This linked post has been doing the rounds recently and there seems to be near-unanimous acceptance of its central message. Clever use of Big Data got Trump into the White House.
While the article is a bit scary and it does make you want to unplug yourself from the Internet grid, I think a bit of healthy scepticism is called for. In whose interest is it to claim that Big Data can be utilised to control any outcome at all? It’s a growing industry and the phrase Big Data currently makes ears prick up at Board Rooms across the world. There’s money to be made – and the way to do it is terrifyingly simple.
If you are in any data mining organisation at all – no matter what level you operate – you can truthfully bask in the glow of acclaimed Big Data success stories. In other industries you have to prove what *you* personally do or have done is tremendous enough to get business. However, Big Data is an industry which people barely understand and if you get the ear of a corporate decision maker, you can make claims that sound amazing – even if you had nothing to do with any of it.
Consider the fictitious case of Eric – a 25 year old average ability non-league footballer that looks a bit like Lionel Messi. Eric shows footage of Messi doing amazing things for Barcelona and then by strong implication promises a third tier football club chairman similar outcomes in return tor large amounts of money. The Chairman, hopelessly bewildered by the modern world of football, hires Eric. The manager and staff face palm themselves sore when they hear the news and Eric is announced as their new expensive megastar player. The Chairlman is the boss and has bought into all of this, so the operating staff have to go along with it.
The narrative of the linked article also makes me a sceptic. A scene is set as in a novel and the expertise of the main character paints a picture of someone who is part of a very smart, efficient machine that is controlling things. We tend to like stories, so a mild-mannered data analyst can easily be cast as Superman. The other unifying element of such stories is the anxiety that it’s already too late. By the time you read this…you’ll be DEAD. All that’s required is the next line – “unless you pay me to make all this work for you”.
At the Cambridge Analytical level, I am sure that the finest minds are algorithming their way to millions and we should be respectful and wary. Yet strategic decision makers at even very large companies are horrendously out of touch. The business approach is – if you can’t be bothered to learn, just outsource to someone who seems to be saying the right things. It shows the shareholders that you are Doing Something.
There is also a scary underlying theme here that Trump is a genius on the quiet, who hitched his wagon to big data early on, knowing its impact. Poink! That’s the sound of my eyebrow being a-raised so high that it hit the light fitting.
I’m not scared. I think the big data claims are overstated and in a few years, we’ll chuckle. Having said, I’m such a great, big-hearted all-round guy that I can and do change my mind on things – if there is a convincing reason to do so. A meringue?
Another angle on this is who’s to say the Democrats (and the Republicans) weren’t doing this as well. Which leads to the other problem with managements who outsource things they can’t be bothered to do themselves.
What if the Democrats were told from their data analysis results that Hilary was not a good candidate for them but decided the analysis must be flawed, because “Hey, Hilary’s really smart, she’s capable. She has a track record. We don’t like that upstart Sanders. They’re outsiders. They’ve got it wrong.” And similarly the Republicans’ data analysis told them Trump’s views had a lot of support and he needed to be watched, but they pooh-poohed it because “Hah! He’s a bigmouthed upstart who’ll alienate everybody. Just you wait and see. Our preferred candidates have nothing to fear from the likes of him. What do these people know anyway.”
Big organisations don’t like big changes. Huge organisations even more so. They commission reports on all sorts of things, major and minor, all the time. And they often subsequently disregard the recommendations as being too difficult/disruptive/costly/threatening to their own positions.
I think the OCEAN approach to people is interesting, and could buy into the idea of a campaign team selectively releasing adverts with quite different messages to target voters according to their hot buttons. How often are we told that, with Facebook etc, we are the product not the customer? The advertisers must be doing something with the data they harvest – why not sell politicians as well as other products/services?
The dramatic telling of the story makes it more of a page-turner, it’s true. But the authors describe what the Trump campaign actually did, and presumably interviewed Kosinski – some basis for credibility. Where it enters into conjecture is the speculation about the effect of using such profiling. Did it work? Well. Trump does trumpet about the forgotten men and women – could those be the various and contrasting demographics that Steve Bannon and co managed to carefully cultivate to swing a majority in enough states to with the electoral college?
Trump doesn’t need to have been the mastermind behind it all. It could have been a symbiotic relationship exploiting his overblown, unfocused rhetorical style and the backroom boys prompting him with simple messages to throw out.
Here’s an updated link – the original, unofficial translation has been deleted now
http://motherboard.vice.com/read/big-data-cambridge-analytica-brexit-trump
edit:
“win the electoral college”
For what it’s worth I did the Cambridge Analytics FB profile scan thing and if my result is anything to go by none of us has much to worry about. Generic, obvious and vague, reminded me of a horoscope.
The trouble with journalists writing about complex tech stuff is they never understand it. It’s a shame it’s always the arts graduates who have the job of explaining the science graduates’ work to the public.
(By the way news just in: Trump isn’t a genius. This is a guy who doesn’t read and gets his policy ideas from cable news. There’s no way he personally understands the concept of big data, even in the extreme abstract)
I’ve seen some demonstrations of Big Data and its practical outputs in the course of my job. It’s real and it works. There are some extremely smart companies out there, and they’re getting smarter all the time.
I agree and have seen some very clever marketing using these methods. Yet I have also seen many also-rans who claim ginormous success. A couple of questions reveal that their part in these successes were minimal (at best) and their prime motivation is to build up the momentum, increase the value of the company until they actually have to deliver on promises – and then sell it.
The industry is entirely unregulated. You can promise amazing outcomes and walk away before delivery day. I met one youngish guy who claimed substantial personal wealth and involvement in just about any company you can name. I asked him how he could spread himself so widely. His answer was that specialises in startups and without a flicker of self awareness said that he gets bored easily – he pumps up a new company’s profile and then moves on when it gets “boring”. I admire his chutzpah but I know never to deal with him.
It’s not that clever. It continually tries to sell me things I wouldn’t be interested in in a million years, doesn’t try to sell me things anyone who knew me for more than half an hour could guess I’d be interested in, and thinks it is quite likely I’d like to buy something I just bought. Useful for trends but individual targeting….nah. It’s effective in things like prompting in app purchases in games, but I don’t play them anyway.
Yebbut. You have a brain and you make your own mind up. If you’re trying to appeal to those who don’t fall into this category, it’s dynamite.
Anyone who has spent a few hundred thousand dollars on targeted FB advertising will tell you that FB keep the most actionable personal data to themselves. There are many companies selling Cambridge Analytica-alike snake oil and their solutions don’t work
Zuckerberg / Sandberg 2020!
I’m with Monbiot on this: “Looking back on our time, we might see Facebook as one of the worst calamities to have befallen humankind.”
Although of course he said that on Twitter.
I actually think Twitter is worse. Facebook might know more about us (although only as much as we tell it) but Twitter is where we’re starting to evolve away all our social goodness.
It’s weird, it’s like road rage. People who would never have the guts to call you x y or z in person feel perfectly safe to do it from the safety of their keyboard and everyone retreats into their own bubble of I’m right (and not only are you wrong you’re a c*nt and I hate you).
It’s not just Twitter, it happened mildly on here on the Women’s March thread, but Twitter seems to be where it’s most out of control. I don’t tweet myself but I followed lots of people for a while and had to get out, it was doing my head in. They weren’t even celebrities but they were retweeting constantly and it was just such a downer. Really upsetting amounts of hate and negativity.
I honestly wonder whether we wouldn’t be better turning the internet of people off and just keeping it to fridges and Netflix.
Netflix and chill? I’d settle for that
Wahey! I set em up…