Here's how to successfully deploy data-driven marketing

To start on a positive note - many brands have made huge strides when it comes to implementing data-driven marketing. Desire has been turned into execution: data teams have been set up and also the systems have been brought in house to successfully collect, access and process data online and offline. Brands are thus increasingly in control and have valuable knowledge about their customers and prospects. They often use this for marketing activation - reaching and moving consumers 1 on 1 through online and offline channels and/or gathering in-depth dashboard insights to better understand them.

So far so good, you might say, but the last step in the data roadmap turns out to be the trickiest and that is getting returns from that data. I talk to brands a lot and I get back time and again that they feel they are not putting those data and insights to their best use.

Why is that? I think there are a number of misconceptions about data-driven marketing that underlie this observation.

The first misconception is immediately the most persistent - we would like to have perfect data, but perfect data does not exist. We collect data about human behavior, perceptions and motives. But nothing is as complex to explain as human behavior. Data are a means to explain and predict this behavior. If human behavior is difficult to measure, then it is also impossible to collect perfect data.

We would all like to earn back at least as much after that made data investment. Ideally, we want to attach a numerator to that. Fine for the CFO, but also to demonstrate why it is necessary to set up data teams. At such a time, the assumption is that that data euro will pay for itself 1 to 1 and that the effect is immediately reflected in the P&L. However, data investments must be interpreted in and to other ways than looking at linear effects, otherwise it will be settled in the wrong way. How to do it, I explain later in this article.

Following on from the first misconception: use a "glass half full" and not a "glass half empty" view when it comes to valuing data. I still sometimes see a tendency for data departments to dwell on what is not yet working well. This often leads to that 'train not moving' and getting stuck in the development process. It is important to start working with limited data and build use cases, even if it is not yet perfect. This way you gain experience and make 'flying hours'.

This last one is not so much a misconception, it is more of an observation. Many brands have a separate data team that often stands as a separate silo in the organization. Somewhere else are the marketing or brand management teams that ideally will leverage that data knowledge.