Dynamic Segmentation for E-commerce

Dynamic Segmentation for E-commerce

What is Dynamic Segmentation

As we all may guess, online shopping behaviour is quite dynamic in the sense that a visitor – who is even at the very beginning of the purchase funnel – has the chance to visit tens of online stores and view a few hundreds of products. Given the variety and depth of the online product portfolio (Amazon offers 200M products only on its US store), we can imagine that the shoppers starts to experience a ‘choice overload problem’ as the psychology scholars put it.

When the online shopper is bombarded with several product options, he/she tends to dynamically change her shopping behaviour. When she first visits the e-commerce site, she may be looking for a casual dress and can easily end up buying a pair of trousers for her son.

Understanding and profiling the online visitors in this kind of an environment is crucial in converting to shoppers and then ensuring repeat purchases. At this point, dynamic segmentation comes into play! The traditional segmentation methodology locates the users in segments by looking at their previous actions whereas in dynamic segmentation, the segments are formed in real-time and a person could move from one segment to another at every click.

A Use Case on Dynamic Segmentation

Let’s try to elaborate on this with an example. We have Katie here who is searching for Christmas presents for her beloved ones. For this, she visits a department store’s online shop and finds her way through the categories. On Monday, she looks for a present for her mom and on Tuesday morning she views some kids clothing. That same Tuesday evening, she gets an e-mail from a men’s only retailer and visits their site in hopes of finding the perfect present for her husband.

Katie is a 34 year old, college graduate and lives in suburbia. That’s what she stated while signing-up for an e-commerce website. Her demographics stay the same whereas her purchase behaviour changes even a few times a day.

Traditional segmentation models are not enough to capture these behavioural changes in real-time but dynamic segmentation is able to come up with a model in which every single click can be defined and attributed to a new segment. There may be some other customers on this same site who share the same behavioural pattern with that of Katie’s and all these people form a segment and in some cases a segment of one. On a big online retailer, at a given time, thousands of dynamic segments can be formed.

The uses of dynamic segmentation are currently being investigated in the field of e-commerce. When powered with complex event processing and personalization algorithms, dynamic segmentation could translate into a huge success for the online retailers.

Want to Learn More? Talk To One Of Our Personalization Experts


Ali Selim Aytuna

Co-founder @Perzonalization, Problem Solver, Maker by Heart, Nature Lover