How Google is destroying trust in its own results

Posted on 12. Apr, 2012 by in Digital Marketing, Thoughts



Google is showing more sites with Rich Snippets review data in its results. It allows sites to use structured markup to code customer review statistics on their pages. This is all well and good except that site owners have started to take liberties.   What Google appears to think users say on seeing rich snippets such as the above: “Wow! This company has 957 reviews with an average rating of 9.9 out of 10! They must be an awesome loan company despite the APRs of 14348% being offered on their homepage. Thanks, Google!” What users actually say: “Hmmm. I don’t trust these Google star thingys. Google sucks.” The number of sites doing this appears to be increasing unchecked. Admittedly I base this on “stuff I see” rather than any statistical analysis but here’s an example to make my point: In the top 100 results for the term “payday loans” on Google UK I see 12 sites who have results with Rich Snippets. The average rating for these sites is 9 out of 10 (adjusted for sites who use marks out of 5 etc.) Considering the difficulty of keeping customers happy when you have to turn them down for a loan or collect money when they default (especially at APRs of several thousand percent) those figures already seem suspicious to me. The average number of reviews per site is 405. The average number I can actually find online is not much more than 10 per site. For most I couldn’t find any.

Great reviews – these companies must be incredible!

If these amazing reviews existed about your company you’d want to link to them, wouldn’t you? You’d be proud and consider them an asset to your company and a tribute to your staff and great business model. Wouldn’t you? There’s nothing embarrassing in a score of 9.7 out of 10. Reading genuine individuals saying what great service they got could only increase conversions. It’s strange then that there is so little evidence that many of the above mentioned reviews even exist. Only 2 of the sites using rich snippets link to any of the reviews they claim to have received. The two sites who link to their reviews link to reviews which lack credibility, in my opinion. One (www.paydayloans.co.uk) links to a slightly comical page containing 5 one line reviews along the lines of “wow love it” from “jon”. The other (www.quickquid.co.uk) has made more effort and links to Google Shopping seller reviews which are aggregated from the likes of ReviewCentre.com and TrustPilot. This seems reasonable enough except that of these reviews, most of the 5 star ones feel very forced and unconvincing. The lower ratings, on the other hand, feel frustrated and genuine. Perhaps that’s just me being cynical. In fairness to these sites, at least they have reviews, genuine or otherwise. Quickquid.co.uk is the best of a bad bunch. QuickQuid Reviews

[update - for more information on QuickQuid.co.uk, see the end of this post] As bad as these two might seem, the other sites were much, much worse. Not one links to or publishes any of their reviews. The review statistics they use is generally hidden around or under the footer (again – embarrassed of your 9.7 out of 10?) In most cases I couldn’t find any reviews online, in some cases there was a lot more negativity than a score of near 10 out of 10 would suggest. 247 Moneybox has a rating of 8.5 based on 929 reviews. ReviewCentre.com gives them 3.2 based on 66 reviews, loanfinder.co.uk gives them 2 out of 5 based on 16 reviews (although it might be run by a rival) and on TrustPilot they get 7.7 out of 13 reviews. The other 834 reviews are, erm, hard to find.   I presume the integrity of and trust in Google’s Rich Snippets is severely compromised by abuse of Rich Snippets but  it would appear that Google is allowing it to happen. Something is definitely amiss, or at least there’s something else going on behind the scenes at Google.

So, what’s going on? I attended a Google Webmaster EDU seminar recently in London where Googlers Jack Menzel and Pierre Far answered questions on structured markup. Jack’s purpose for being there seemed to be to promote Schema.org, Rich Snippets etc. A question was asked about whether Google wants sites to publish the reviews which are marked up in their snippets or whether that doesn’t matter. I was surprised that the answer given was that Google wants you to have to reviews on the page but if they are elsewhere but genuine, that’s also fine. Fair enough – but there’s no mechanism for verification, doesn’t that just mean I can put any numbers I want in?

It would appear so.

Google has a lot of experience in avoiding spam in their results, so it’s surprising that they’ve allowed this to happen so widely. I’d be surprised if Google were not aware of the scale of the problem. I don’t know what Google is up to but here are a some theories about why they might be allowing trust in Rich Snippets to be eroded in this way. From sensible to paranoid, in that order:

Building A Better Algorithm Theory

There is a special Rich Snippets report page to report spammy rich snippets which suggests they are aware of the problem and are collecting data on sites that are abusing the system. By identifying less trustworthy sites, Google can build more sophisticated algorithms which understand the signals that trustworthy and untrustworthy sites with Rich Snippets give off. When improving its linkspam detection, Google put out requests for SEOs to report linkspam. Requesting Rich Snippet spam reports would be following a similar pattern of behaviour, the only difference being that they are, to some extent, encouraging the spam in order to kill it.

The Dangle A Carrot Theory

Google wants to get site owners implementing structured markup more and it’s possible that uptake has been slower than Google would like. By allowing sites to take a few liberties and gain traffic as a result, site owners and SEOs get the word out that schema.org is a wonderful way to get traffic.

The Admiral Akbar / Tin Foil Hat Theory

It’s a trap. With all the talk of Google “levelling the playing ground” with regards to “over-optimisation” (by the way, Matt Cutts did not say “over-optimisation penalty”, that was searchengineland.com), this could be a great way to trap the spammier SEOs and their clients.

The evening after I wrote this blog, I received an email from Enova, the parent company of QuickQuid.co.uk whose reviews I mentioned above. They seemed to appreciate my point of view but were keen to give their side of things, which I’m delighted to do. They point out that they are large multinational company employing hundreds of people in their customer services department. I won’t go into the numbers but it’s a huuuge amount of money that they lend.

They asked (quite nicely) to point out that they say their reviews are genuine and that they do not give out incentives to write them. I’m thinking that if Enova are putting the same effort into their customer services as they are to their online reputation management, their reviews may well be genuine.

Enova say they “actively request positive and negative feedback from customers and prospective customers.” Here are their reviews: QuickQuid Reviews 

The point of the article was not really about the spammy reviews/non-reviews, it’s about the integrity of Google’s Rich Snippets stars system. I’ve asked Enova for a comment on whether they feel the ratings they work hard for are being compromised by competitor spam.

by Richard Falconer

Tags: , , , , , , ,

  • http://twitter.com/rtavs Ross Tavendale

    I’ve said it once and I’ll say it again. Top article Rich! 

    I like the “it’s a trap” analogy but I haven”t made my mind up if this is a good or bad move by the big G.

    I love the idea of proper mark up being a must and like the fact that SEO is becoming harder as it weeds out the spammers and the snake oil salesmen from the industry. 

    I would love to see some split testing with review mark-up in the future.  The only problem is that it is so hard to build out while keeping all variables the same. 

  • http://www.ydeveloper.com/e-smart-ecommerce-suite.html eCommerce

    It’s true that the rich snippets in Google search results can be personalized. It means the rating and reviews should be a owners choice.

  • http://twitter.com/MrHeyes Simon Heyes

    I think the team over at Bruce Clay have the same sentiment… 
    http://www.bruceclay.com/blog/2012/04/google-rich-snippet-spam/

    • http://www.bigmouthmedia.com bigmouthmedia

      appove

  • rich_falconer

    Interesting one that Simon. I’ll explain later ;-)

  • Pingback: Why Is Google Allowing Rich Snippet Spam? | jhWebWorks | Columbus Ohio Web Design, Development, SEO, Social Media

  • Spook SEO

    I agree with Richard on this one and Google should do something to mitigate this issue. Trust is Everything. How can you go to a dentist that you don’t trust? make sense, right?