BrightLocal published the sixteenth Local Consumer Review Survey on 11 February 2026: 1,002 US adults on a SurveyMonkey panel, nineteen charts, and a set of headline numbers that will be quoted in pitch decks all year. We opened every chart and read it against the text written around it.

Most of the findings hold. Six do not match the prose that describes them, and the loudest one of all, the arrival of AI, compares two different questionnaires. This is a reading of the data for a small business in Europe, a market the survey does not cover at all.

What the survey measured, and whose market it describes

This is a self-reported online panel: 1,002 adults in the United States, set against 1,026 the year before. It records what people say they do about reviews, not what anyone watched them do, and it contains no respondents from Poland, Ukraine or anywhere else in the European Union.

The age quotas are even, at 22% aged 18 to 29, 25% aged 30 to 44, 28% aged 45 to 59 and 25% aged 60 or over. That balance is worth knowing, because review habits differ sharply by age and a skewed panel would show up here first.

The sample size is not the weak point here; self-report is. Research on the intention-behaviour gap keeps finding that stated intentions predict observed conduct far more weakly than they predict what people report about themselves. Read a line like “31% will only use a business rated 4.5 or above” as a stated rule, not a measured threshold.

BrightLocal sells reputation software. That does not make the data wrong, and the methodology is published openly, which is more than most vendor research offers. It does mean the framing leans towards “reviews matter more than ever”.

Almost everyone reads reviews, but the list of places got longer

Ninety-seven percent read reviews for local businesses and the average person consults six different sites. Google leads every platform by a wide margin at 71%, though that sits twelve points below last year, while Facebook, Yelp, YouTube, Apple Maps and a row of directories all gained ground.

BrightLocal bar chart titled Video platforms and AI are rapidly reshaping local recommendations once dominated by traditional review sites, comparing 2026 and 2025 usage for 25 platforms from Google at 71 percent down to other generative AI tools at 4 percent
The full platform table. Note the dashes in the 2025 column against ChatGPT, Google AI Mode, Gemini, Microsoft Copilot and Claude. Chart: BrightLocal, Local Consumer Review Survey 2026, reproduced under the terms on that page.

Facebook reaches 49%, Yelp 44% and YouTube 43%. Apple Maps almost doubled to 27%, Trustpilot rose to 21%, and Healthgrades and Yellow Pages more than doubled off small bases. Attention is spreading over more platforms rather than moving from one to the next.

The exception is local news sites, down from 48% to 29%. BrightLocal attributes Google’s own slide partly to supply: its SMB Marketing Report 2025 found only 35% of small businesses have a Business Profile at all, so a large share of searches meet no Google reviews to read.

The 45% AI figure is a net across six tools, not one habit

The most quoted number in the report, a leap from 6% to 45% for generative AI, is an aggregate. The platform chart lists five named AI tools with no 2025 bar whatsoever, meaning they were not offered as answer options that year. Taken alone, the biggest of them reaches 31%.

Chart titled The 45% AI figure, tool by tool, showing ChatGPT 31 percent, Google AI Mode 23 percent, Gemini 17 percent, Microsoft Copilot 14 percent and Claude 9 percent each marked new in the 2026 list, and other generative AI tools at 4 percent
Five of the six had no answer option in 2025, so the 6% to 45% comparison is not like for like.

That does not make AI irrelevant to local discovery. It means the honest version of the claim is narrower: roughly three in ten say they have used ChatGPT for a local recommendation, and the combined reach of all six tools is 45%, measured for the first time this year.

Trust is a plurality rather than a majority. Forty percent agree they trust AI for local business recommendations while 32% disagree and 28% sit in the middle. Asked whether they trust AI as much as online reviews, 42% agree and 33% disagree.

Review summaries are the quieter story. Eighty-two percent read the AI-generated summary above a set of reviews, and 23% would decide on that summary alone. Google now generates these in Maps with Gemini, which makes the aggregate impression of your reviews, not any single review, the thing a customer reads first.

The star bar rose, and the chart behind it does not fully add up

Thirty-one percent say they will only use a business rated 4.5 or better, up from seventeen. That particular move is large and clean. The cumulative figure built on top of it is not: the same chart’s 2025 bars add up to 50%, while the report quotes 55% for the identical measure.

Grouped bar chart titled The minimum star rating consumers say they need, comparing 2026 and 2025: 4.5 stars or above 31 versus 17 percent, at least 4 stars 37 versus 33, at least 3.5 stars 14 versus 17, below 3.5 stars 9 versus 6
The four plotted options sum to 91% in 2026, so at least one answer is missing from the chart.

There is a second gap. The text says 10% will only consider a five-star business, but no five-star row appears on the chart, which is why the plotted answers stop at 91%. The 68% headline for “four stars or more” therefore excludes the strictest group of all.

