Online review statistics you can cite: traced to the source
Online review statistics you can cite: figures traced to the survey or rule behind each, with year, sample and link, plus the repeated ones we could not trace.
Written by OpenRated Team You need a number. The deck is due, or the blog post, or the pitch to the owner who still thinks reviews are a vanity metric, and you want to write "97% of consumers read reviews" with a source you could defend if someone asked. So you search for online review statistics and get pages of percentages, each credited to a company name, none to a survey with a year and a sample. Cite one of those and you are repeating a roundup of a roundup.
This post holds only the figures we traced to the page that published them, with who was asked, how many, and when. There are thirteen, from three sources, and a short section on the repeated figures we could not trace, so you know which ones to leave out of your deck.
Online review statistics that trace to a survey or a rule
Three sources passed the test. BrightLocal's Local Consumer Review Survey 2026 asked 1,002 US adults through SurveyMonkey and was published on 11 February 2026; BrightLocal sells local SEO and review software, so it is a vendor-funded survey, which matters for the questions it chose to ask and not for the arithmetic. Pew Research Center's report on online reviews, published 19 December 2016, draws on a national survey of 4,787 US adults conducted from 24 November to 21 December 2015, and Pew has no commercial interest in the answer; it is the oldest figure here and still the most independent. The Federal Trade Commission's press release of 14 August 2024 announced its rule on fake reviews, which gives you the regulator's position in its own words.
The table is the citable summary. Each section after it says what the figure does and does not tell you.
| Figure | Who was asked | Source |
|---|---|---|
| 97% read reviews for local businesses | 1,002 US adults | BrightLocal, 2026 |
| 41% always read reviews when searching for a business (29% the year before) | 1,002 US adults | BrightLocal, 2026 |
| 85% say positive reviews make them more likely to use a business | 1,002 US adults | BrightLocal, 2026 |
| 49% trust online reviews as much as personal recommendations | 1,002 US adults | BrightLocal, 2026 |
| 31% only use businesses rated 4.5 stars or higher (17% the year before) | 1,002 US adults | BrightLocal, 2026 |
| 47% will not use a business with fewer than 20 reviews | 1,002 US adults | BrightLocal, 2026 |
| 74% look for reviews written in the last three months | 1,002 US adults | BrightLocal, 2026 |
| 69% wrote a review in the past 12 months | 1,002 US adults | BrightLocal, 2026 |
| 78% were asked to write a review; 83% of those asked wrote one | 1,002 US adults | BrightLocal, 2026 |
| 97% say businesses should face punishment for fake reviews | 1,002 US adults | BrightLocal, 2026 |
| 82% at least sometimes read ratings or reviews before a first purchase; 40% always or almost always | 4,787 US adults, surveyed late 2015 | Pew Research Center, published 2016 |
| 48% of review readers say it is often hard to tell if reviews are truthful and unbiased | the review readers among 4,787 US adults, surveyed late 2015 | Pew Research Center, published 2016 |
| Around one in ten nearly always post their own reviews | 4,787 US adults, surveyed late 2015 | Pew Research Center, published 2016 |
The FTC rule is not a percentage, so it sits in its own section below.
Do customers read online reviews before they buy?
Yes, by every survey here, and the two sources agree more than their ten-year gap suggests. BrightLocal's 2026 edition has 97% of its 1,002 US adults reading reviews for local businesses, and 41% saying they always do when searching for a business, up from 29% in the previous edition. Pew's survey, run in late 2015 and published in 2016, asked a different question, about buying something for the first time, and found 82% of US adults read ratings or reviews at least sometimes, with 40% doing so always or almost always.
The two "always" figures, 41% and 40%, are the ones to cite for a habit. The 97% is a ceiling: it counts anyone who reads a review ever, which is nearly everyone with a phone. If you put it on a slide, say "read reviews for local businesses", because that is the question asked. "97% of consumers read reviews" with no object is the version that spreads and the version nobody can trace.
Pew's age split is useful when your customer is older. In the 2015 fieldwork, 53% of 18 to 29 year olds always or almost always read reviews before a first purchase, against 23% of those 65 and over. Those are self-reported habits, and they are a decade old, so treat the gap as real and the exact numbers as dated.
What customers do with the reviews they read
This is the group of figures a business can act on, and the one where the surveys moved most between editions.
Positive reviews still work: 85% of BrightLocal's respondents say positive reviews make them more likely to use a business. Trust is lower than the roundups suggest: 49% say they trust online reviews as much as personal recommendations, which means a slight majority do not. Pew's readers in 2015 were split the same way, with 48% saying it is often hard to tell if online reviews are truthful and unbiased.
The thresholds are the figures to take to your team. In the 2026 edition, 31% only use businesses rated 4.5 stars or higher, up from 17% a year earlier; 47% will not use a business with fewer than 20 reviews; and 74% look for reviews written in the last three months. Put those three together and the target for a business is a rating above 4.5, more than 20 reviews, and a few new ones every month. How many you need in your own category, and how to get there, is a post of its own: how many reviews a business needs.
