The Value of Segmentation

Web Analytics combines multiple Micro disciplines under a huge performance umbrella of Analytics – But what exactly is Segmentation within this area?

Segmentation is one of the most important Possibilities within the Analytics-Universe and therefore one of the disciplines that is worked on the most intense and with the highest development power within Analytics vendors-space.

Segmentation has no limits 

There are almost unlimited Segmentation possibilities, meaning combination and filtering of separate metrics or attributes. Before approaching those capabilities in detail, it seems to be important to think about a definition of Segmentation. There are already a lot of diverse sources about a definition (for example on this blog: Segmentation Definition by Thomas Koschel, Head of Analytics at Immobilienscout24 Web Analytics Blog by Thomas Koschel). 

In my point of view:

Segmentation describes the act of separating a digital overall picture in single attributes and the adaptation of one or more of those sub-groups for the use of a more detailed examination.

So, segmentation is a daily support activity for the analysis of current or future marketing actions. Especially with the segmentation of Visitors (New Visitors, Returning Visitors, First Buyers, Repeated Buyers) it is a very useful possibility to work on details and execute special communication and/or acquisition arrangements.

Visitor vs Visits vs Channels

Often, the segmentation in visitor or visits will be used: During a visitor review every visits will be attributed to that visitor. On the other hand the visit review is only checking this a single visit for the segmentation purposes.. Furthermore, often it is segmented to each channel: Entry channels, conversion channels etc.

Segmentation is well established in the Web Analytics world as a necessary and important feature. Therefore, it made me all the more filled with disbelief when I saw the following graphic:

Matrix showing segmentation in Web Analytics

This summary was written by a blogger who builds his knowledge on the content of a report from eConsultancy (Conversion Rate Optimization Report by e-consultancy).  Once we have learned how important and simple segmentation is to achieve desired results, the positioning appears just above the average “value” line and also operates in the area of “unused” attribute yet. According to this report, A / B Testing and customer journey analysis are the most widely used and most valuable disciplines.

I need to ask you two questions:

  1. Are A/B Testing and customer journey so valuable because they are new and so many analysts deal with it, which in turn leads to an exploitation of the benefits of these disciplines?
  2. Why is segmentation so little used, although it a) was available much earlier in the web analytics industry and b) is implemented much simpler than A / B Testing?
Subscribe by Email
About Ralf Haberich

Ralf Haberich, Chief Commerical Officer (CCO) for Webtrekk is globally responsible for Sales & Marketing and in charge of leading the international team.

In addition Ralf holds the honorary position of (co-)chairman for the Digital Analytics Association (DAA, formerly known as WAA). In this role, he drives the activities of the association in the German market, pushing the subject matter Website Analysis in the media, as well as relevant target groups in business and politics. He is the chief editor of the well know German Web Analytics Blog “Web Analytics Europa” and author of the Web Analytics and Business Intelligence book “FUTURE DIGITAL BUSINESS”. He also holds the position of lecturer at the German Online-Marketing Academy and other Universities. Prior to Webtrekk he worked as General Manager for comScore and Marketing Communications Manager at Panasonic.

Learn more about Ralf Haberich
Read Ralf's latest articles
Have a question? Contact Ralf here


  1. Each Web analytics data inference should point to objective. Data needs to be segmented based on the object

  2. I would not agree that segmentation is easy (unless it is a plain technical segmentation e.g. by browser version, time, paid vs unpaid search).
    First, it needs to understand what customer and prospects the business is aiming for, and then to figure out how this is reflected on a website (or other channel) and in the analytics system. At least for websites you have visit based and visitor based segments (one visitor may come by different reasons, and one visit reason can apply for multiple visitor types – are they all the same?).

    Visit based segmentation is easier, since the WA stream data is by nature based on visits. But just think of the increasing “not-provided” keywords from organic searches which make it harder to understand why a visit happened.

    Visitor based segmentation is more complicated (as the (returning) visitor recognition is compared to identifying a visit). Marketing usually has a clear understanding of its customer segments (if not, then they have homework to do..) but getting this captured by WA is not easy. And when integrating different data sources (mobile, VoC, social, ..) this can get far more complex.

    Segmentation requires a thorough process across different stakeholders to match the marketing strategy with the respective digital channel. Then the technical WA setup/implementation has to be done accordingly. Only now meaningful segmented reports are delivered, but analytics and optimizations have not started yet. All this is not coming with a tool, it is additional work and by this (imho) often not done by reasons of time and budget/resources.

Speak Your Mind