By Stéphane Amarsy,

The world of marketing has been radically transformed. Once all about linearity, stability, testing and control, the digital era has made everything today much more complex.  Marketing has become non-linear, unstable, more predictable and entirely measurable.

At the same time, the way companies understand their consumers has evolved.  While consumers were previously seen as acting according to reason and knowledge, their irrationality is finally being recognized. Paradoxically, however, this does not prevent us from predicting their choices. Each consumer must now be understood as a particular entity and relevant techniques, such as mathematics or neuroscience, must be used to pave the way for omnichannel and hyper-personalized communication.


Filter bubbles have their limits

“Freedom of choice” is an important concept to many. But it’s undermined by the modern paradox of having too many choices but not enough time to make them. Algorithms guide our decisions more or less consciously. Taking their lead isn’t laziness; it has more to do with optimizing the three pillars of time, pleasure and proximity.  Once again, the ease with which we accept this virtual “support” is remarkable. And it’s just the start of something that will radically change our relationship with others and our understanding of what guidance entails. All of these approaches, however, have one weak spot: the filter bubble. The more brands learn about a consumer and the more they are offered experiences based on both observed and predicted tastes, desires and behaviors, the more likely they will be gradually locked into a bubble and prevented from having any alternative experiences. As an example, users of on-demand music or movies have found that recommendation engines suggest content that matches their tastes but only in part. As with social media, you only see content close to your filter bubble, which in turn determines your tastes, concerns and habits.


A similar approach is used in the world of online dating, where “matches” follow and resemble one other with increasing predictability. With these applications, spontaneity tends to disappear. The risk of spending countless hours browsing profiles for common ground is that dating has almost become a series of scientific experiments. Not surprisingly, the results can be somewhat underwhelming: “We had so much in common that we had nothing left to say.”


Mistakes can be the key to creativity

While these approaches offer consumers a more personalized experience, they are only effective over a short period. They will inevitably lead to a consumption bubble as consumers become increasingly changeable and fickle. It’s necessary to take risks and make some mistakes to ensure sustainability. On the other hand, mistakes must not be random. They must be strategic and allow for effective measurement so that new consumer needs can be identified. And they must also generate material—data—to help algorithms learn.


According to the way artificial intelligence is perceived today, it’s a nice little inside joke. Artificial intelligence, like every business, avoids risk-taking. This is absurd unless we are willing to accept a loss in terms of experience and knowledge. Mistakes pave the way for opportunities, providing genuinely sustainable operational efficiency. And, if used intelligently, they are the key to creativity.