Get Free Quote
« Using Data to Drive... | Blog home | Google Social... »
April 12, 2012

Why Is Google Allowing Rich Snippet Spam?

Print Friendly
Estimated reading time:
3 minutes
Top takeaways:
• One of the new microformats for the semantic Web for reviews is now being abused.
• Google has a Rich Snippets spam reporting form, but the jury is out on if it’s being enforced. 
• Should Google stop showing starred reviews in the SERPs until it can get a handle on verification?

A couple months ago, I read a great article on how to get extra rows of stars in Google Places by inserting the new hReview markup code onto a page of your site. One of the new* microformats for the in-progress Semantic Web, hReview lets webmasters markup customer reviews for products and/or services so search engines can understand what product or service is being reviewed, who’s reviewing it (or how many reviews there are if it’s aggregated) and what rating they’ve given. The rating shows up on SERPs as a row of golden stars.

*(Update 4/13: As Aaron Bradley points out in the comments below, hReview dates from 2005 so it’s hardly “new”)

The hReview format just looked too good to be true, especially because it could work for any listing on the SERP, not just local listings. I thought to myself, all you have to do is stick a scrap of code on your site and Google will give you a row of stars beneath your SERP listing?

My immediate reaction was, “I’ve got to tell all my clients!”

Followed by, “Wait . . . what’s keeping anyone from just giving themselves whatever rating they want?”

And finally: “This is going to get abused, so fast.”

Turns out, I was late to the party with that line of thinking; as early as June last year people were predicting Schema.org markup would get spammed.

And they were right.

Obviously Fake Aggregated Ratings

I’d all but forgotten about the hReview markup until last month, when I came across some starred reviews while doing unrelated research. Here’s the entry I found (censored because I don’t want to give the site any undeserved publicity):

Wow! What a great rating! And so many votes! How’d they get more than 8,000 people to rate them?

The answer, of course, is that they didn’t. Well, I can’t say for certain, but my SEO-sense sure is tingling. This isn’t some Fortune 500 company that might have the reach to achieve this; the site in question has almost zero social media presence. And considering how the actual site is crammed with other negative trust indicators, I feel pretty confident in calling shenanigans.

This isn’t some minor issue; a new study shows that most consumers trust online reviews as much as personal recommendations. By displaying stars in the SERPs, Google implicitly grants authority and legitimacy to any site containing the hReview markup.

Out of curiosity, I went to Google’s rich snippets spam reporting form and reported the site a few weeks ago. Then this morning I checked to see if it worked. Here’s what I found:

Yup, the starred review is still showing, though for some reason, their rating went down by 0.2 points and they seem to have lost exactly 7,000 votes. Funny.

But has Google slapped them with any kind of penalty? Nope. The site ranks No. 1 for some queries with a not-inconsiderable amount of search activity.

(From what I’ve read, it may take up to five weeks for Google to take action against reported spam, so I’ll keep an eye on it.)

A little further down on the same page we find another recklessly high aggregated rating:

Google Sending Mixed Signals

OK. But so what? The problem isn’t the reviews themselves. The problem is that they don’t have to be tied to anything. There’s no kind of third-party verification required.

What’s worse, Google ads sometimes show a row of stars tied to third-party reviews:

Yes, these two entries actually show up like this. How can users tell which is real and which is fake? How would users even know to suspect one might be fake?

The stars in the PPC ad are actually tied to something outside the site, as you can see by clicking on the anchor text, “113 reviews”:

I suppose these could all be spam too, but that would require a whole lot more effort.

But let’s say you really do have actual customer reviews. Is there any way you can demonstrate their authenticity to a greater degree than the fake reviews?

Well, as David Naylor points out in his article on Google’s crackdown on rich snippet spam, some sites with fake reviews are referencing an internal or external review page.

Customer reviews on your site can be manipulated with ease, so that’s out. One of the sites in Naylor’s article above references an external review site, but there are plenty of sites where you can pay for fake reviews.

So even if you have actual reviews from flesh-and-blood customers, there’s no way to authenticate it that couldn’t simply be spammed.

What Gives, Google?

The bottom line is that Google should stop showing starred reviews on the SERPs, at least until it comes up with some sort of verification system. It’s just too easy to spam, and the potential rewards greatly outweigh whatever risk there may be. And even if your reviews are all on the up-and-up, there’s no way to demonstrate that. By continuing to display rows of stars in the SERPs, Google is implicitly encouraging spammers.

What do you think? Is this really that big an issue, or just much ado about nothing?

