Google on a Review Takedown Tear

Legitimate reviews appear to be collateral damage from a stepped up Google effort to address review spam.

Google on a Review Takedown Tear
Photo by Denis LORAIN / Unsplash

Google Product Experts Joy Hawkins, Ben Fisher and Amy Toman have pointed out a recent anecdotal uptick in complaints from businesses that Google is taking down their reviews. A number of Google Product Experts, myself included, have received out-of-forum emails asking for help.

This lead me to update my tracking of missing review complaints in the Google Business Profile Forums. Small business complaints about Google taking down too many reviews or not catching fake reviews tend to be a very good proxy for how aggressive Google's algorithm is towards spam reviews.

Uptick in Review Takedown Complaints

Review takedown complaints rose dramatically in March of this year. April and May saw historically high levels, albeit not as high as March. While June is only 8 days old, it would appear that unless Google dials down their spam removal algo complaints might reach even higher levels.

Notice that the uptick actually started in February

Why Is This Happening Now?

Google has clearly unleashed an updated ML-based review spam algo that is catching a great many more reviews. However, as we have noticed in the past, with any statistical AI/ML approach to review takedowns, some percentage of legitimate reviews are also affected. Generally speaking (but not always), complaints in the forums reflect the honest, smaller businesses that are struggling to get a few reviews only to see them not showing. Complaints typically increase anytime Google gets more serious about addressing its fake review problem.

While possible that it is a bug, it's more likely that this is an artifact of Google having a poorly trained AI on the loose and cranking up the dial on what constitutes a pattern for identifying review attributes that are similar to the known spammy reviews in the training set. That is to say "it is a feature, not a bug".

I don't think that the March timeframe for Google's increased review scrutiny is random or coincidental. In late January, the FTC issued a review abuse fine to Fashion Nova and put review platforms on notice about needing to have reasonable procedures to identify fake or suspicious reviews after publication.

Shortly after that, in early February, Google published a blog post touting their ability to "moderate reviews with the help of machine learning." At the time I perceived the post by Google as mostly performative but obviously Google decided to actually create some proof of their compliance with the FTC requirements vis-a-vis fake reviews, albeit in a typically Googley way. Whether this heavy handed approach is actually"reasonable" is another question.

What Businesses Can Do

First and foremost businesses need to understand that the reality of Google review takedown AI all too often throws out at least some part of the baby with the bath water. Given thqt the process is fully automated and AI based, losing reviews at Google is just "part of the game" and to be expected whenever they do in fact crank up the dials.

However, if you want to make an effort to recover these reviews, the first step is to reach out to Google support with the content of the review. Sometimes the business will have received an email from Google alerting them of a new review before it is nuked or, if you know the client well, you can ask them to send you a copy of the review from your business or their Maps contribution profile where it will still be visible to them.

After five business days, if Google support has not recovered the review, you can escalate the issue to the Google Business Profile forum and ask a Product Expert there to escalate the case. Be sure to include both the content of the missing reviews and the case number if you have one.  

A business should really assess the time that they are willing to invest in recovering lost reviews. All too often it is better to just put in place a consistent review-ask program and keep getting new reviews on a regular basis. Reviews are a long term game that should be viewed over a five-year horizon. Losing a few reviews, if you are gaining more every month, will not have a huge impact. Taking this approach is frequently a more efficient use of your time.