TLDR
- A bad review is normal. In booteek's analysis of Google star breakdowns at 838 UK restaurants and bars (measured September 2026), 95% had at least one 1-star review.
- The money moves with your average rating. A Harvard Business School working paper on Seattle restaurants (Yelp, 2003 to 2009) found a one-star rise went with a 5 to 9% rise in revenue, driven by independents.
- booteek's recommendation is to reply to every negative review, because the next guest reads your reply. Then use the table below to work out your own exposure from your own covers and average spend.
Do negative reviews really cost a restaurant or bar money?
The best evidence says yes, and it says the cost runs through your average star rating. A Harvard Business School working paper by Michael Luca studied Seattle restaurants from 2003 to 2009 and found that a one-star rise in a restaurant's Yelp rating went with a 5 to 9% rise in revenue, and that the effect was driven by independent restaurants, not chains.
The second study measured tables. Michael Anderson and Jeremy Magruder, economists at UC Berkeley, looked at San Francisco restaurants in 2010. An extra half-star on Yelp made a restaurant's 7 pm tables 19 percentage points more likely to be fully booked (*The Economic Journal*, 2012).
Picture what that means on a Friday. Two couples on the same street are choosing where to eat and drink at half past six. Both open Google Maps and see your venue beside the one opposite, with a star average next to each name before any review is read.
That is the moment a star average earns or loses a table. So the question owners ask is a fair one: what is a bad review actually costing me? Neither study answers it for a UK bar on Google. Here is exactly what each one measured.
| Luca (HBS working paper, revised 2016) | Anderson & Magruder (2012) | |
|---|---|---|
| Place | Seattle, USA | San Francisco, USA |
| Platform | Yelp | Yelp |
| Data from | 2003 to 2009 | July to October 2010 |
| Venues | Restaurants | Restaurants on one online booking site |
| What it measured | Quarterly revenue | Whether 7 pm tables for four sold out |
| The finding | One star up, revenue up 5 to 9% | Half a star up, 19 percentage points more sell-outs |
| Who it applied to | Independents (chains showed no effect) | Restaurants with online booking |
| What it does not cover | Bars, Google, the UK, owner replies | Revenue, bars, Google, the UK, owner replies |
Two cautions travel with those findings. Both studies measure a change in the average rating, so neither tells you what one review costs. And both are about Yelp in the United States more than a decade ago. They show the direction and the scale of the effect. They do not give you a UK figure, and we found no published one.
How common are 1-star reviews for UK restaurants and bars?
Almost universal in our data. Across 838 UK restaurants and bars whose full Google star breakdown we hold (measured September 2026), 95% had at least one 1-star review — and so did 84% of those averaging 4.7 stars or more. One- and two-star reviews were a median 5% of a venue's Google reviews.
Put plainly, the typical venue in that study had about one review in twenty at one or two stars, and the best-rated venues were barely less exposed than the rest.
The count comes from the star breakdown Google publishes for each profile, so it covers every review on the profile and carries no bias from which reviews we happened to capture. Its bias is in the venues. The 838 are the ones whose breakdown we hold, 565 of them in Manchester, and they run slightly more established than the other UK venues we hold. The sample is selected, not random, and it is a single snapshot taken between March and September 2026.
It also says nothing about money. It is the exposure half of this article. The revenue half rests on the two studies above, exactly as scoped there.
How much does one bad review move your star rating?
Very little on its own. For illustration, on a profile with 400 reviews, one new 1-star review lowers the average by less than a hundredth of a star, whatever the average was before. That is arithmetic, and it holds for any rating.
A cluster does far more. For illustration again, 20 new 1-star reviews on the same 400-review profile, averaging four and a half stars, would take about a sixth of a star off. A bad month after a kitchen change or a short-staffed summer can do that.
The average matters twice. It is the first thing a guest reads, and Google's own local ranking help page says "more reviews and positive ratings can help your business's local ranking". Google says local results are mainly based on relevance, distance and popularity, and it files reviews under prominence, one of the things that decide which venues it shows at all.
