Recently, it turned out that at C-level, there’s no other role replaced as often as the CMO. It makes sense. Replacing the Chief Marketing Officer is the perfect illustration of the effect of having too little data and too much gut feeling in the boardroom. Our marketing isn't working; maybe we should try it with another CMO?

Today, marketing has entered a new phase. For years, the (academic) debate about whether marketing is a pseudoscience raged on and on. Many argued it is pseudoscience because you could only give a scientific explanation after doing things you did purely on gut feeling.

These days, things are different. We have a high degree of predictability thanks to machine learning and AI. We've never been so close to certainty. We're moving towards predictability, checking whether a marketing or sales strategy has worked and why. Or why not? The capability to predict makes marketing much more efficient. You waste less money because of the ability to optimise campaigns continuously.

This is one of the reasons why we at iO set up marketing mix modelling for many of our clients. Essentially, we thoroughly audit their marketing investments over the past three years. We look at which campaigns were set up and what impact they had. This way, we can accurately predict how clients can make the best use of their available budget.

The mistake many organisations keep making? They rely on fixed, annual playbooks: campaign X in the summer, campaign Y in the winter. The trick is to go further than just making a few small tweaks to the script. The greater the predictability, the more we dare to suggest radically changing course. And the more data those companies have at their disposal, the more accurate the predictions will become.

Where does predictability begin? With creating order out of chaos. With classic marketing methods, you had just a few KPIs. You could look at how many times a campaign was broadcast on TV or appeared in the newspaper, and how many people those media reached. Or you could measure whether your market share was growing among a certain target group. 

When digital marketing first landed, chaos followed swiftly. Suddenly you had thousands, even millions of data points. Which, out of all those data points, of all those metrics, are the business-related KPIs?

Today’s buzzword: data mesh. For years, everyone wanted one platform, one database where all data comes together. Today, a kind of acceptance has come to be. Giants like Microsoft have been building large data lakes for years, but even they have realised that there is and can be more. So what should you do? Don't just throw everything in one barrel; make connections between all those different barrels.

Let different data sources and methodologies for collecting data coexist, and make sure you put everything in place to let the data flow from one database to another. Everything depends on setup and ambition, but it’s a movement worth committing to.

Data mesh for the automotive industry

That granular data - granular because the data is scattered everywhere like tiny grains - shows in the work iO has done before for large automotive brands. In the past, customers had to fill out a paper survey in the garage, and dealers received maybe five responses a month. Today, millions of customers configure their cars on just about any car brand's website.

Hundreds of thousands of those people give direct feedback on their positive or less positive experiences during configuration. These car manufacturers had years worth of valuable data but did nothing with it. The responses were text in different languages. We used NLP (Natural Language Processing) and a self-teaching computer model to convert that text into data. With the customer responses the car manufacturer receives today, they can optimise the online configurator for tomorrow.

The Data & Intelligence teams at iO align fully with the latest trends and evolutions. In fact, we have D&I teams in all our campuses and every country. Those teams work closely together and share experiences and cases. 

They are all organised similarly, with the same four profiles. 

• You have the data scientist, a kind of mad professor who, using smart models and rapidly evolving AI tooling, ensures that the databases fill with new, intelligent data. That’s crucial, especially when Google retires third-party cookies next year. 

• The data engineers connect the data locations. Their job has become extremely exciting because of data mesh: they solve several thousand-piece puzzles daily.

• The third profile is the data analyst, who extracts the relevant information from all those data locations. 

• The insights consultants translate data into learnings and conclusions. 

Our added value lies in combining these four domains: science, engineering, analytics, and insights. These are often four islands or silos — we’ve built bridges between them.

Data is in many different places today. The faster you can connect those barrels and fill them with the correct data, the stronger your insights. That’s how we help companies build a data advantage - increasing the competitive advantage as we go.