Volume matters alongside the average. Forty-seven percent will not use a business with fewer than twenty reviews and only 9% are comfortable with five or fewer. That second figure is the one that bites a new business or a new location.

The recency window moved from months to weeks

Consumers pulled their definition of a current review inwards. Asked how recently a review must have been left to affect their decision, more people chose the past week and the past fortnight than a year ago, while every window from one month outwards lost share.

Grouped bar chart titled How recent a review has to be to sway a decision, comparing 2026 and 2025: past week 18 versus 11 percent, past two weeks 14 versus 9, past month 24 versus 27, past three months 17 versus 21, past six months 10 versus 13, past year 7 versus 11, recency does not matter 9 versus 8
Single choice. The short windows grew and the long ones shrank.

Adding the two shortest answers gives 32% who want something from the last fortnight, against 20% last year. That is the finding with an operational consequence: a quarterly push that lands thirty reviews in a fortnight and nothing for ten weeks now reads as stale for most of the quarter.

Fewer people are writing reviews, especially critical ones

The supply side moved the other way. Sixty-nine percent left a review over the past twelve months, yet the share writing up a bad experience dropped six points and the group who say they never would has edged upward. Businesses now want more reviews from a population writing fewer of them.

Grouped bar chart titled Who wrote a review in the last 12 months, comparing 2026 and 2025: yes about a positive experience 60 versus 64 percent, yes about a negative experience 29 versus 35, no but might in future 25 versus 25, no and never would 6 versus 4
The fall in negative reviews clears the survey's error bar. The fall in positive ones does not.

Most reviewers are occasional: 75% wrote between one and ten reviews over the year. Here the write-up and the chart part company again, because the text claims 20% wrote more than ten while the plotted buckets add to 26%, and it puts the heaviest reviewers at 7% where the chart gives 10%.

Asking still works, and one pair of numbers in the report looks contradictory until you check the denominators. Seventy-eight percent were asked for a review; 65% of everyone surveyed wrote one after being asked, which is 83% of the people who were actually asked.

BrightLocal chart titled Review Requests Get Results if Businesses Keep Asking, showing 78 percent were asked and 22 percent were not, with the 78 split into always left a review 28 percent, left a review more than half the time 20 percent, less than half the time 17 percent, and did not write a review 13 percent
Both the 65% and the 83% are correct; they are shares of different groups. Chart: BrightLocal, Local Consumer Review Survey 2026, reproduced under the terms on that page.

Same-day replies tripled, and so did indifference

Nineteen percent now want an answer on the day they post, against six a year ago. The tidy reading is that everyone grew impatient. The chart shows something less tidy: the two-to-three-day and within-a-week groups both shrank, and the group expecting no answer whatsoever grew as well.

Grouped bar chart titled How fast consumers expect a reply to their review, comparing 2026 and 2025: same day 19 versus 6 percent, following day 13 versus 12, two to three days 31 versus 38, within a week 18 versus 25, two weeks 4 versus 5, a month 1 versus 2, whenever as long as they reply 3 versus 5, would not expect a reply 11 versus 7
Single choice. The middle of the distribution is emptying in both directions.

Eighty-nine percent expect a reply of some kind, and 81% want it inside a week. What the response chart adds is that replying selectively barely helps: answering only positive reviews scores 45% and only negative ones 47%, against 80% for answering everything.

BrightLocal chart titled Each Review Needs a Response, but Generic Replies Do More Harm Than Good, showing likely, not likely and no impact shares for responding to all reviews, only to negative, only to positive, with a templated answer, and not at all
Templated replies put half of respondents off, a worse score than not replying at all. Chart: BrightLocal, Local Consumer Review Survey 2026, reproduced under the terms on that page.

A generic answer makes 50% unlikely to use the business, while silence makes 42% unlikely, so a copy-pasted thank you tests worse than saying nothing. And 31% remain likely to use a business that never replies at all.

After a good review, most people go looking for your website

A positive review rarely closes anything by itself. Two thirds of readers carry on researching and a third move towards a purchase. The most common next step is opening the company’s own site, picked by 54%, up from 32% when the question was last put in 2019.

Beyond the website, 37% read more reviews on the same platform, another 37% go to a different platform, and 24% check social channels. Meanwhile 22% simply keep looking at other businesses, the cost of a weak next step.

That is where the funnel breaks for small firms. BrightLocal’s own SMB research puts the share of small businesses with a dedicated website at 40%, so for a majority the most popular action after a good review leads nowhere useful.

The chart’s purchase column is labelled “visit the business’s website” at 31%, while the surrounding text calls the same figure visiting the business location. One of the two is a slip, so we would not build a plan on that number.

