supersieben_logo_2022
coffee_black
coffee_black

CONTACT

Social

Language

© 2024 Supersieben GmbH & Co. KG
Methods & Tools: Target Group Machine

Target Group Machine

Relevant similarities instead of irrelevant differences.
Read more
A data-driven approach to realistic target group segmentation

Who can, should or must my brand talk to?

The classic way to segment target groups works like this: you look for one or more characteristics (age, consumption intensity, gender …) and sort people accordingly. In other words, you think up boxes. And then you divide people into those.
The advantage: At first glance, this always works.
The downside: How do you know if the people you put in a box really belong together? For example, do “heavy users, male, 45+” really form a cohesive group?

Relevant similarities instead of irrelevant differences.

With the target group machine, we have succeeded in creating an algorithm that imitates the natural grouping behavior of people and therefore produces more realistic results.
The Target Group Machine does not look for what separates people from each other, but for what connects them to each other. The algorithm works with up to 400 properties, but instead of forming boxes from them a priori, it determines the combinations that arise all by themselves. Just as people with matching characteristics organize themselves into groups in real life.

This is how it works

Thinking
without boxes.

Up to 400 properties

Which characteristics and attributes should be observed by the algorithm depends on your case: personality type, consumer behavior, favorite color, … everything is possible. The more the better.

1.000 test subjects
We work with national and international online panels. Again, the more the merrier. This makes it easier to analyze subgroups.
n = 1,000 is the minimum.
Algorithm Magic

The algorithm uses Big Data methods to search for statistical twins (and nephews, cousins, and great aunts) among all the traits and their expressions. Groups are formed on the basis of these commonalities.

Name commonalities

We determine what exactly connects these groups with each other with the support of artificial intelligence.

Extrapolate

Since we work with large (and usually also representatively selected) samples, we can extrapolate the results to the population.

The result:
Real-life, real-world target groups

Imagine you could describe someone with 10, 20, 100 or more characteristics. Wouldn’t that give you an extremely precise picture on the basis of which you could develop offers, measures and communication that fit the target group so well in a way that has not been possible before?
All this is provided by the target group machine.