What these figures will not do is prove that reviews caused a sale. Every one is a person saying what they would do. The observed figures, where a research team watched conversion change as reviews appeared, come from the Spiegel Research Center and are in our post on why reviews matter to a business; cite those when the question is revenue, and these when the question is behaviour.
How many customers leave a review, and does asking work?
More than the folklore says. In BrightLocal's 2026 edition, 69% of respondents wrote a review in the past 12 months. Pew's 2015 figure measures a stronger habit: around one in ten Americans nearly always post their own reviews of products, services and restaurants. The two are not in conflict; one counts anyone who wrote a review in a year, the other counts people who do it almost every time.
Asking is the lever. 78% of BrightLocal's respondents had been asked to write a review, and 83% of those asked wrote one. 28% say they will always write a review if asked. For a business that has never asked, those are the figures that justify the hour it takes to set up a request: the steps are in how to get more reviews, and the tools that send requests by SMS and email, watch several sites from one inbox and flag suspect reviews are compared on our reputation and review management listing.
Two limits. A survey of US adults says nothing about your customers in particular; a plumber's customers and a SaaS buyer's procurement team are not the same population, and neither is 1,002 people on SurveyMonkey. And asking does not fix a product people dislike. A request sent to an unhappy customer produces an honest one-star review, which is the system working.
How common are fake reviews, and what do the rules say?
Two questions people add to this search are "how common are fake reviews" and "how many reviews are fake". No figure in this post answers either, because we traced no independent measurement of it. What exists is of two kinds: platforms publishing how many reviews they removed, which is self-reported and counts what their own filters caught; and surveys asking consumers whether they believe they have seen a fake, which measures suspicion. A figure like "x% of reviews are fake" with no method behind it is the one to leave out of your deck.
What a survey can measure is the appetite for enforcement. In BrightLocal's 2026 edition, 97% of respondents say businesses should face punishment for fake reviews. On the regulator's side, the FTC announced its final rule on fake reviews and testimonials on 14 August 2024, by a 5-0 vote, effective 60 days after publication in the Federal Register. The FTC's release lists what the rule prohibits:
- Fake or false consumer reviews, consumer testimonials and celebrity testimonials, including AI-generated fake reviews and reviews that misrepresent the reviewer's experience.
- Buying positive or negative reviews, where payment is conditioned on the sentiment of the review.
- Insider reviews and testimonials without the required disclosure of a material connection.
- Company-controlled review websites presented as independent.
- Review suppression through unfounded legal threats, intimidation or false accusations.
- Misuse of fake social media indicators, such as bot-generated followers.
The release says the rule lets the FTC seek civil penalties against knowing violators. Two things to hold onto: "buying reviews" is defined by payment conditioned on sentiment, so an incentive for any honest review is a different question, which is it legal to pay for reviews works through; and this is a United States rule as the FTC published it, not legal advice, and other countries have their own rules that this post does not cover.
Repeated figures we could not trace
Statistics roundups repeat figures that fall into patterns, and the patterns are more useful to you than the individual numbers, because they tell you what to check on any statistics page.
A percentage credited to a company name. "According to" a software vendor or a consultancy, with no survey, no year and no sample. The company may have run a survey or may have written its own roundup; the page does not say, and a reader cannot tell without doing the tracing the page should have done.
A survey figure with no edition. One page credits a figure to "experts from Bright Local". BrightLocal publishes its survey every year and the figures move: the share who only use businesses rated 4.5 or higher went from 17% to 31% between the last two editions. An undated figure from an annual survey can be several editions old and still be presented as this year's.
A headline figure we could not trace. One page opens with 98% of shoppers reading reviews before buying and a 35% revenue lift from responding to a quarter of reviews; neither has been traced to a study here, so neither is in the table above, and that is the test: a figure you cannot follow to its source is a figure you cannot defend.
A platform's own enforcement count. The share of fake reviews a platform says it caught before publication is a number the platform produced about itself. It is worth reading, and it is not an independent measure of how common fake reviews are.
How to read a review statistic before it goes in your deck
Six checks, in the order that rules a figure out fastest.
- Who ran it, and who paid. A vendor's survey is not wrong, and the questions it asks are the ones its product answers. Say "a survey by a review software vendor" when you cite one.
- Who was asked, and how many. 1,002 US adults is a fine sample and still not "consumers worldwide". A figure with no sample has no population, and you cannot say whom it describes.
- Self-reported or observed. "Say they read reviews" is a stated habit; conversion measured on a retail site is behaviour. Both are useful. They answer different questions, and the observed one is rarer.
- Which edition, which year. For an annual survey, name the edition. If the number moved a lot between editions, say so, because next year's reader will check.
- A share of whom. "83% wrote a review when asked" is a share of the people who were asked, not of everyone. Read the denominator before you copy the numerator.
- Does the link reach the survey. If the citation lands on another roundup, you have not reached the source yet. Keep clicking or drop the figure.
What to do with these figures
Cite the table, with the source and the year, and your deck will survive the one person in the room who checks. Then act on the three thresholds the 2026 survey sets, a 4.5 rating, 20 reviews and something recent, by asking the customers who are already happy: that is the whole of how to get more reviews, and it is the next thing for your business to do this week.
