Yelpsters

There's been a lot of opinion expressed on both sides for this topic - most of the feedback is overwhelmingly negative but there's a substantial number of positions which are quite positive.

However looking at specific examples I've been able to witness first hand seem to weigh in on the former, although in reading up on this one there's been some comedic uses of the review sites - such as a bistro in the US offering a discount to customers providing negative reviews. There's an unknown amount of trade for former "SEO Consultants" who are now plying their trade to review site optimisation too, inferring that there's a sizeable economy surrounding these platforms.

Not only did it make a mockery of the review system on Yelp specifically but it highlights how abusive some of the platforms can appear. For example we got a local tradesman to fit the new carpets in our home and whilst talking to him he specifically asked us to read his reviews on Facebook, not Yelp.

Of course a comment like that is going to peak my natural curiosity so I dug a little deeper whilst the carpets were being laid, uncovering all sorts of fun & games (and stress) had trying to deal with the Yelp reviews. Having a quick look at his business Yelp page I can see what he means.

Of the 23 reviews that I could see at the time, 12 had been removed by Yelp as violating it's "terms of service", and 9 were "not recommended". This left two valid and "recommended" reviews - as far as Yelp itself were concerned - that were viable views on the business involved.

I couldn't understand the criteria match involved as the two remaining reviews were from Qype user accounts, and Qype hasn't existed as a platform since it was absorbed into Yelp in October 2013. To my mind more recent reviews over the previous twelve months are more likely to be representative of a business than something logged about a problem in three years previously.

For example, a restaurant can change hands and improve it's quality of service as a result; or a museum changes exhibits regularly providing differing levels of engagement over a longer period of time. TripAdvisor solves this problem by refusing any reviews over 12 months old - although they may still be visible their priority and importance in the overall business rating are lowered slightly too. On the face of it that makes more sense to me and allows business to atone for any past mistakes (as well as ensuring they don't rely on previously high ratings to boost their search rankings).

Yelp does not do this from any of the business reviews I've seen. It tries to hide the majority of reviews with a greyed link at the bottom of the page (enlarged in this image for ease of reference).

Suspect Carpety Reviews

The rest of the reviews - which are almost all positive from what I've read - are hidden in a "...reviews that are not currently recommended" moniker, and labelled essentially as irrelevant.

Yelp makes the following statement about this logic: 

Yelp statement on "Not recommended" reviews page

Now there's a few interesting sentences in there I'd like to focus on which don't seem to tie-in with a common-sensical approach. The first (highlighted yellow) seems to indicate that the reviews are assessed on quality and reliability combined with user activity. I don't understand how that's possible in Pete's case as there are a number of reviews in the last twelve months from people who have created a Yelp account purely to commend Pete's carpeting ability. 

But the two reviews which are being counted also indicate those users only ever made one review, and are from 2013 - neither those reviews or most of those on the "not recommended" page have any Yelp "friends" which may be the algorithm metric involved. Doesn't seem to tally. So the Qype user reviews giving Pete 5 stars are "not recommended", but the Qype reviewers panning Pete are on the front page.

Pete doesn't advertise on Yelp and rejected their sales teams suggestions that he buy a business account, shortly afterwards he says the problems really started. Customer reviews got moved away from the front page and his Yelp star rating shown alongside the search result dropped to a one star. Of course Yelp vehemently reject any suggestion that the two events are linked but Yelp only have a 2 star rating on their own website.

After Pete fitted the carpets I added my own review for his business on Yelp - creating a new account in the process. I also added a review for another business I'd used in the previous six months just to see if that was the deciding factor but, as if by magic, my review went from the front page to the "not recommended" page within 24 hours.

So going back to the Yelp statement the blue highlighted note about advertisers getting preferential treatment cannot really be proved or disproved without actually seeing the underlying algorithm first hand. However the majority of positive reviews for this specific business would nominally give Pete a  4.8 out of 5 (4 star service). The reviews de-listed because they "infringed the terms and conditions" would provide a small increase over that.