So the evidence points to two costs, both through the average: fewer guests may choose you from the list, and the list may show you less often. One review barely moves either. A run of them can move both, and a reply cannot undo that: it changes what the next reader sees, not the average or the ranking.
Do unanswered negative reviews hurt more than answered ones?
We found no measurement of that for UK restaurants and bars, and neither study above looked at owner replies. What is measured is how many people read reviews before they choose.
In BrightLocal's 2026 Local Consumer Review Survey of 1,002 US adults, 97% said they read online reviews of local businesses and 41% said they always do. That is a US sample, about local businesses in general, and it records what people say they would do. It is a US figure, and we found no UK equivalent.
Google's own guidance on replies is modest and worth quoting exactly. "When you reply to customer reviews, it shows that you value their feedback. Positive reviews and helpful replies can help your business stand out." Google does not say a reply changes your ranking, and we won't claim it does.
What a reply changes is the page. An unanswered 1-star review is the last word under your name for every guest who scrolls that far. In booteek's view, a reply that names the problem, owns it and says what changed turns the same review into evidence that someone is running the place.
A reply also cannot lift your average. Only new reviews do that, and new reviews come from guests who had a good night and were asked. For the other side of this, how quickly UK owners actually reply and what booteek's reply-time study found, read how fast restaurants and bars reply to Google reviews.
How can you work out what bad reviews cost your own venue?
Multiply the covers you think a weaker rating costs you each week by your average spend, then multiply by 52. Every number in the right-hand column below is invented for illustration. Replace them with your own.
| Step | Where to find it | For illustration only |
|---|---|---|
| 1. Covers in a typical week | Your till or booking system | 300 |
| 2. Average spend per cover, in pounds | Weekly takings divided by covers | 25 |
| 3. Covers you believe you lose each week | Your own estimate (see below) | 6 |
| 4. Weekly cost, in pounds | Step 3 multiplied by step 2 | 150 |
| 5. Yearly cost, in pounds | Step 4 multiplied by 52 | 7,800 |
A bar can run the same sum on drinks or rounds per head instead of covers. The method holds either way.
Step 3 is the hard one, and it has to be your own assumption. No study gives a UK figure for Google, and the Seattle range was measured on revenue, per full star, on Yelp. Two ways to put a number on it:
- Compare covers or bookings for the four weeks after a run of bad reviews with the same four weeks last year. A clear gap, with nothing else changed, is your first estimate.
- Watch your busiest booking slot. Anderson and Magruder measured whether 7 pm tables sold out, and you can check whether yours still do.
Then run the sum twice, at a low and a high estimate for step 3. In the illustration, six lost covers a week at an average spend of 25 pounds comes to a little under eight thousand pounds a year. The range is the honest answer, and it tells you how much time a problem deserves.
What should you do about a negative review this week?
Reply to it within a window you can keep, and report it only if it breaks the rules. Both cost nothing but time. Here are booteek's recommended steps, in the order that protects the average.
- Reply to every unanswered 1- and 2-star review, newest first. Write for the next guest reading the page. Thank the reviewer, name the specific problem they named, own your part and say what has changed. Our guide to responding to 1-star reviews has a "say this, never say this" table.
- Sign it with your name. A named owner reads as a person who will be there next Friday. A reply from "the team" reads as a template.
- Take the detail offline. Give an email address or a phone number, and settle the rest in private. The public reply only has to show that you listened and acted.
- Report only what breaks the rules. You can't delete a customer's review. You can report one, and Google removes it only if it breaks policy. The routes, the waiting times and the UK law are in our fake review action plan.
- Set a reply window and keep it. Our recommendation is to pick a window your rota can hold every week, then treat a missed one like a missed delivery.
- Ask every guest for a review the same way. You can ask every customer for a review. You can't ask only the happy ones. A card with the bill, or one line on the receipt, is enough.
Step 6 is the only one you control that lifts the average. The others stop a bad review doing more damage than it has to.
What will not protect your star rating?
