Chasing Big Data Unicorns

This piece appeared in the May 10 print edition of B&T Magazine.

It’s 2013: welcome to the future.

In the past few years the advertising industry has transitioned from handshakes, lunches, and the big idea, to a data-driven, real-time world where endless creative options are dynamically tested, optimised and iterated to make every marketing dollar more accountable and effective. We’re already putting radio on DSPs, and in a few more years TV buying will be fully automated, print will be dead, and the desks of three-quarters of the people reading this will be in the cloud.

Big data is at the heart of this. We’re finally able to take billions of pieces of information and create super accurate models of audiences. We can put the right ad in front of the right person at the right time – and then we can measure how effective we’ve been.

And it’s simple – building audience profiles is now as easy as hitting a button. An example: based on a few basic data points, I can tell the following about you:

“You have a great need for other people to like and admire you. You have a tendency to be critical of yourself. You have a great deal of unused capacity which you have not turned to your advantage. While you have some personality weaknesses, you are generally able to compensate for them. You pride yourself as an independent thinker and do not accept others’ statements without satisfactory proof. At times you are extroverted, affable, sociable, while at other times you are introverted, wary, reserved. Some of your aspirations tend to be pretty unrealistic. Security is one of your major goals in life.”

Welcome to the future. It’s exciting, yes?

At least it would be if it was all true.

Big data has been promising to revolutionise everything in our world for years now, but nobody seems to know exactly how. Behavioural economics professor Dan Ariely summed it up wonderfully - “Big data is like teenage sex: everyone talks about it, nobody really knows how to do it, everyone thinks everyone else is doing it, so everyone claims they are doing it.”

There are countless promises being made but very little actually happening. And even less making a real positive impact to this industry. The description of ‘you’ above wasn’t written by big data. It was written in 1948. It was written by a psychologist named Bertram Forer, and it describes accurately around 85% of people. It highlights perfectly the challenge that big data faces - we’re all unique in our preferences, thoughts, and motivations, yet we’re also exactly the same.

This individual uniqueness challenges the notion that we can use big data to understand precisely how people react to communications. Our understanding of how people form preferences and make purchasing decisions seem to be driven only by the latest theories that are supplanted every six weeks.

The reality is that there is so much data that it’s possible to tell any story you want, as long as no one actually looks behind the curtain. Nobody knows how a print ad or a Facebook Like or a YouTube video view will impact an individual when they’re standing in front of the shelf or hovering over a ‘Buy now’ button. And those who claim to almost always have a vested interest in the outcome they’re championing.

Using data to predict the mindset of people seems to be the unicorn that the industry is chasing, and we’ve somehow fooled ourselves into thinking it’s a possibility.

This stuff is really hard. And repeatedly saying ‘Big Data’ and tossing around a lot of numbers doesn’t make it any easier.

Predicting more rational aspects of a person in the real world is however, a very different proposition to predicting emotional aspects. And this is where the opportunity with data truly lies.

Using data to predict how people act physically is perfectly feasible. Albert-László Barabási’s research into scale-free networks (the kind of social networks in which almost all communications are transmitted) has shown that it’s possible to predict where you will be at 3 p.m. tomorrow with 93% accuracy, as long as we have the historical data of where you have been previously. The same logic correctly posits that cinema ticket sales go up in blistering hot weather, and that if you’ve tweeted about #QandA for the past 4 weeks, you’ll likely be watching again this week.

This isn’t a new idea. It’s just an idea that’s been largely ignored. As an HBR article on Big Data Hype proclaimed last year:

“Activities that are governed by physics and precise laws like the force of gravity can be predicted to an amazing degree. Predictive analytics can figure out how to land on Mars, but not who will buy a Mars bar.”

If big data is actually going to make an impact, we need to be better at identifying the opportunities and being honest about what’s possible. We need to stop looking to data to solve the problems we want it to solve, and understand the problems it’s best suited to solve.

It’s in the real world, the rational, physical, and measurable world that big data is interesting and powerful right now. It’s what we’re seeing in products like Nest and Google Now, and it’s the powerful innovation that the advertising and media industry seems to be ignoring.

At MediaCom lately we’ve been working on a few projects that explore the value of real-time physical data and the power of bringing that data to the point of decision-making.

mTrigger is a tool we’ve built that allows us to switch campaigns on and off, modify search bids and keywords, run mobile offers, and modify website creative in real time based on any external data trigger. We can completely change the experience of a brand in any digital medium when it’s raining, when an Australian cricketer has hit a six, when the trains are running late or even when cat mentions on Twitter by users located in Newcastle happen to be higher than normal.

We’ve also created mBuzzTV, an extension of our in-house social monitoring tool, that’s focused purely on TV shows. The Top 40 shows on TV are constantly tracked and given an mBuzz Score, an indication of the talkability of that show. We’re beginning to use these scores as a predictor of ratings in future weeks, and understanding which people keep coming back to a particular show and why they talk about it.

This use of data is exciting. It’s exciting because it’s real, not theoretical. This stuff is far easier than chasing unicorns, and helps us to get better at making decisions. Because if you can’t connect data to the point at which you make a decision, it’s useless.

By discovering and understanding simple metrics that create decisions and actions that make a measurable and positive difference, data can be powerful, not just big.

- May 2013