The AI-ization of Media Expertise

Artificial intelligence has become an indispensable aspect of media planning. The Serviceplan Group too relies greatly on the power of data science and machine learning. But if you want to engage in media planning that is both creative and effective, you have to employ not only your computer but also people who can offer knowledge and experience in a wide variety of disciplines. Through the concept of the “House of Communication”, the Serviceplan Group enables its experts to work hand in hand in an integrated way. In this interview, Andrea Malgara, Managing Partner of Mediaplus (Germany), and Marcus Ambrus, Managing Partner of Plan.Net Business Intelligence (Germany), explain how AI and media expertise interact.

When will we see the one advertising algorithm that manages and explains everything?

Andrea Malgara: No algorithm can do that – the issues are too complex and technology is just one component among many. Not so long ago, service providers such as Blackwood 7 and SAP XM tried to capture our market through the power of their programmers alone. As we know, it didn’t work. It’s not only about technology but also equally about combining it with experienced experts’ knowledge of the advertising effectiveness can have and the special features of individual types of media: in other words, integrating people and machines. The decisive factor is the willingness to invest properly in knowledge and advertising effectiveness research.

AI is considered to be the new silver bullet in media planning. Is that really the case?

Marcus Ambrus: That’s why we are investing so heavily in AI because it’s where we see huge added value. But this kind of value is different from the one that is usually talked about in public debates. Media agencies have been working with regressions and marketing-mix models for over 20 years. This has always been AI, it just wasn’t called that. What has changed compared to before is the enormous amount of data that we are dealing with today and that can no longer be managed manually. AI’s great achievement is not about developing the one super algorithm that can achieve advertising effectiveness. That’s not the point and it’s not even possible. What is important, however, is productivity. Clearly, AI can turbo-charge productivity, helping us to make faster and better decisions – in planning, during the purchase process, as well as when delivering personalized content in real time.

Malgara: Today, as a media agency, you have to be able to process ever larger quantities of data. If you’re not in control of that, you won’t be able to control the personalization of advertising media that customers demand either. Without smart data, AI, programmatic advertising, personalized configuration of creativity and automated processes, you won’t be able to survive as a media agency.

In Germany, Mediaplus is working very closely with its sister company, Plan.Net Business Intelligence, in this respect. How does this work in practice?

Malgara: Business intelligence is a very important factor in our success. A few years ago we launched Brand Investor, our industry’s first genuinely effective, AI-driven media planning tool. What is even more important, however, is that we are constantly adapting it to the changing demands of the market. Most recently, we expanded Brand Investor to include advertising media such as performance-based, search and out-of-home (OOH). In the case of radio and OOH, we make a distinction between analogue or digital. YouTube is also included. Brand Investor is now used by over 150 clients and controls the communications mix via 19 different channels. But we must stress quite clearly time and again – it’s not a question of technology alone but about the combination of technology and media expertise. NASA obviously has a lot more data analysts than we do, but it would never be able to develop something like Brand Investor. The skill lies in combining our media know-how with automation. In a way, it is about the “AI-izing” of media expertise.

Ambrus: After seven years of experience with Business Intelligence, I feel that the key to success lies in the collaboration between mathematicians and media planners. The better the collaboration, the better the AI applications as well. Our experts have a role to play in helping Mediaplus and Plan.Net win more pitches and expand their existing budgets. There are certainly many factors that contribute to making Mediaplus successful and why it wins a lot of pitches even against Group M & Co, but Business Intelligence is undoubtedly essential. One thing I’m really proud of is that we haven’t lost a single project to Accenture, SAP or anyone else in the last seven years. When it comes to subjects such as software, marketing automation or AI applications, we are the first port of call.

So, the job of media planners is just feeding the machine with data and know-how?

Ambrus: No, media planners don’t feed Brand Investor; they use it. People remain at the helm, defining the goals and evaluating the results. Basically, Brand Investor is like a pocket calculator – and someone has to operate it. AI enables media planners to be able to guarantee much faster and more reliably that a certain plan will have a certain impact. And it allows us to simulate many different scenarios.

How reliable is this “pocket calculator” in practice? How precisely can you predict the impact of marketing campaigns?

Ambrus: Its variance is less than one per cent over three months – that’s really good. Brand Investor is currently the only comprehensive intermedia planning system that I can use to simulate precisely which planning would have the greatest impact on either brand recognition or sales.

Malgara: The computer didn’t do that by itself. Its variance is less than one per cent because so much of our experts’ knowledge and experience has been programmed into the AI.

If the variance is less than one per cent, you already have the perfect planning tool. What's next?

Ambrus: This subject is still much bigger! It’s not just about budget plans but about marketing automation in general. And as far as that goes, to be honest, we’re still pretty much at the beginning, even though we’re much further ahead of the competition on all these issues. Overall, marketing has become incredibly complex: it’s not just about media planning but also, for example, about the interaction of advertising, customer relationship management, corporate websites and pricing policy. AI can help with all of these issues. Brand Investor is an important component but by no means the only one.