Nothing that buys or filters reviews. Those moves break Google's rules, and UK law now bans fake reviews outright.
Since 6 April 2025, UK law has banned fake reviews outright. Writing, buying or commissioning a fake review is banned, and so is hiding the bad reviews while you show the good ones. Google's review content policy bans offering anything of value for a review, down to a discount or a free drink, and bans offering it for taking a bad one down.
That rules out the moves that tempt owners after a bad month. A free round in exchange for a 5-star review is an incentive. Sending the review link only to the tables that smiled is selective asking. A "removal service" that promises to make 1-star reviews vanish is selling a result it cannot deliver, because Google removes a review only when it breaks policy, whoever asks.
Flooding a profile with friends' 5-star reviews carries a further risk. Google can stop a profile receiving new reviews and show a warning to readers when it removes fake ones. That costs far more than the review you were trying to bury.
What does booteek do about unanswered negative reviews?
booteek is the AI agent that works on your Google Business Profile with you, and the reply is where it helps with bad reviews. On the review in front of you, on Google Business Profile and other review platforms, the booteek Chrome extension drafts a reply. You read it, edit it against the steps above, copy it and post it yourself.
It does not post anything for you, and the table above is still yours to fill in. Plans and what each includes are on the pricing page.
The next step is a short one. Open your Google Business Profile, answer the newest 1- or 2-star review that has no reply, then fill in the first two rows of the table with last month's covers and takings.
Where do these figures come from?
- booteek star-breakdown study (booteek, measured September 2026) — the 95%, 84% and median 5% figures. The sample is 838 UK restaurants and bars whose full Google star breakdown we hold, Manchester-heavy (565 of them in Manchester). The venues are selected, not random, and run slightly more established than the other UK venues we hold. The data is Google's own star breakdown for each whole profile, one snapshot per venue captured between March and September 2026, so the counts are a slight floor. It does not show any effect on revenue, and it is not a trend.
- Luca, "Reviews, Reputation, and Revenue: The Case of Yelp.com" (Harvard Business School working paper 12-016, revised 2016) — the 5 to 9% rise in revenue for a one-star rise in Yelp rating, driven by independent restaurants. Seattle restaurants on Yelp, 2003–2009. A working paper, not peer-reviewed. Read the paper (PDF)
- Anderson and Magruder, "Learning from the Crowd" (*The Economic Journal*, 2012) — the 19 percentage points more sell-outs of 7 pm tables for an extra half-star. San Francisco restaurants on one online booking site, Yelp, July to October 2010. Read the authors' manuscript (PDF)
- Local Consumer Review Survey 2026 (BrightLocal, February 2026) — 97% read online reviews and 41% always do. 1,002 US adults, about local businesses in general. It records what people say they would do. Read the survey
- Tips to improve your local ranking on Google (Google Business Profile Help, read 28 September 2026) — that local results are mainly based on relevance, distance and popularity, that reviews sit under prominence, that more reviews and positive ratings can help local ranking, and the quoted guidance on replies. Read Google's page
- Prohibited and restricted content (Google Maps content policy, read 28 September 2026) — the ban on offering anything of value for a review, or for taking a bad one down. Read Google's policy
- Report inappropriate reviews on your Business Profile (Google Business Profile Help, read 28 September 2026) — that you can report a review and Google removes it only if it breaks policy. Read Google's page
- Business Profile restrictions for policy violations (Google Business Profile Help, read 28 September 2026) — that Google can stop a profile receiving new reviews and show a warning to readers. Read Google's page
- Digital Markets, Competition and Consumers Act 2024, Schedule 20 (UK Parliament, legislation.gov.uk) — the ban on businesses that write, buy or commission fake reviews, or hide the bad reviews while showing the good ones, in force since 6 April 2025. Read the law
- Fake reviews guidance, CMA208 (Competition and Markets Authority, April 2025) — that you can ask every customer for a review, and that asking only the happy ones is banned. Read the guidance
- The cost table — illustration only, with no source. Every number in its right-hand column is invented to show the method.