Eagle eyed Phil Rozek of Local Visibility pointed out this example of a business with 4 reviews that is showing the new star treatment in the main search result for jewelry boston:

I thought that odd as previously it had seemed that 5 reviews was the limit to get the star treatment. Well it is odd. It seems that sometimes it is four reviews and sometimes it is five reviews. Go figure.

Here is a local carousel for a search (restaurants ellicottville, ny) where a listing with 4 reviews doesn’t have stars but 5 reviews does:

And another for a search for jewelry orlando which also shows no stars with four reviews and stars with five:

There is some factor that causes Google to consider 4 enough to show the stars. What it is, is not exactly clear. Ideas?

Please consider leaving a comment as your input will help me (& everyone else) better understand and learn about local.

How Many Reviews to Get the Star Treatment? Somewhere Between 4 and 5
by Mike Blumenthal

10 thoughts on “How Many Reviews to Get the Star Treatment? Somewhere Between 4 and 5”

Perhaps “owl-eyed” would have been most accurate, given the crazy nighttime hours I usually keep 🙂

My best guess as to why Beverly Hills Jewelers (from the example you show) has 4 reviews but no stars: one of those “reviews” isn’t a review, but rather only a rating. It’s also true that Goldsmith has 3 reviews and 2 ratings-only, but one of those ratings was posted by a verified Google Plus user. So my hypothesis – until someone turns it into Swiss cheese – is that Google doesn’t count ratings (AKA stars without review text) that weren’t written by verified Plus users. I think everything else counts toward the 4-review threshold for having the stars appear.

@Phil

Hmmm…. sounds pretty complicated. We need more data. The current set of 1 is hard to draw conclusions from. 🙂

“Best guess” doesn’t necessarily mean “good guess” 🙂

I’ve got 5 reviews, and still no stars

And I’ve been trying for weeks to get photos showing on my Google plus business page

I just did a search for plumber and a simple review of the 7 pack shows –

Reviews with no stars-
3 had 3 Google reviews
1 had 4 Google reviews

Reviews with stars-
1 had 6 Google reviews
1 had 7 Google reviews
1 had 19 Google reviews

Using the same keyword (plumber), I checked for results in Google+ Local. I was served a single page that had 399 businesses listed. A few of the listings had bad catergories/keywords (locksmith, flooring, belts and carpet cleaning) .

Most of the addresses were within 15 miles of my address and a few outliers (11) were between 20 – 25 miles away.

Of the 399 businesses listed, 19 had 5 or more reviews and stars showing, 72 had 1 – 4 reviews and no stars showing. The remainder of the listings (308) did not have any reviews.

At least now I know that I need 5… or maybe 4 reviews to actually get them in the first place. I think Phil is onto something there, does it have to be a verified Google user for your rating to actually ‘count’?

I strongly suspect that Google is using a Bayesian average – http://en.wikipedia.org/wiki/Bayesian_average. It might be the right thing to do but it sure is frustrating! A business with one hundred 5 star reviews should (in my view) get a better score than a business with four 5 star reviews. This is one way to make that happen.

I have a website with over 700 reviews on other sites but only 5 on google maps local listing. The 700 reviews come out at 4.8 stars in shopping (that I pay for) but the google maps listing (with four 5 star views and one 4 star) still doesn’t have a score!

The Bayesian score might even get a little more complex if google are taking in to account the mindsets of the individual reviewers. If Bob has reviewed 10 different businesses and given them all 5 stars, then Bob might be an extremely positive person. If Sally has reviewed 10 businesses, 9 of which are 2 star and 1 is 5 star, then perhaps this 5 star review from Sally is effectively “worth more” than Bob’s. Just a thought to make this even more complicated than everyone would probably assume. But does support the fact that the same ten 5 star reviews may have 4.6 one week and 4.5 the next purely because Bob has written yet another positive review.

Could very well be.

What I don’t understand is that I’ve got 5, 5 star reviews and my and it says “4.8 out of 5”. Can you explain?

@joe google both looks across your review corpus across the web and factors those in AND probably does something like a Bayesian average where they use a statistical model to predict your score based on a larger sample.

Perhaps “owl-eyed” would have been most accurate, given the crazy nighttime hours I usually keep 🙂

My best guess as to why Beverly Hills Jewelers (from the example you show) has 4 reviews but no stars: one of those “reviews” isn’t a review, but rather only a rating. It’s also true that Goldsmith has 3 reviews and 2 ratings-only, but one of those ratings was posted by a verified Google Plus user. So my hypothesis – until someone turns it into Swiss cheese – is that Google doesn’t count ratings (AKA stars without review text) that weren’t written by verified Plus users. I think everything else counts toward the 4-review threshold for having the stars appear.

@Phil

Hmmm…. sounds pretty complicated. We need more data. The current set of 1 is hard to draw conclusions from. 🙂

“Best guess” doesn’t necessarily mean “good guess” 🙂

I’ve got 5 reviews, and still no stars

And I’ve been trying for weeks to get photos showing on my Google plus business page

I just did a search for plumber and a simple review of the 7 pack shows –

Reviews with no stars-

3 had 3 Google reviews

1 had 4 Google reviews

Reviews with stars-

1 had 6 Google reviews

1 had 7 Google reviews

1 had 19 Google reviews

(example)

Using the same keyword (plumber), I checked for results in Google+ Local. I was served a single page that had 399 businesses listed. A few of the listings had bad catergories/keywords (locksmith, flooring, belts and carpet cleaning) .

Most of the addresses were within 15 miles of my address and a few outliers (11) were between 20 – 25 miles away.

Of the 399 businesses listed, 19 had 5 or more reviews and stars showing, 72 had 1 – 4 reviews and no stars showing. The remainder of the listings (308) did not have any reviews.

https://plus.google.com/u/0/local/Los%20Angeles%2C%20CA/s/plumber

At least now I know that I need 5… or maybe 4 reviews to actually get them in the first place. I think Phil is onto something there, does it have to be a verified Google user for your rating to actually ‘count’?

I strongly suspect that Google is using a Bayesian average – http://en.wikipedia.org/wiki/Bayesian_average. It might be the right thing to do but it sure is frustrating! A business with one hundred 5 star reviews should (in my view) get a better score than a business with four 5 star reviews. This is one way to make that happen.

I have a website with over 700 reviews on other sites but only 5 on google maps local listing. The 700 reviews come out at 4.8 stars in shopping (that I pay for) but the google maps listing (with four 5 star views and one 4 star) still doesn’t have a score!

The Bayesian score might even get a little more complex if google are taking in to account the mindsets of the individual reviewers. If Bob has reviewed 10 different businesses and given them all 5 stars, then Bob might be an extremely positive person. If Sally has reviewed 10 businesses, 9 of which are 2 star and 1 is 5 star, then perhaps this 5 star review from Sally is effectively “worth more” than Bob’s. Just a thought to make this even more complicated than everyone would probably assume. But does support the fact that the same ten 5 star reviews may have 4.6 one week and 4.5 the next purely because Bob has written yet another positive review.

Could very well be.

What I don’t understand is that I’ve got 5, 5 star reviews and my and it says “4.8 out of 5”. Can you explain?

@joe google both looks across your review corpus across the web and factors those in AND probably does something like a Bayesian average where they use a statistical model to predict your score based on a larger sample.