Print Friendly




23 responses to “Why Is Google Allowing Rich Snippet Spam?”

  1. Aaron Bradley writes:

    Good article, Bob, but just to correct one point for your readers. You say:

    “One of the new microformats for the in-progress Semantic Web, hReview….”

    I had to check the date on your post to make sure it had indeed been published today. hReview was introduced in April, 2005 – nearly seven years ago – so I would not call this “new.”

    While you don’t quite say it directly it is indeed schema.org microdata (and in particular AggregateRating) that’s been identified as the main culprit where ratings spam has been recently been detected.

    Given the noise made over faux reviews of late, I do expect to see action taken by Google fairly soon to fine-tune how review snippets are granted. However, your suggestion that “Google should stop showing starred reviews on the SERPs” is unlikely to be followed, and indeed this sort of all-or-nothing approach to rich snippet spam is not required (according to this stricture we would, for example, no longer see ratings attached to well-established and sites with trusted reivews like Amazon).

    I agree to an extent that “some sort of verification system” will need to be put in place, but my “to some extent” qualification is because
    Google doesn’t simply take review rich markup on face value at present. It’s clear from what’s been said (see the Inside Search blog posting in Jan. 2012 on “30 search quality highlights”) that Google is enabling review rich snippets on a site-by-site basis, so there is something like a “verification system” in place, albeit an inadequate one. In all likelihood we can expect to see further trust and proof mechanisms put in place for reviews, probably as another algorithmic layer specifically for review content, on top of simple site trust.

  2. Bob Meinke writes:

    Ah, thanks for the correction. I didn’t realize hReview had been around so long.

    My understanding is that Google was originally whitelisting sites for rich snippets on an individual basis, but now they’ll display rich snippets for any old site. Yoast.com did a good post on it.

    Well, OK, maybe I’m a little hasty to advocate an all-or-nothing approach, but I think it’ll be a huge problem until they reinstate a whitelist or find some other verification system.

    Thanks again for the great comment.

  3. Richard Falconer writes:

    I think it’s great that a consensus seems to be forming in both the UK and US that this spam is a bit out of control. I posted yesterday with some similar points about UK SERPs: http://blog.bigmouthmedia.com/2012/04/12/google-ruining-own-results/

    Well done Bob Meinke /BruceClay for also speaking out :-)

  4. Andy Kuiper writes:

    Interesting article Bob :-)

  5. Bob Meinke writes:

    Richard –
    Thanks for commenting. Looks like our posts are hammering on the same nail – it’s so bizarre that Google is allowing this to happen. But I really like your Admiral Ackbar “It’s a trap!” theory. Just crazy enough that it might be true. . . .

  6. Bucheron writes:

    Hi Bob

    The problem lies in the third party verification.

    How could someone or something verify the votes ? They can use proxies to fake the IP, with differents referer and configuration, or even use Tor which would be quicker.

    I think there’s nothing to do on that point, if we don’t want to also penalise the “good guys”.

  7. Arun Agrawal - Ebizindia writes:

    Google takes an automated approach to all such things and therefore has to activate the listings automatically subject to manual verification wherever in doubt (or randomly).

    However I believe when they find people misusing it, they might blacklist the sites with respect to just the stars or the whole listing itself.

    Most of my sites and clients using the system are doing just fine but we do refer to a legitimate (but not 3rd party) review base.

  8. Paul Gailey writes:

    Bob – It is an issue but I suspect the whitelisting model from before is not as scaleable as the current setup, even if that means part reliance on other SEOs using the report tool to police the web. I can’t see how Google can otherwise support a structured markup web and be selective with reviews to cater for spammers. That said I think the risk is higher for spammers who get do get snippet denied compared to having their paid backlinks deindexed, in that the penalty applies direct to their domain and not just their paid network. Play with fire, get burnt.

  9. Zach writes:

    Nice read Bob.

    Still trying to understand why they are allowing as well. But I am not complaining either, CTR is way up :)

  10. Chris writes:

    Google should also only give credit to the site that is the original source of the reviews. Take for example, Google’s book pages. site:books.google.com The bulk of their rich snippets are from another source (goodreads.com). There’s very little original review/ratings on their own pages, yet by partnering with Goodreads they now have ratings and rich snippets for thousands of books. Seems like they should only show up for Goodreads in the organic rating and Google Books should wait until they’ve accumulated their own content.

  11. Ellen writes:

    I just came across a case of review abuse while doing a search for “local seo services” where the top result is using the aggregate reviews microdata format with no reviews whatsoever on their site.

