planung&analyse - Algorithmus oder Bauchgefühl - 4/2019
The sheer number of touchpoints found along the customer journey today calls for sophisticated touchpoint management. Christoph Spengler and Catherine Ammann of Accelerom explain how an algorithm can help to discover the relevant touchpoints and use them correctly.
Experience shows that companies today manage well over 100 touchpoints along the customer journey. The attempt to be present at all touchpoints is not only a Sisyphean task, it’s also hard to achieve given ever tighter budgets.
But focusing means making decisions. It’s well known that taking monetary decisions based on thin evidence leads to paralyzing uncertainty, and usually the same questions: Are we really using the right analogue and digital touchpoints? Have we allocated sufficient budgets to the various touchpoints?
More and more often, gut feelings and facts diverge widely. For years we’ve been asking our clients the same question: What are the ten most important touchpoints on the customer journey for your customers when they buy your product or a comparable one? It’s sobering to compare the “inside” and “outside” views. On average, only six of the ten most important touchpoints for customers are guessed correctly. What’s the key to overcoming these challenges? Holistic and measurable touchpoint management that takes into account every phase of the customer journey.
The goal of holistic touchpoint management is to focus on the customer. Empirical and systematic evaluation of the touchpoints along the customer journey is essential here. Customers come into contact with a wide variety of touchpoints as they progress along their individual customer journey. The spectrum ranges from posters and web shops to personal consultations.
In order to assess the impact of the different touchpoints, we use uniform units of measurement. This uniform measurement of all touchpoints makes it possible to record and compare the impact of analogue and digital touchpoints, as well as Owned, Paid and Earned touchpoints. Market research provides the database for this approach. In addition to looking at customer behaviour, touchpoints are also queried, which is performed using five different metrics, with each metric representing a phase of the customer journey.
The first two phases of the customer journey deliver information on the passive and active reach of a touchpoint. While passive reach (awareness) answers the question of which touchpoints are involved when customers become aware, active reach (consideration) shows which touchpoints customers use to gather more information. Based on these two reaches, which are given in percentages, it is also possible to calculate the total audience, i.e. the combined reach of several touchpoints.
In the three subsequent phases, information value, transaction value and attractiveness value come into play. These three values are given on a scale of 0 to 100, and their average value represents the relevance of a touchpoint (the touchpoint value). Put simply, relevance corresponds to emotional preference from the customer's point of view, and thus provides information on whether the touchpoint is popular. This approach was developed and validated with the Department of Communication and Media Research at the University of Zurich. While these values reflect the customer view, they only describe the impact of the touchpoint.
Next, it’s clearly interesting to look at implementation, to see what combination of touchpoints can be used to cover the customer journey effectively when geared to a specific target or target group. Certainly not an easy task when you consider that with 80 touchpoints there are theoretically 1024 possible combinations. It’s simply not possible to compare all these possible combinations within a manageable period of time. The solution is to use smart algorithms. Optimal touchpoint mixes differ depending on the goal set. Three elements are taken into account when configuring the algorithm:
The first step is to select the touchpoints. Basically, it makes sense to focus on those touchpoints that can be controlled by the company. Touchpoints that are difficult to influence, such as recommendations from friends, are therefore not taken into account.
The second step is to select the target group for which the optimal touchpoint mix is to be calculated.
Finally, the different phases of the customer journey have to be weighted. For example, if you want to increase brand awareness, the optimal mix will have a stronger weighting of the phases at the beginning of the customer journey. Given that the optimal solution generated by the algorithm cannot always be implemented exactly in daily business, such findings are extremely relevant for corporate management.
We use genetic algorithms to get the optimal touchpoint mix in the most time-efficient manner. Genetic algorithms can determine the best solution from a very large number of possible solutions. Their functionality is based on the Darwinian evolutionary principle. The starting point is a population consisting of several genotypes. Each genotype is a randomly generated variant of a touchpoint mix. Based on the idea of survival of the fittest, selection ensures that the second generation following this first population is just as well adapted, or even better adapted, to the challenges. From generation to generation new mixes are generated, compared and improved, until the optimal mix emerges.
An algorithm-based solution sounds promising. But is this really the best solution? We developed a case study to put the algorithm to the test and answer questions about it. Over the past three years, more than 300 post-grad students in Switzerland and Germany have solved the case study. But none of them have come up with a better solution than the algorithm’s solution.
Another exciting finding is that the students' proposals differ greatly. Few were close to the best solution, and the majority were far away from it. This wide distribution can quickly become a disaster for a company: sometimes the communication strategy works, sometimes it doesn't. It’s impossible to work out the reasons for this retrospectively. But now, thanks to the empirical, algorithm-based touchpoint method, concrete questions can be answered precisely. Strategies and campaigns become not only measurable, but comparable.
Various practical examples show that complexity can be overcome using smart simplification while simultaneously achieving greater market success. No matter whether you’re a financial institution, a transport company, or a retailer: the algorithm can calculate the optimal touchpoint mix for every industry.
This article was originally published in German in planung&analyse 4/2019.