Consumers are as predictable as the weather
Most international blog experts agree: 2015 turned out to be the year of the big data hangover. Every year we generate about as much new data as the total data mass produced by humankind ever before. We evolved from big data to exponential data and our human brain (let alone our human organizations) really no longer can and wants to cope with this. All that is left is a data hangover and disillusionment as to what we can actually do with the increasing data flow. So let’s make it official for 2016: the big data bubble has bursted. There are too many reasons as to the why – in a bit of literature that you can digest in one toilet visit – but here are some considerations.
Too much data, too few insights
In the past few years, many companies, brands and organizations invested astronomical amounts of money in gathering and capturing data. Some had a plan (of what to gather and what not), but the majority didn’t. The main motivation was to stop lagging behind. Or no, even worse, because of the fear of being surpassed by competitors or by new transformational or disruptive challengers. Today, this results in a whole lot of data. The main merit of big data as a corporate trend is that, indeed, it has entailed a lot of data. The talent and brains required to turn this huge pile of data into insights, however, appears to be scarcer. In the end, big data in a consumer context only result in a headache, so it seems. Ending up with potent and actionable insights seems to be exception rather than rule.
Comparable? Don’t think so
Thé promise of big data in a commercial context is increased sales, by smart predictions of what you want, based on how you behave and on the pattern of similar people. Only – and I don’t know if you are experiencing the same thing – I keep getting TripAdvisor and Booking.com suggestions for locations I have just returned from or for places that would never ever even make it onto my shortlist. In the meantime, Spotify is shuffling my very own and carefully selected playlists, in a smart way so to speak: apparently the Swedish music streamer takes into account the beats per minute and the mood per single, but after listening to my playlists for a while, I realize that 20% of my own choices never even make it into the shuffle. Basically, the mechanics behind this are so smart that it becomes stupid because artists with the wrong mood or the incorrect beats per minute are never selected. This also means that those artists make less money as being listened to equals making money. Fact is: I am not comparable with other, comparable, listeners.
People are like the weather
Ever since there have been people, they have tried to predict the weather. In 650 before the Current Era, the weather was forecast in Babylonia, based on cloud patterns and astrology. In 350 BCE, Aristotle wrote Meteorologica (Meteorology), considered to be a standard on weather and nature until the Middle Ages. Back then, weather forecasts were based mainly on studying and recognizing patterns – as they still are today. Since the 19th century, we have adopted a rather professional and scientific approach; humankind keeps track of all possible weather patterns and variables. In other words: 2 centuries of systematic data: Extremely Very Big Data. And what does this result in? When everything is stable, we can forecast the weather 3 days ahead at most; but sometimes even a few hours ahead is a challenge. And above all: despite all weather forecast anchors, rain radars and rain alerts, the weather forecast is wrong 50% of the time. Maybe because, just like humankind, nature is totally unpredictable.
People cannot be boxed
If I say: “British, multimillionaire, male, born in the 40s, married to a famous woman and adulterer, chased by paparazzi, owns several houses across the globe, passionate about alternative medicine and homeopathy, has had books written about him…,” the Big-Data guru or marketer in you must think: what a nicely predictable segment!? Yet… nothing could be further from the truth:
Prince Charles and Ozzy Osbourne in one and the same big-data box?
It is an illusion to think that we will once become predictable beings, based on this kind of segment thinking and the what other people did logic. The essence of big data is to box people, which is not (or no longer) possible.
That’s it for what we have all seen coming for a while now. But then, what could do the trick? What could we do to avoid a data hangover?
The self-curating user
Step 1 is to respect the human, personal and self-composed choice. Most of today’s big data models start from a top-down model which calculates in the background what may or may not (conditional tense) be suited for you. A few failed suggestions later, you, the experienced surfer, no longer pay attention. When I add friends on Facebook, I do so to see their status updates in my newsfeed, not to let Facebook use big data formulas to decide which of their updates should end up in my feed or not (having to use these conditional tenses is the start of the decline for big data). I’d rather have the power to compose, adapt etc. my own newsfeed and to program my own algorithm: “Just give me everything Philip posts, the music updates and tips added by Eve and the ‘almost weekend’ updates by Aykan and Ben. And I would like to always get Bruno’s and Loutfi’s pictures at the top too”. We should be more respectful towards one’s own online choices. If I choose to add a number to my playlist, I want as much chance of hearing that song as all the other songs. No need for expensive and complicated formulas. The power should go to the self-curating and self-programming user.
Mood and context interfaces
But of course it’s not that simple. For example: lately I’m having trouble deciding what could be a great city trip for me. I’ve been on about 20 in the past 10 years and it’s hard to find destination number 21, as the word-of-mouth around me has long since dried out and is exhausted now. But rather than the clumsy and idiotic recommendations by TripAdvisor, booking.com and other trip providers, I want to enter my wishes in an interface: simple and fun for me to do, intelligent and supported by smart big data for the provider who can offer it. And nope, I’m not referring to the need for a TV in my hotel room and/or a sauna or swimming pool nearby. I’m talking about emotion, context, mood, ambiance, offer, flight duration… I want destinations offered to me based on my mood of the moment, my expectations for that particular weekend, my definition of a cool destination for that time in that particular company. Something which is unpredictable (I would even find that prediction unpleasant), even for my friends or household members.
From big data to (Artificial) Intelligent Data
Artificial Intelligence, as applied for example in robotics since the 50s, offers a lot more options, of course. In essence it is not top-down at all. It’s all about a two-way relationship between man and machine. It is not unthinkable that, in a (maybe not so) near future, we have personal assistant robots who will thoroughly filter the data world (read: the Internet of things or the cloud), based on artificial intelligence rather than on algorithms. Artificial intelligence leaves the option to learn from each other, which brings us in a situation where man and robot grow closer. Predictability will of course become somewhat more feasible. A bit like how our friends are good judges of which concert, film or teambuilding activity we may like. We’re almost there, but (really) not quite yet.
From capturing to curating
In the meantime, the challenge for companies which gather a lot of data today lies in translating the data into insights for tomorrow. It’s about time for companies to really evaluate the quality of their insights and the people who are responsible for them. Big data must be freed from the IT silo and there is no time to waste. So in the coming years, the investment should not so much go to capturing data as to curating data across different departments. The first step is to get to insights, which is possible by keeping an open mind, contracting a mix of specialists and generalists and getting in touch with experts on the matter. And mainly also by giving more levels of freedom to the users or consumers who should be in a position to help set the algorithms. Furthermore the insights should also be distributed and used, of course. They need to create an impact and be actionable.
And don’t forget: Francis Bacon’s ever valid and current aphorism is ‘Scientia potenta est’, not ‘Data potenta est’.