Ninety-three percent have bought something after reading a positive review, 27% have spent over $1,000 and 13% over $5,000. Seventy percent have also regretted a purchase made after reading reviews.

What consumers want done about fake reviews, and what the law already does

Consumers put the burden on platforms first: 65% say review sites should catch fakes and 49% say the business receiving them should. The punishments they favour run from a platform ban all the way to prison. Regulators, meanwhile, have already acted, and in Europe further than these respondents assume.

BrightLocal bar chart titled Consumers Call for Fake Review Punishments from Bans to Jail Time: banned from review platforms 57 percent, losing customers 54, removal from Google search results 46, fines 37, losing their business license 20, criminal charges or prison 16, no consequences 2
Two percent of respondents want no consequences at all. Chart: BrightLocal, Local Consumer Review Survey 2026, reproduced under the terms on that page.

There are two more discrepancies here. The text gives 63% for platform responsibility where the chart shows 65%, and it says 3% want no consequences, which would make the headline 97%; the chart shows 2%, so the correct headline is 98%.

For a European business the legal position is stricter than platform policy, and it is what actually governs the practices this survey describes.

Chart titled What the law already says about reviews, market by market, listing the EU Omnibus Directive 2019/2161, Poland from 1 January 2023, UOKiK penalties against sellers of reviews, the UK DMCC Act 2024 from 6 April 2025, the US FTC rule 16 CFR Part 465 from 21 October 2024, and Google's policy in every market
The survey covers the United States. In the EU the binding rules are the local ones.

Enforcement is not theoretical: the European Commission’s coordinated sweep of 223 sites found at least 55% potentially in breach of EU consumer law, with 144 unable to demonstrate that they check review authenticity. In Poland, UOKiK has issued decisions against companies trading in fabricated opinions, including fake comments posted to Google Maps.

Platforms remove at scale too. Google blocked or removed over 240 million policy-violating reviews in 2024, along with 12 million fake Business Profiles, while Trustpilot removed 4.5 million fake reviews, 7.4% of everything submitted to it that year.

Only the discount fell out of review incentives

Incentives did not go away, they narrowed. The share offered a discount in exchange for a review dropped nine points to 27%, but loyalty points rose slightly, gifts held, and cash offers stayed flat at 20%. Eleven percent were offered something specifically for leaving positive feedback.

The share offered nothing at all moved from 36% to 41%, so the practice is receding slowly rather than collapsing. The steady 20% for cash is the awkward one, because Google prohibits reviews paid for in cash or in kind and can stop a profile receiving new reviews.

In the EU the exposure is larger than a policy strike. Paying for a positive review, or commissioning reviews at all, falls under the blacklist of unfair commercial practices, where national penalties reach up to 10% of turnover.

Which year-on-year moves are large enough to trust

A poll of a thousand people carries an error bar, and comparing two such polls carries a wider one. At these sample sizes a gap between the two waves needs to reach roughly 4.4 points before it can be separated from sampling noise. Several of this year’s talking points do not get there.

Chart titled Which year-on-year moves clear the survey's own error bar, listing fourteen measures with their 2025 and 2026 values and the change in percentage points, with eight highlighted as clearing the threshold and six shown as inside it
Our calculation: the two samples are 1,002 and 1,026 respondents, giving a threshold of about 4.4 points.

Comfortably above the line: local news at nineteen points, the 4.5-star requirement at fourteen, Apple Maps and same-day replies at thirteen each, Google at twelve, and the willingness to write when asked, also twelve.

Below it, and better described as unchanged: the fall in reviews sought from the past month, the drift in the three-month window, the rise in people who would never write a review, and the four-point dip in positive reviews written. The last of those is the one most likely to be over-read, our own earlier point about shrinking supply included.

What this changes for a small business in the EU

Very little transfers directly, because the survey measures American consumers living under American law. The direction does transfer: a steady flow of recent reviews beats a stockpile, replies need to be quick and specific rather than templated, and whatever sits behind the review has to work when someone arrives.

Keep invitations continuous instead of seasonal, so the newest review is always days rather than months old. Reply to everything within a day, in your own words, and rotate wording rather than pasting one formula. Make sure the website, opening hours and contact route survive the visit that 54% will make. And treat cash, discounts and any request for a specifically positive review as off limits, since in the EU that is not a policy question but a legal one.

For the operational side of all this, our guide to running Google reviews covers requests, replies and problem cases; the profile checklist covers the fields customers rely on, and the Google Maps SEO guide covers where reviews sit among ranking signals.

None of the above was measured in Europe. If your market is Poland, Ukraine or anywhere else in the EU, this survey is a hypothesis to test against your own review data, not a set of benchmarks to adopt.

Sources