Shopper Analytics vs. Free Will. How much can we predict people’s behavior?

The future is here. The huge amount of shopper data generated every minute(*) in retailers all over the world is allowing Watson-like machines to predict what we are going to buy, where and when, pushing us to buy more, more often. And this is not going to get better. Collected data will include different types of behaviors (not only transactions, but digital interactions, social influence, physical movements in and out store…), and machines will increase their power to a point where by 2030 a $1000 computer will be a thousand times more powerful than a single human brain. 

2030 is just in 15 years’ time. Do you remember where you were 15 years ago? Facebook didn’t exist, the iPhone didn’t exist, the Wikipedia didn’t exist. Our main priority as humans was extending the Year field from 2 characters to 4.

This exponential increase in computational power in the next years will come with an exponential accuracy in behavior prediction, to a point where most of our behavior will be predictable. And if it’s predictable, it’s subject to influence. Where does free will stand, then? Will we lose any power in our purchase decisions?

Fortunately, technology will dramatically change how we purchase today and how we make decisions, because most of them will not be made by us anymore. Why spending 2 hours every weekend in a huge supermarket with wide choice and promotions when most of the purchase is exactly the same than the previous week? Technology can help me identify what products I really need and what are the most relevant offers related to them. Our personal shopping wallet would decide if to take a 3×2 offer, depending on the space available at home and the cash flow and the future expenses of the family. Everything would be optimized to the shoppers’ benefit, and influencing the wallets would be as useless as trying to trick Google for positioning some pages in search results.

It will not be a machine against human battle, the first trying to influence the second, but a machine-machine one, and there, little influence can be done, as decisions will be mathematically taken based on the best possible scenario for the shopper, and the products will be delivered home without the need of spending time navigating the aisles. No aisles, no assortment, no category management, just automated shopping.

On the other hand, we don’t always buy what we need, do we? Sometimes we do buy things we don’t need but we want (certainly we can convince ourselves that we need them). These unplanned, impulsive purchases might be related to a specific event, stimulus or a simple whim.

How a machine can predict a whim? Well, it’ll probably not, unless we are truly connected with machines, something not happening before 2045. In the meantime stimuli will need to be relevant in order to produce the desired effects, and there, Big Data can again provide huge advantages.

Automated Shopping will cut short many of the current predictive analytics strategies, but then shoppers, brands and retailers will have the opportunity tofocus on the sweet spot, the purchases that somehow can’t be so easily predicted, the ones where the value is. For everybody.

Retailers and brands could push innovations to your personal shopping wallet, which would choose the ones that are really relevant to your interests and your profile (not only the purchase one, but your whole persona). Personal wallets will not have any interest but provide you the best options. It’s like someone would filter all the advertisement floating around and delivering the 5% that really fits your interests. With such relevance the chance to buy would increase exponentially.

As technology is evolving so fast, current shopper predictive models will be outdated very soon, not because a lack of accuracy, but because a significant shift in how people buy. We all better get prepared.

(*) In 2014, emnos processed more than 65000 transactions per minute worldwide.