Werbewoche - für Köpfe in der Kommunikation, 2018/06
300 students at Masters level in Germany and Switzerland played with a case study against an algorithm-generated solution. Human or machine, who can develop the best integrated marketing strategies and campaigns?
At its core, the recipe for an integrated marketing strategy is quite simple: select the most suitable touchpoints along the customer journey tailored to the company's objectives and the target group. To ensure that the messages actually appear on the radar of (potential) customers, certain measures must always be provided with sufficient budget. Digitalization and automation will help to significantly increase productivity. So far so good.
Complex and dynamic reality
With increasing digitalization, there has long been a paradigm “be present at all touchpoints". Understandably, this has led to a high level of operational complexity. It gets difficult when budgets are spread over more and more posts - until the activities fall below the effectiveness threshold. This approach also leaves unanswered the questions of effectiveness: doing the right things, and efficiency: doing things right. A recently published study from Vienna (Österreichisches Gallup Institut Dr. Karmasin GmbH, 2018) once again shows the extent of this challenge. When it comes to advertising: "Some 60 percent of marketing managers surveyed said they couldn’t ensure sufficient advertising exposure for all the relevant touchpoints.” Complexity can’t be completely avoided; as Einstein said: "Everything should be as simple as possible, but not simpler.” In the new reality, however, this requires courage and total transparency on impacts.
Seeking the best solution
When we first worked with optimisation algorithms a few years ago, we were led by a very simple idea: "Mirror, mirror on the wall, what’s the best solution of them all?" Specifically, the algorithm had to calculate the best mix of activities to cover the customer journey across the widest variety of digital and analogue touchpoints. We knew very well that we were breaking new ground with this demanding challenge.
At the same time, we were convinced from the start that such a factual basis would make life considerably easier for every decision-maker. The findings not only provide decision-making certainty, but also enable a quantum leap in areas like success management, impact maximisation, risk minimisation, and ROI calculation. Finally we’d have the big picture complete with numbers! Admittedly, offering an algorithm to calculate the "best solution" is quite a claim. Superlative advertising! Bullshit bingo?
Several million options
In order to find the "best solution", i.e. the optimum mix, there are theoretically numerous possible combinations. With 80 touchpoints, there are 1024 combinations. That’s a lot of work. Another central requirement: everything must be measurable and verifiable. This is essential, as the data on which the calculation is based includes qualitative and quantitative assessment criteria for digital and analogue touchpoints from every marketing category: sales, marketing, media and services.
The most burning question
At the end of the day, we’re all interested in one thing above all: is the solution calculated by the algorithm really the best? To put the machine solutions through their paces, we developed a case study based on real data. The task for the participants at the various universities was therefore to develop concrete communication and sales strategies for a product launch. Using a comprehensive system of indicators, the proposals were measured and compared. This ensured that all the mixes could be assessed on a uniform scale.
Human against machine
Since then, over 300 students at Master's level in Germany and Switzerland have solved the above-mentioned case study during their further education studies. You could call it pitting brains and gut instincts against the algorithm. The majority of the students were experienced marketing specialists who had already developed and implemented many strategies and campaigns themselves. Our scenario engine generated key figures on which to take decisions in real time for each selected mix.
And the result? The machine solutions beat all the mix solutions by the participants. In a few cases, participants succeeded in getting very close to the "optimal mix" in purely numerical terms. The scattering loss was considerable across all solutions.
The algorithm passed the acid test
The central task of the algorithm is to show how customer journeys can best be covered comprehensively in a networked "multi-cross-omni-channel world" using an "optimal mix". It shows where there’s potential for improvement, and how this can be exploited in concrete terms. This forms the basis for targeted market development and impact-oriented allocation. Brand experiences can be optimally designed to inspire and retain customers. The result: well-founded and reliable decision-making, with a focus on the really relevant touchpoints.