By Stéphane Amarsy,
CEO of D-AIM

With the use of data science becoming more mainstream, businesses can now follow an effective marketing strategy by prioritizing closeness with each individual consumer.

Predictive marketing: taking a new approach

Predictive tools are both concerning and fascinating in equal measure. But unlike prophetic predictions, these forecasts are fact-based and can be measured by probability. The magic of predictive marketing is that it moves from intuition—or even irrationality—to data and the probabilities of success. However, predictive marketing demands a radically new approach: Targeting is no longer based on what we think or assume, or even on marketing studies such as surveys and panels, but on the data and its contents.

The more data we have, the more predictable everything will be, which in turn boosts the quality of our forecasts. The objective in marketing is not to achieve certainty in forecasting but to improve upon an empirical approach. The principle is simple. For example, by choosing customers with a high purchase intention towards a certain coat, you will obtain a much higher purchase rate for the product than if you targeted customers based on your experience or targeted all your customers. It also reduces your cost per contact—sonly some customers will be targeted—and prevents customers from becoming annoyed with offers they aren’t interested in. Widespread use of this information will make it possible to build customer intimacy with relevant products and services.

The key point in this approach involves working at the individual level and taking each consumer’s singularity into account. The error—sadly commonplace, even with predictive tools—is when we reconstruct a notion of segmentation to summarize a trend whose projection at the individual level is often wrong. Why would we use such tools merely to reproduce work habits from the past? Instead, we must take the plunge and accept that each client has their own history, desires and possible futures.

 

Redefining the role of the data scientist

Data science is undoubtedly the perfect tool for this approach. Before now, it was regularly seen as the purview of data scientists—the stars of modern marketing. But times have changed. While there will always instances when a human being can do better than AI, time remains a deciding factor. Even the most talented data scientists can only deal with a limited number of predictive models. As such, value creation becomes problematic.  The aim is no longer to use a limited number of predictive models but to quickly increase self-learning and automation. Data scientists can then refocus their time and energy on spotting new opportunities. This twofold approach is not only ethical but also adds value. For example, my company has developed software that lets us do in 15 minutes what a data scientist does in 4 days. Having delegated the production to AI, our data scientists now have more time for carrying out analyses.

 

If you look more closely, you will see that 80% of the customer relationship chain can be entrusted to AI. Sharing tasks related to data science is central to successful customer marketing. Indiscriminately bombarding customers with identical offers has become a thankless, futile and counter-productive task. What’s more, it annoys the consumer inside all of us. Without predictive analytics, the very idea of building closeness in customer relationships would be unfathomable. This transformation won’t be easy. Marketing departments must be willing to change and their teams ready to offer support. In a certain way, they will need to unlearn the old way of doing things and take a new approach where collaboration with algorithms is vital. Sometimes you have to let go and embrace the transformation.