Video Deep Dive: Does Review Sentiment Impact Rank?
Mike Blumenthal

This is the 5th installment of of 2016 of our Deep Dive Into Local series in 2016. For the week ending Monday, February 8th, Mary Bowling and Mike Blumenthal shared their thoughts about the previous week in local. The complete video, including links and commentary on critical happenings of the previous week is posted in the Local U forums (paywall). In the second half of that video, they take a deeper strategic and tactical dive into one interesting area that caught their attention during the week.

In this discussion, we look at the role of reviews in rating website quality and if sentiment impacts rank.

Mike B: With that, let’s transition into the Deep Dive. And talking about people, places, and things, we’ll discuss the role of reviews and review quality, review sentiment, and in Local. Why don’t you kick that off, Mary?

Mary: So what prompted us to talk about this today is if you read the newest Search Quality Rater Guidelines from Google, for the first time, it talks quite a bit about reputation and how reputation should help you determine the quality of a business and a website. Let me find a quote here from it.

For example, it’s one of the directions that the raters are given is, “You must also look for reputation information about the website. We need to find out what outside independent sources say about the website.” It talks about how big brands are expected to have reviews, whereas if a small business doesn’t have reviews, that’s not necessarily a sign of bad reputation. So I really suggest that everyone takes a look at these, the guidelines in general, and looks specifically for these little snippets in there talking about reputation. Because, to me, it indicates that even though Google may not be able to do a very good job of incorporating review sentiment into its guidelines, it very much wants to. And perhaps having humans review this and give the algorithm some ideas of what it considers a reputable website.

Mike B: A reputable, yeah. A reputable website, just from the external observation. I don’t know if they’re using sentiment directly in the algorithm. I mean, Google has trouble, I think, to some extent with the noisiness of sentiment. I do a lot of Google surveys where you do open ended questions, and they have a huge trouble collating those questions accurately. We also saw when they were scraping review content and third party content, bringing in a GMB, they had a lot of spurious and noise unrelated things that they thought were reviews just because it had the word good in it or the word bad in it, when it really wasn’t a review at all. It was just a page on a website.

So I think they still are having some problems with the sentiment, but if they take the approach of trying to assess good quality websites that correlate with high quality reviews, that certainly adds a human element to it that may give those with good quality signals, externally, an edge. And then, also in the patent on entities, I think we spoke about this several months ago. Barbara Starr wrote a good article on it, and she mentioned, or the patent mentioned, that they were looking at professional opinion as well as regular crowd sourced opinion.

It isn’t hard to imagine either that they would rely more heavily on, for example, Google Local guide opinion than just regular reviews, or perhaps even Google Yelp Elite as opposed to just regular reviews. Because both of those, as well as professional reviews, are relatively easy to identify, for example, a review with the New York Times or San Francisco Chronicle, from a professional. So it’s conceivable to me that they’re also bringing in those elements into the review space, particularly in food and hotel space, where they’re more prominent.

Mary: Yes, most definitely.

Mike B: And movies.

Mary: And this kind of all reinforces Yelp review-er algorithm that it’s had all these years, that maybe that was the way to go to begin with rather than actually looking to what’s being said in the reviews but who’s saying it. But I do think it makes sense from Google’s point of view to want to reward businesses who have good reputations by boosting them up in the search results because, you know, who wants to go to the worst Chinese restaurant in town? None of us.

So possibly, those types of reviews might only surface in the future if you’re looking at a particular brand. But I do think it’s very interesting. I think it’s appropriate. Like you, I don’t think it’s here yet, but I think it’s a really clear signal from Google that they would like to incorporate review sentiment into the algorithms and that’s yet another reason why people need to be working on getting a body of more good reviews over time and developing their reputation rather than just dealing with reviews one at a time.

Mike B: Great summary. With that, I think we’ll wrap it up. I just wanted to mention that Local U Seattle is sold out. MozCon Local in Seattle is sold out. We still have seats available for the Local U Advanced in Wlliamsburg on the night of March 4 and 5, and through an interchange with one of the potential attendees, David has agreed to take on the charge of making sure that there are a lot of good IPAs at the gathering, Friday night. So if you’re an IPA fan, it’ll be a particularly enjoyable weekend.

So with that, I think we’ll say good-bye. Thank you very much for another week in Local.

Mary: Bye-bye.

Posted in