/ Technology

The CMA: can you trust what you read in online reviews?

magnifying glass on laptop

The Competition and Markets Authority today published a report on online reviews and endorsements. Here’s the CMA’s Senior Director, Nisha Arora, reflecting on the debate their report has already started to generate.

The most important message from our report is that reviews and endorsements are really valuable to consumers.

That said, there’s work to be done if we want to preserve consumers’ trust. We’ve received information about fake reviews, genuine negative reviews not being published and a lack of clear disclosure of paid endorsements in blogs, vlogs and other online publications.

That’s why we’ve opened an investigation into a number of companies in connection with potential non-disclosure of paid endorsements. And we’re looking into other concerns that have been raised to determine whether we should be taking further action.

Reaction to our report

When I did the rounds on TV and radio this morning the same important questions came up repeatedly:

  • Can we believe what we’re reading in reviews and endorsements? Given the concerns we’ve identified, I can’t vouch for every review, blog and tweet. Nevertheless, the outcomes for users speak for themselves. Eight out of 10 people who use online reviews and blogs find that the thing they buy matches their expectations.
  • What can be done to address your concerns? While the CMA will play an active role, the main responsibility lies with businesses in the sector. It’s critical for them to maintain your trust and ensure that you continue to get good outcomes when they use these tools. We’ve tried to help by producing information for businesses explaining what they need to do to comply with the law.
  • What do you suggest people should do? There’s no sure fire way for you to tell if a review is real or fake – or indeed to be certain that the site that they are looking at is publishing all reviews. However, our research on the way that people use reviews and endorsements suggests that some of you are already pretty savvy about using online reviews and endorsements. For instance, testing what they read with other information sources including reviews, blogs and friends and family.

Online reviews can provide valuable information and it’s important that your trust in them is maintained. ​That’s why we’re continuing our work in these sectors.

What’s your experience of online reviews? Have you been suspicious of a review that’s been published? Have you ever had problems posting your own user reviews?

Which? Conversation provides guest spots to external contributors. This is from Nisha Arora, Senior Director Consumer at the Competition and Markets Authority. All opinions expressed here are her own, not necessarily those of Which?.


I listened to Nisha on Radio 4 this morning and was distinctly unimpressed with the CMA. When asked what hard evidence the CMA had of fake reviews, she had no information to give. I am not suggesting there are no fake reviews, and I am strongly against faking “independent” endorsements, but I would have liked a more prepared response to support the CMA’s assertions.

Looking quickly through the report there seems to be “reports of” and maybes, but no hard evidence. Just the suspicions of some (unquantified – 1-2%, maybe higher) malpractice that perpetrators should address while CMA continues to investigate.

Perhaps it was all a bit early to go for publicity? Perhaps I am ill-informed and being unfair.

Oh, and please, can we try to ban people on the airwaves from beginning sentences in reply to questions with “So….”. Perhaps that put me in an unsympathetic frame of mind..


I always ignore reviews if there only a few (say less than 10) reviewing a product I am interested in. If there are hundreds then I am more likely to believe them. I’m sure fake reviews are everywhere.

I’ve posted negative reviews on Argos only to have them removed.


I was wondering where this issue figured in the CMA’s work schedule of the top ten important consumer concerns.

My opinion on reviews has always been that they are worth what you pay for them.


I look for reviews that contain detailed information and explain some of the disadvantages of a product as well as making positive comments.

Some people take a great deal of effort to provide us with useful information that retailers omit.


Yes, I agree with that but it has been widely alleged that in the hotel sector some of the adverse reviews have been posted by competitors.


I was thinking mainly about household goods, John. You are right about hotels.


Just looking at the restaurant and cafe reviews on TripAdviser for my local touristy town, they all get 4.5-5 stars with the most recently opened or revamped ones getting the highest scores due to a burst of recent reviews from locals.
Considering the wide variety of establishments covered absolutely of no use at all for making a decision,


Unfortunately Which? is not guiltless in the reviews stakes.

Before I explain that I should mention that I have been following this area for several years collecting examples of these deceptions from computer and consumer sites. I have highlighted a potential case on Which? reader reviews some months back. I have also suggested to Which? how to make their reader review system more robust and reliable.

Which? carries out product testing and subscribers can view the results and if subscribers are owners they can add their own review. This is very important as Which? does not employ test labs for longevity/durability trials on appliances.

Anyone browsing the Web can read these Readers Reviews. Unfortunately , and particularly where it concerns Best Buys, Which? is not responding to Readers Reviews. In what might be described as the star exhibit we have a product where over 20 people out of 22 with valid scoring give it the lowest rating of one star. The comment now is this should be a Do Not Buy not a Best Buy.

Given the offending item is around £20 you would think an investigation would not be too difficult to arrange and given the two common faults expessed seem design faults and the average time to fault around 6-12 months in normal usage not too difficult to arrange.

Which?’s unique selling point has been independent testing but it must back this up with taking the input from readers seriously so its reputation is enhanced not weakened.


I would expect that the reviews of products written by Which? members are more trustworthy and useful than those published on retailers’ websites but in many cases there are few comments on products, whereas Amazon may have numerous reviews of a product.

I agree that Which? should look at information posted on its own websites. I was very disappointed that the recent article about LED bulbs in Which? magazine made no mention of the problems of radio interference and premature failure mentioned by so many contributors to Which? Conversation.