Dastardly Data

This week I went on the excellent Gambling Files podcast to talk about the recent controversy over SkyBet’s misuse of data – which you can catch here.

I played no part in the activity, nor the investigation by Cracked Labs that expertly uncovered it (on behalf of Clean Up Gambling), but I know data, CRM and gambling malpractice, so I was deemed expert enough to comment.

I tried to listen to the episode, but unexpectedly couldn’t stand the sound of my own voice. So, I thought I’d take the chance to put my thoughts down in the blog without all the ums, ahs and knob jokes.

In short, SkyBet have been sharing their customer information with dozens of third party providers, and some have been returning the favour. That, in itself, is not a big deal. Every app and website does it.

Furthermore, SkyBet clearly state they are going to do it in their Ts & Cs (even if no-one reads them). They are transparent that data will be shared, and if you don’t like it, you can just bloody well go away. However, they are more opaque about who the data will be shared with, and for what purpose.

Some of their data partners are legitimate and of no cause for concern. They have an obligation to prevent money laundering/fraud, and so use agencies for that purpose.

Likewise, there are services that perform age verification checks. There are tens of thousands of underage problem gamblers in the UK, but ‘reputable’ operators are pretty good about checking that stuff when it comes to taking online bets.

There are also excellent services out there that will help a busy operator to identify potential problem gambling behaviour, but based on this sorry tale, I don’t think any of them are involved.

No problem so far. Everything is fine, legal and best-practice compliant.

A little more dubiously though, SkyBet share a lot of information (apparently in real time) with companies that are almost certainly advertising partners. Google and Facebook were listed as recipients and they’re #1 and #2 in the world when it comes to touting products and services for money.

I encourage a degree of precision advertising, but I don’t think social media platforms have any business knowing when I’ve made a deposit to a gambling site. Or, for that matter, if I’ve booked a flight or bought a kettle from Argos. From here on, privacy is being invaded.

But where SkyBet really crossed the line, is in pursuing their own little affordability checks. Not the good responsible gambling kind where less affluent customers are restricted from spending money they don’t have. No, the exact opposite of that.

SkyBet were, and presumably still are, receiving information from partners about how much money their customers can afford to lose. The investigation found they had gathered information on mortgage payments and credit ratings, which allows an algorithm to calculate how much more they might squeeze from you.

That’s a bad enough breach of privacy when you’re talking about regular customers, but this investigation was built on a GDPR request for a known problem gambler. A person who had a history of self-exclusion, and whose pattern of betting was unmistakably that of an addict.

A person who has admitted that their negative relationship with gambling led them to suicidal thoughts. SkyBet almost certainly didn’t know that last bit, but they would have known that this person was vulnerable and what that entails.

At this point, I’m going to take the opportunity to misquote the great moral philosopher Joey Tribbiani: “SkyBet have gone so far beyond the line, that the line… the line is just a dot to them”.

I don’t know how they justify this behaviour, but ignorance can be no excuse. I have been critical of the general standard of CRM activity, both in and outside of gambling, but I have to grimly take my hat off to SkyBet here. The execution of this activity is impressive.

They understood exactly who this customer was and how valuable they could be (much more so than they had any right to). From analytics on their communications, they also knew what buttons to press to elicit further custom. That’s a technical CRM master class.

Where it falls down, is in the complete moral bankruptcy involved in exploiting an obviously troubled person for further profit. Like a maffia thug, SkyBet had worked out exactly how many punches it could land to leave the customer broken but breathing. It’s stunningly poor practice.

The complete lack of compassion and decency makes me think, and even hope, that this is an error of machine learning. Perhaps somewhere in the bowels of a server room in Leeds, an unsupervised algorithm worked out that the most cost effective way to make money is to target players that have self-excluded, because they are the most responsive to communications and tend to spend the most money.

That’s exactly the kind of emotionless and erroneous insight that artificial intelligence might reach. In his excellent book The Data Detective, Tim Harford describes an algorithm that could tell photos of huskies and wolves apart with the same accuracy as any human.

It turned out the computer was ignoring the canids completely and simply looking for snow in the background of the pictures. An ostensibly correct result was being obtained, but the methodology was horribly flawed.

The alternative explanation is that the leadership of SkyBet is evil and greedy. This situation, and their previous (possibly related) disaster of emailing over 100K problem gamblers with a free-spins offer, does hint that Skeletor or Cruella might be in charge.

Regardless of how this was allowed to happen, someone needs to be fired and exiled from the industry. If gambling wants to clean up and improve its reputation, then the bad guys need to go. Somewhere at SkyBet there is a culprit still in a job; and by extension, at Flutter Entertainment too. The master has a responsibility to control its dog.

If the industry needs a mantra to help avoid these problems in future, then look no further than the wisdom of Jon Bruford, the learned host of Gambling Files, who said: “Let’s just not be sh#t”

Well spoken, sir.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s