Traditional Segmentation vs Predictive Segmentation

28 October 2020

Why are traditional segmentations no longer useful for marketing actions?

As you know, segmenting your target audience is fundamental to the success of a campaign. Developing targets is one of the most delicate tasks facing a marketing manager and without doubt, one of the key points in planning marketing and sales strategies.

Traditionally, we carry out the divisions according to 3 or 4 types of segmentation:

  • Geographic: segmentations based on geolocation (country, city, C.P….)
  • Demographic: segmentations based on a person’s age or gender.
  • Socioeconomic: segmentations based on economic capacity or lifestyle.
  • Behavioural: segmentation based on user behaviour.

We use traditional segmentation when we are going to carry out a campaign, whether it is a display, email or telesales campaign, and we produce different messages for each subset. The success of a campaign is mainly based on getting the right message to the members of each group.

This type of segmentation is also used to develop personal shoppers. A personal shopper is the ideal customer to whom we address our advertising messages. It is the typical user for whom we have developed a product or for whom we want to cover a need with our services.

It must be taken into account that, despite suffering variations, the personal shopper is created when we are developing the product, before it even goes to market, and is based on previous market research or according to the target where we have seen a need.


How does AI replace traditional segmentation?

The reason why traditional segmentation is no longer useful is because it’s outdated. Nowadays, thanks to new data processing techniques we can produce much more accurate segmentations. We can even reach 1 to 1 communications with our clients.


Machine Learning technology applied to marketing and sales allows us to predict customer behaviour, so thanks to predictive analytic algorithms we can segment our database according to the interest the customer has for the company.

This is extremely useful for companies that are treating large amount of leads, i.e: Universities and training centres.

This new paradigm outperforms the traditional segmentation method, as it does not distinguish by user profile but by success profile, i.e. we can segment the database according to the campaign objective we set ourselves.

For instance, we can make an initial segmentation by monetary value (CLV – Customer Lifetime Value) we can predict what a client will spend on our business. But we can also segment by low risk (churn rate) and detect those customers who are more likely to drop out.

As you can see, at no time do we talk about the type of customer or user profile. Instead, AI allows us to carry out multivariate analyses (between profiles and behaviour) and convert them into success profiles (purchase, repurchase, churn, registration ….).

The results of conducting campaigns based on success profiles instead of traditional profiles allows companies to increase various metrics (total benefit, conversion, increased brand loyalty or reduced churn, among others). Some studies point to the benefit of using predictive analytics in campaigns always above 15%.

And you, are you still using traditional segmentation?

The results

Get ahead of the competition by contacting the most interested students first. 

+40 %
-25 %
Enrolment costs
enrolment increase united kingdom

Want to see it in action?

Useful articles