Your reviews expire. Three-quarters of customers won't look at anything older than three months — and a machine now reads your profile before your customer does.
Download the full report with benchmarks for your vertical
Get a DemoMost operators think of reviews as something you accumulate. Get to a hundred, hold a 4.8, and the job is done. That model died this year. Reviews are no longer a balance you build — they're a stock that spoils, and the shelf life is about ninety days.
74% of consumers now only consider reviews written in the last three months, and 32% only look at the last two weeks — up from 20% a year ago. That means a 4.9 built patiently over six years, whose newest review is from spring, reads to a prospect exactly like a business with no reviews at all. You didn't lose the rating. You lost the only part of it anyone still looks at.
These figures come from a single named survey of 1,002 U.S. adults published in February, with the year-over-year comparisons the publisher reported. What stands out isn't any one number — it's that all of them hardened in the same direction inside twelve months.
74% of consumers only consider reviews from the last three months. 32% want something from the last two weeks, up from 20% a year earlier, and 18% are only swayed by the last seven days.
68% will only use a business rated 4.0 or higher — up from 55% a year ago. 31% now hold out for 4.5 or better, nearly double last year's 17%. The floor rose 13 points in one year.
45% used an AI tool for a local business recommendation in the past year, up from 6%. Among people aged 30 to 44 it's 64%. Over the same period, use of Google reviews fell from 83% to 71%.
89% expect owners to respond to reviews. 80% are more likely to use a business that replies to all of them, 42% are unlikely to use one that never replies, and 19% now expect a same-day response — up from 6%.
In one year, AI went from a rounding error to the third-most-used tool for finding a local business. The instinct is to read that as reviews mattering less. The data says the opposite — and the distinction decides what you do about it.
Among people who use AI for recommendations, 88% fact-check what it tells them and 97% at least sometimes verify against real reviews — 42% always go to the review platform itself. Only 23% stop at the summary. So the profile didn't stop mattering; it became the source material for a second reader who summarizes you to the first. A thin or stale profile doesn't just lose the customer. It gives the machine nothing to say about you.
Route service has a structural advantage here that almost nobody uses. A restaurant gets one shot at a review. You are at the same house twelve to fifty times a year. None of what follows costs anything, and the first four can start on your next route day.
Not on a schedule, and not by blast. The visit where you caught the problem early or fixed something they'd noticed — that's the ask. You get dozens of those moments a year per customer.
Fifty chances a year, not oneA handful of fresh reviews every quarter beats forty in one burst and then silence. Steady beats big, because the filter is recency, not total.
Clears the 74% recency cutoff81% expect a response inside seven days. On anything negative, reply the same day — 19% now expect same-day, triple last year. A reply is also the only part of the profile you control outright.
89% expect an answer“If you mention the service and the town, it helps people nearby find us.” Detail is what a human scans for and what a summary has to work from. “Great job” gives neither of them anything.
Gives the summary something to saySince October 2024 the FTC rule on consumer reviews has carried civil penalties up to $51,744 per violation. A discount for leaving any review is fine. Conditioning it on a five-star review is not, and neither is suppressing negatives.
Up to $51,744 per violationYou are already doing the work that produces good reviews. The gap isn't service quality — it's that satisfied customers almost never post unprompted, and the ones who did post for you two years ago no longer count.
A review request that fires automatically when the tech marks the stop complete
→ The ask lands while the clean pool is still in front of them
→ Fresh reviews arrive continuously instead of in once-a-year bursts
Proof of service attached to the visit the customer is reviewing
→ Specific reviews, which is what both the reader and the summary need
→ A rating that stays above the four-star filter as the book grows
Ninety days from now, the reviews you have today will not count. The ones you ask for this month will.
Consumer review behavior, recency, star-rating thresholds, and response expectations from the BrightLocal Local Consumer Review Survey 2026, published February 11, 2026 — a representative panel of 1,002 U.S. adult consumers surveyed via SurveyMonkey, with year-over-year comparisons as reported by the publisher. AI usage, trust, and verification figures from the companion AI release of the same survey, published March 10, 2026; figures describing AI users are drawn from the 455 respondents who had used an AI tool for a local business recommendation in the previous twelve months. Review-regulation figures refer to the Federal Trade Commission's final Rule on the Use of Consumer Reviews and Testimonials, 16 CFR Part 465, effective October 21, 2024, which carries civil penalties of up to $51,744 per violation, adjusted annually for inflation. This report is not legal advice; consult counsel on your own review practices. Consumer survey figures describe local businesses generally, not route service specifically, and no ProValet operator data was used in this issue.
See what review requests triggered on service completion look like on your routes. Book a 15-minute walkthrough.