Why does Yelp filter my reviews?

find those reviews

Congratulations – your business received a glowing review on Yelp! Just as you feel the waves of validation tingling through your bod, though, that positive review disappears…

What happened?

Most likely, your positive review was filtered out by Yelp’s automated recommendation software.  (You can see filtered reviews by clicking on the link below the first page of reviews that says “… other reviews that are not currently recommended”).

Yelp recommends certain reviews based on a number of factors. Yelp’s software looks at quality, reliability and activity – if a user is new, not active or does not post consistently useful reviews, their reviews may not be recommended.

Although your positive review came from an authentic customer, the reviewer probably isn’t a user who is established on the Yelp community. Yelp software will also filter reviews if it detects anomalies in the language. Specifically, Yelp will not recommend a review if it is written like a rant or rave because that type of language can be an indicator of a biased opinion, whether it’s coming from your mom or a disgruntled former employee.

Believe it or not, Yelp’s ability to detect fake reviews is 98% accurate… Just kidding. You shouldn’t believe that. We have no idea.

Yelp’s software also updates on a daily basis, as it collects new information. What was a relevant review last month may be considered a stale review tomorrow and no longer be front and center.

Your positive review may have disappeared because your happy customer deleted her Yelp account. This is rare, but there is a chance that she decided to clean house and slim down her internet presence. This probably has nothing to do with your business, but there is a better chance that your positive review was filtered out.

Appealing to Yelp to bring back a positive review back is probably not going to be fruitful because their software (though not perfect) is still the best way for them to systematically recommend or not recommend reviews without making a human judgement call. So if you find that a positive review on your business’s Yelp page is missing, your best course of action is to keep doing good work.


Use HootSuite to Monitor Reviews

HootSuite announcement

Over the past few months, businesses big and small have asked us, “Does Reputology monitor social media like Twitter and Facebook???”

At the time, our answer was “No”.  And now… Well, our answer is still “No”.  But…

HARK!  Powered by the new Reputology app, HootSuite can track online reviews on all the major review sites, including Google+ Local, Foursquare, Yahoo Local, Citysearch, Yellowpages, Urbanspoon, Opentable and more.  

No need to log in to yet another dashboard.  Now, you can monitor reviews in HootSuite along with other social media.  Key features:

  • Monitor reviews for all major review sites, for multiple storefronts, all from one place.
  • Customize each stream by rating, review site, and location.
  • Track how reviews are followed up with by assigning a “status” to a review.
  • Not only monitor reviews on general business directories.  Also monitor industry specific review sites for restaurants, hotels, auto dealers, health professionals, etc.  (You name it, Reputology can track it).

Showing is better than describing.  Check out the screenshots below.

Screenshot #1: Monitor reviews with all your social media channels in HootSuite.


hootsuite screenshot 1

Screenshot #2: Customize a stream however you want.  Create multiple streams, too.


hootsuite screenshot 2

Getting set up is easy.  Here’s how you can start monitoring your online reviews in HootSuite:

  • Already a HootSuite user?  Use this link to Install Reputology now.

  • If you don’t have HootSuite account, sign up here to install Reputology

Minnesota Nice

Hey y’all.  Some of you may have caught me on Minnesota Public Radio this morning sharing a “shallow” dive analysis on Minnesota restaurant reviews and why they rate higher than the other metros from our regional analysis (see blog post titled We Made It to Regionals!).  At the time we  wrote the regional analysis, we didn’t have any good hypotheses for why Minnesota restaurants came out on top.  But after more snooping around, we found some evidence that points to a little bit of “Minnesota Nice” influencing the reviews.  (For those of you who are not familiar with the phrase, it describes stereotypical behavior of native Minnesotans to be courteous and mild-mannered).  Here’s the “shallow” dive analysis shared on the air.

Passive Aggressive Behavior, Quantified

The Minnesota Nice stereotype encompasses passive aggressive behavior.  Ya know, “I really liked your blahblahblah, BUT maybe next time you can meeglemorf instead?  Just a suggestion!”  Or something like that.  So one hypothesis is that maybe Minnesotans don’t give lower ratings out of politeness, even though they do have complaints.

To check this out, we did a quick count of reviews that contained conjunction words like “but”, “although” and “however” that would suggest a critique was on its way after a compliment, and we did this for 3 (out of 5) star reviews because those seemed like they could have been rated lower had a tougher critic been writing the review.  For Minnesota, 58% of 3 star reviews contained at least one of those words.  For Georgia, Massachusetts and Los Angeles, that number was 48%.  Passive aggressive, indeed.

Non-Minnesotans Give Lower Ratings to Minnesota Restaurants

A spot check of the Yelp reviews for 15 Minnesota restaurants showed that non-Minnesotans rated those restaurants lower — an average of 3.27 out of 5 stars versus 3.4 overall.   Now, 15 data points is not a lot of data, and 0.13 is not a huge spread.  But it points the same way as the other analyses — Minnesotans leave higher ratings.  And it takes away from the argument that Minnesota restaurants in our regional analysis were simply better than those in the other metros.

Cultural differences is probably not the only factor accounting for this spread, but it’s our best (or at least most interesting) guess at this point.  Perhaps an analysis of Massholes is our next step? 😉