Is it creepy when brands pester you on social media?

By @edent

You're sat in a pub, chatting with your mates. You start to moan about how the cheap lager they serve gives you a headache. All of a sudden, a stranger runs up to you and says:

"Oh no! Headaches? Have you tried the refreshing taste of Pepsi® Cola? It's the Flavour That Keeps Giving™!"

I suspect you would complain to the bar manager and then find a new watering hole. You might, perhaps, tell the stranger never to contact you again.

And, yet, this is what happens fairly regularly on Twitter. I was complaining to my ISP (Virgin) when a different ISP (Sky) butted in to the conversation to sell their wares.

WHAT THE JUDDERING FUCK? Why would anyone want a brand hijacking a conversation like that? Is that sort of unsolicited electronic marketing even legal?

How are they doing this?

Companies like Sky are using BrandsEye to mine social media to send you unsolicited advertising.

We mine opinion data from social media to provide actionable insights for your organisation.

Their whole shtick is "find people complaining about your competitors - then bombard them with adverts." Innovative… And no mention of legal obligations or data protection.

A whole bunch of marketing bullshit.

In fairness, BrandsEye have a privacy policy where they promise "We will treat your information as confidential." Except for all these people we're happy to give it to. Interesting use of the word "confidential" there!

We will treat your information as confidential. However, we may disclose your Public Social Data to other third parties for the purposes set out below or for any additional purposes approved by you: Our Customers: our customers which include banks, supermarkets, manufacturers and other organizations; Internal Third Parties: other companies in the BrandsEye group of companies who will perform analysis and/or improvement of the Services and our internal tools and processes.

They do let you opt-out of their creepy marketing database (not that I ever opted in).

If you would like to exercise any of your rights above, please contact us at and provide us with your full details (including name, telephone number, address and email and the precise nature of your request and/or grievance).

So I dropped a Subject Access Request to [email protected] - to see what fascinating insights they had on me. It was truly pathetic.

What data did they have on me?

A few days later, they sent me 104 JSON files of my data. Some only contained a single tweet, like this one:

Here's the full data in the file:

[{ "link": "", "published": "2017-06-24T19:34:24.000+0000", "authorLocation": "Oxford, UK", "authorTimezone": "London", "city": { "name": "Oxford" }, "region": { "name": "England" }, "country": { "name": "United Kingdom" }, "language": { "name": "English" }, "category": { "label": "Consumer" }, "gender": { "label":"Male" }

Not much, is it? My location, gender, and that I'm a "consumer". What amazing insights they must be generating for trusted brands. FFS…

Other files held dozens of Tweets. Often on a single theme. Here's one which is just me talking about buses:

[{ "link": "", "published": "2018-11-13T05:47:03.000+0000", "extract": "@mskarengibb it is about a 30 minute walk into the city centre. Or, there's a bus stop just outside the venue. Google Maps will show you which bus to catch.", }, { "link": "", "published": "2018-11-12T07:56:53.000+0000", "extract": "@mskarengibb you'll need to go into the centre of Oxford and catch the busy from the Gloucester Green bus station,\nOxford OX1 2BX\n", }, { "link": "", "published": "2018-06-28T15:35:53.000+0000", "extract": "@TfL please can I get a response to this?\nHaving a bus stop listed in the wrong location is very annoying. Thanks.", }, { "link": "", "published": "2018-06-27T14:47:19.000+0000", "extract": "Some idle thoughts while I wait for a bus…\n\nDo you notice when you receive a \"Nudge\"?\nIf so, how does it make you feel?\n\nI'm talking about those messages which include things like \"95% of people pay their bill on time\" or \"people in [your postcode] tend to donate £X\" etc.", }, { "link": "", "published": "2018-06-27T06:24:14.000+0000", "extract": "*sigh* How do I view the latest NaPTAN database & see who is responsible for incorrect information?\nI assume it is TfL's responsibility to update bus stops in the capital.", }, { "link": "", "published": "2018-06-25T15:58:36.000+0000", "extract": "@OxfordBusCo this *still* isn't fixed. Signs up at the bus stop - but your app is still wrong.\nI reported this a month ago, you keep promising to correct it.\nCan you please tell me what's going on?", }, { "link": "", "published": "2018-06-18T15:27:16.000+0000", "extract": "@OxfordBusCo the Marylebone Road bus stop listed in your app is wrong. When will it be fixed?", }, { "link": "", "published": "2018-06-18T07:52:07.000+0000", "extract": "@OxfordBusCo it is now Monday, and the app still isn't fixed.\nYour website says one thing, the app says another, the physical bus stops say a third.\nYour customers are confused.\nPlease tell me when this will be fixed?", }, { "link": "", "published": "2018-06-12T12:49:40.000+0000", "extract": "Do you like open data? \uD83D\uDCCA\nLove statistics about bus travel? \uD83D\uDE8C\nCrave files in Open Document Format? \uD83D\uDD13\n\nDo we have a website for you! \n\n", }]

I can only assume that a bus company wants to know what sad, middle-aged blokes think about buses in Oxford. Hopefully my dazzling Tweets will provide value added insights into their corporate synergies.

How are they doing this?

BrandEye's website claims that they use "a proprietary mix of AI and crowdsourcing".

I'm going to call shenanigans on that. As far as I can tell, they're using the bog-standard Twitter API to mine keywords and locations. They might be using some basic Natural Language Processing to weed out the obvious errors. But judging from their LinkedIn page, their "crowdsourcing" is cheaply paid students in Kenya, and freelancers in Jordan, Venezuela, Mauritius, and Mumbai.

It's the old SpinVox trick - sell AI magic to customers, and get Mechanical Turks to do the real work.

Is this creepy?

OK, so BrandsEye don't appear to be doing a great job of mining social media. But is what they're doing ethical or useful?

On the one hand, I spew my tweets out into the ether. Anyone with an Internet connection has the technical capability to read them. But, those Tweets contain my personal data - be it my text, location, or images.

We've moved to a world where companies have to get an explicit opt-in to process our data and to use it for marketing.

I would argue that I have not opted-in to Sky's marketing. I have given no consent for them to send me electronic spam - no matter how highly targetted. Similarly, I haven't told BrandsEye that they can process my personal data. It isn't my job to opt-out of every half-baked business; it is their job to convince me to opt-in.

Even if the law were different, and companies could spam with impunity, is this an effective marketing strategy? Superficially, I get the appeal of searching for people complaining about your competitors - and then targetting them. But does it work?

Using special AI Algorithms and BlockChain Technology, I spent half-an-hour looking through Sky's BrandsEye-powered marketing Tweets.

Dozens of Tweets from Sky, pleading for customers.

It looks so needy and pathetic. Hundreds of spammy message. The only replies to Sky seem to be telling them to piss off. I'm sure once in a while they hit a receptive mark - but is it really worth annoying so many people?

The future is not brands sliding into your conversations like they're your mates.