviews; they hunt for them to test how the business responds and whether the complaint pattern is a deal-breaker.
3\. The 30-second skim
Average time spent reading reviews was 94 seconds in e-commerce, 112 seconds in hospitality, and 158 seconds in home services. The shape of the curve is consistent: a 30-second initial skim, then a deeper read of two to four full reviews if the skim raises a question.
Average review read time across the cohort: 121 seconds. The shoppers who converted spent 1.7x longer reading reviews than those who bounced.
4\. Reading the owner response, not just the review
57% of shoppers who clicked into a negative review also read the owner response. When a response was present and resolution-oriented, conversion intent on the post-survey was 2.1x higher than when the response was missing or defensive.
5\. Recency check before trust
64% of shoppers checked the date of at least one review. Reviews older than 90 days were treated as background noise; reviews from the last 30 days carried 3.4x more weight in the post-survey trust rating.
6\. Star average pattern matching
Shoppers do not read the headline star average; they read the distribution. 81% glanced at the star bar histogram before scrolling. Profiles with a 4.6 average and a clean distribution outperformed profiles with a 4.8 average and a suspiciously empty 1-2 star band by 19% on conversion.
Shoppers who converted spent 1.7x longer reading reviews than shoppers who bounced. Reviews are not a checkbox on the way to the buy button; they are the deliberation itself.
7\. The hesitation pause on the contact step
41% of home-services shoppers paused on the contact form for more than 10 seconds and scrolled back to the reviews. The pause-and-rescroll behavior predicted abandonment 67% of the time when the most recent review was older than 30 days.
8\. Cross-channel verification
38% of shoppers checked reviews on a second platform before purchasing. Most-checked combinations: Google + Trustpilot for e-commerce, Google + Yelp for home services, Google + TripAdvisor for hospitality. Cross-channel checking was strongest for first-time customers and almost absent for repeat customers.
9\. The photo-and-detail jackpot
Reviews with photos kept attention 2.8x longer than text-only reviews. Reviews mentioning a specific product attribute or service step were rated 'most useful' 3.1x more often than generic praise. Specific beat positive.
10\. Negative-review immunity threshold
Shoppers tolerate negative reviews up to a clear threshold. Profiles with 6 to 14% negative reviews were trusted at the same rate as profiles with 2% negative reviews; both outperformed profiles with zero negative reviews on perceived authenticity. The tolerance ceiling was 22% negative; above that, conversion intent dropped 41%.
11\. Post-purchase confirmation reading
29% of shoppers returned to read reviews after the purchase. The post-purchase read functioned as cognitive-dissonance management: shoppers were checking that they had made the right call. Profiles with strong recent reviews reduced return rates by 8% in the e-commerce cohort.
