I just noticed a local restaurant has a score of 3.9 with 16 reviews, but if you click in to read the reviews there are actually 19 of them. The average is actually 3.47, so it seems they did not count three of the 1-star reviews towards the rating.

My first though was that there might be newer reviews that didn't get counted yet, but this is not the case. The newest review is 3 months old, and after I added a review the count went up to 17 but the rating didn't update.

How does Google decide which ones to not include in the average rating, if this is actually what they are doing?


If I sort the reviews from oldest to newest, they now show only 16 reviews. I also included a running average:

Oldest  1   1.00
        2   1.50
        5   2.67
        5   3.25
        5   3.60
        4   3.67
        2   3.43
        5   3.63
        5   3.78
        1   3.50
        5   3.64
        4   3.67
        1   3.46
        5   3.57
        5   3.67
Newest  4   3.69

As you can see, at no point did the average ever get up to 3.9, so the idea that the average calculation is stale is out. Perhaps it has something to do with the missing 3 reviews (went from 19 to 16)? But the real average went UP since it was showing 19 reviews, which means the missing 3 were below the average and can't explain the 3.9 score either.

  • Please don't think that programmers are working so hard as to have Google look at the reviews and make decisions. Seriously. Google is compromised of several parts. Search is just one. The knowledge graph is another. If you are seeing a knowledge graph card on the right, then your instinct regarding updating is correct. No search engine is real-time or even close. The knowledge graph and search are busy places. It is extremely likely that the knowledge graph has not updated the entry for the business yet. This may be a periodic batch style process. Who knows? – closetnoc Aug 27 '16 at 4:21
  • @closetnoc but it still doesn't add up. The most recent reviews are 4, 5, and 5. In order to average to the 3.9 rating, you have to keep those and remove three of the 1-star ratings. So this is not a stale-data issue. And I never said anything about programmers... – Sarke Aug 27 '16 at 4:43
  • Also, this doesn't concern the search at all. – Sarke Aug 27 '16 at 4:45
  • Are you talking about the knowledge graph card on the right of the search results? If not, then you may want to clarify where specifically you are the discrepancy. That will help. I assumed you were talking about the knowledge graph card. It is still possible the discrepancy is related. – closetnoc Aug 27 '16 at 5:01
  • I'm talking about Google Places. This info is show in both the knowledge graph and on Google Maps. This is not related to search. – Sarke Aug 27 '16 at 5:14

It's no big secret that Google loves to use algorithms and crazy hard mathematical equations for its return of search results, in both organic and local results. This is no different when Google computes its review average.

MEAN (Average)

The issue here isn't the fact that Google is returning an incorrect review average, the problem is that you are assuming they are using average calculator based on a simple formula called 'MEAN'.


To calculate the Mean

Add the numbers together and divide by the number of numbers. (The sum of values divided by the number of values).

To determine the Median

Arrange the numbers in order, find the middle number. (The middle value when the values are ranked).

To determine the Mode

Count how many times each value occurs the highest is the mode. (The most frequently occurring value)

Bayesian Average, or Similar

Google is using a far more complex formula to work out your average, and many people believe its a modified Bayesian average formula that predicts the probability of next few reviews, a simpler way of getting your head around this is to think, Google is predicting a negative review, It's unrealistic to assume that your business will always be 5.0, 4.6 and so on.

A business could 5 reviews with a average rating of 5.0, while another business may have the exact same number of reviews, and the exact score from visitors, but have a completely different average rating, this is because Google is using Bayesian average or something similar, or a modified Bayesian formula. Google for example does, can, or could take into other factors such as time, date, frequency, volume and even the niche of the business.

Google is far from being transparent when it comes to disclosing algorithm information, so sometimes you just need to accept it, or attempt to understand it the best you can.

It's extremely unlikely that someone is going to pick a 4.7 over a 4.5, Google knows this... People are for more likely consider the amount of reviews a better indicator. The only thing you can do about it is keep up the good work! everyone else is in the same boat.

You will find more information about Bayesian and Google Local Reviews on Google.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.