Should business owners like Pete care about Yelp? I'm not so sure - we did a lot of searching on other platforms such as Check-a-trade and social media; on both of those areas Pete scored highly based on recent customer reviews. Everyone gets to make up for past mistakes.

I also recently had an interesting conversation with another platform provider - Trustpilot.co.uk - who apply reviews of businesses and websites. I thought I'd try something related to my anti-spam activities and lodge appropriately negative (but entirely honest) reviews about two data trading businesses. The two businesses involved are AdView and UK Staff Search - both trading brands of Roxburghe / Dash Marketing who scrape job seeker details from jobs boards such as Jobsite.co.uk, pretend to operate in the US and then sell the personal data to various unsolicited marketing firms back in the UK.

The review went something like this:

Actually not as detailed or specific as it could be

The review omits details about how the company broke the law (DPA & PECR) but notes the basic facts.

Both business lodged a complaint about my reviews, and Trustpilot were of the opinion that despite AdView illegally acquiring my data and then spamming me I didn't qualify as a customer. Despite that being the very definition of an Adview customer they deemed my review in breach of their terms and conditions. And also despite AdView staff writing most of their reviews themselves. Perhaps if word of this post gets to them this review will also be removed by Trustpilot but I was completely unimpressed with their reasoning and position.

However the USS review stood and today someone marked it as "Helpful". That's the real purpose behind these platforms - finding good opinions and using them to choose the right product or service.

Overall I'm going to continue taking all reviews and opinion with a pinch of salt and actually talk to people to see if they know their trade. It's crazy but it might just work.

Trial Result: UK Apollo Group

In a surprise ruling, the judge decided that I could not prove the claim against Apollo.

The reason? I'd forgotten to include in evidence the documents that showed UK Apollo group scraping email addresses from the rest of job seeker data in my bundle. This was necessary as the initial spam was sent to an encoded email address (e.g. jobsite.<date>@<domain>).

So essentially it was not the case where the defence was robust and proven, it was simply an error on my part which failed to cement the facts of the case. UK Apollo would not be able to refer to this year long legal entanglement as a 'victory' as a result, and they also disclosed a lot of facts in a court of law - facts that would be of interest to regulators and in future SARs.

Furthermore, the judge refused to accept evidence of spam emails received after the submission of the date of the claim, which included readily identifiable email addresses. He also noted - but failed to act on - the fact that the first defence was submitted unsigned, and an alleged re-write of the defence was never served to the claimant.

In fact the trial was the first time I'd heard that Apollo had even adhered to the application to strike / re-write judgement.

Despite a number of breaches of CPR by the defendant (duly noted but not enacted by the judge) the defence - such that it was - was allowed to stand, despite a submission related to Denton & others. I even submitted a revised bundle for the second trial after the court ushers directed me to the wrong floor in the first trial - but the judge claimed not to have it in front of him. Good thing I got delivery receipts then.

Very interesting - almost as interesting as the very personal remarks made by the defendants rep, Keith Taylor. He was very angry! He couldn't actually apply a robust defence at all and I'll share the highlights of the comedic vitroil once I get the trial transcribed. You shouldn't laugh too much in a court. At one point he claimed that ICO was getting the law changed to help his company continue to spam people. Of course, no evidence was presented to substantiate any of these statements.

Keith even claimed he charges his time at £1k per day, although judging by the posted accounts none of the companies seem to be charging for many of his consultancy days. Most of the people associated with Apollo appear to have many other jobs too....

Also of note was the defence at one point admitting liability at two points, saying "just find me guilty, m'lud and fine me￡200 so we can all go home.". In court, on record.

However because the district judge excluded the emails following the initial claim document they are not considered part of the claim that has now been judged by the court. One of those spam emails was to an account I have listed as <firstname>.<surname>@<domain> - which is personal data in itself.

I think I can see why the judge did what he did and I've decided not to appeal for a number of reasons (not least the additional costs liability if I get something wrong). So I sent Apollo a shiny, brand new SAR last week - they've read it twice but are yet to respond. He awarded a £55 cost for the defendant - which should cover his petrol home.