    Once the abuse comes to the point where people are questioning Google’s serps then Google will put some policies into place. That then will mean that all of us who are using the micro formats correctly will have to update our sites with whatever the “new” policies end up being.

    My little rant for the day!

  12. Saney writes:

    There is no clear difference between white and black SEO. There-fore, if these methods are not banned – they are legal. These people are experimenting, and there is nothing criminal – except 4-digit num of votes. Btw. Google rich snippets and hReview appeared in 2009.
    To become a professional in SEO, it is necessary to experiment with the bad things.

    I use another method of increasing the CTR for several years, you can write an article about it too :). However, this does not change the world.
    saney.com/tools/google-snippets-generator.html

  13. Andreas writes:

    It really sucks that google is allowing this.
    It appears that they will remove sites with static hreview text after a rich snippet spam request but as long as you have some ajax star rating there you’ll be fine, even though you have a super niche site with 4384 ratings :/
    My competitors all get much more clicks due to their stars and by the time I implement these artificial stars on my site google will start to penalize those pages. Guaranteed.

  14. Richard F writes:

    Bob, this reminds me of Google maps. There are so many fake reviews and locations in certain industries. It’s been years and they still haven’t cleaned that up.

  15. Karan writes:

    You know its funny you mention this….. I just got an email from one of the many syndication services (that recently got slapped) about this very same code.

    Google is allowing this but not for long. Every grey/black hat method has a shelf life before it gets dealt with.

    The problem is that new methods are constantly created so this keeps the spammers perpetually in business to the disadvantage of the honest webmaster.

    It is the basis of Google’s Algo that is the problem. These bandaid methods will not work for long.

    I wrote a post about this on my site. Please check it out and tell me what you think.

    thanks

  16. Milind writes:

    Bob,

    I am from eBrandz. And I am surprised you have made this statement:

    “What’s worse, Google ads sometimes show a row of stars tied to third-party reviews:”.

    The reviews shown besides the PPC Ad come from Google Checkout (Product/ Seller). And it is very difficult to fake these reviews. Why? You have to buy something using your credit card from Google Checkout (now Google Wallet). Only after you purchase anything will Google give you an option to rate the merchant.

    But I do agree with your overall point that reviews and ratings are getting abused.

  17. Bob Meinke writes:

    Milind:

    Thanks for the feedback. I guess I should have been more clear, but I used the eBrandz ad as an example of a row of stars that is dependent upon verifiable reviews that, like you said, are very difficult to fake.

    My complaint is that unscrupulous webmasters can easily get the exact same row of stars, which will erode trust in legitimate starred reviews like eBrandz’s.

  18. Bob writes:

    Thanks for the clarification. Appreciate it!

    I am actually surprised Google did not think this one through. What did they think would happen once they started showing ratings and reviews in search results?

  19. Ryan - RMC Tech writes:

    Interesting stuff – So, I think what I’m getting from this is, you could get potentially better positioning, using rich snippets, if in fact the reviews are legit, and you’re not gaming. Maybe.

    I’m all for it! I do have reviews

  20. Charles Edmunds writes:

    Ellen I think we must have come across the same website or there’s multiple “local seos” that are manipulating the SERPs. It’s going to be interesting to see how Google treats these spammers.

    On the local side it’s a shame because a local business owner trusts Google’s results pages and will likely get a sour taste for SEO “in general” if they get taken for.

  21. Rex writes:

    Hi

    I have just implemented rich snippet in my website. But there was an error. a massive one. My programmer had mistakenly implemented the no. of testimonials as the no. of votes for all the products (more than 400+) and before we can check, the pages are all crawled. Now I am not sure what to do? Should I pull down the rating system altogether and start from scratch. or just keep the present values and allow customers to vote. But the think is that, all the products are having same votes and same star ratings… looking like spam.. will google give penalty for this error?

    Thanks

  22. Bob Meinke writes:

    @Rex:

    Thanks for the question!

    If it’s just been implemented, then your best bet would probably be to do it over, especially if you think it looks like spam. If it was a small error you could just leave it and let your actual customer reviews dilute it, but with more than 400 you’re better off scrapping it for now.

  23. mazdodot writes:

    Hei Bob, I had implemented rich snippet. Test on Google Rich Snippet Tool and no error! Thanks for your article, I now better understand what is Rich Snippet.



Learn SEO
Content Marketing Book
Free Executives Guide To SEO
By continuing to use the site, you agree to the use of cookies. AcceptDo Not Accept
css.php