What is a next-gen customer segmentation?
Customer segmentation is the process of splitting customers into groups based on shared characteristics (usually their demographics and behaviours) so that their experience of the service can be more personalised and they can be marketed to more effectively. It is the best way to avoid a service that assumes one-size-fits-all.
Our approach is different because:
- We use LLM and NLP-powered Information Extraction to pull the most informative features possible from behavioural data. If (with their consent) we can analyse what your customer is reading, watching, listening to and talking about, then we know so much more about their tastes and preferences than we can ever know from plain demographics.
- We use the latest machine learning techniques to first simplify your data and then identify customer groupings.
- Unlike segmentations that are built from surveys, and which then have to be linked in some unsatisfactory manner to your database, our segmentations are always built directly in your database.
- Our segmentations change over time and adapt intelligently to the amount of data that you have for each person.
Why would you need it?
A segmentation is an ideal starting point for building up a data science capabilities as it can be reused in many places. Some examples are:
- As a communication tool within the business (reports can be broken down by segment, discussions can be structured using the segments - “I think we should be targeting the X segment with this new newsletter”)
- As a way of providing a targeted response to customers, e.g. tailoring the content of email communications by segment
- As variables in propensity models and in the customer lifetime models
- As strata for more advanced testing
- As a way of understanding the different customer life cycles (segment X tend to subscribe in response to offers but will cancel once the full price is charged.)
- To profile customers for partners or advertisers.
In addition to this the process of creating the segmentation also provides as a bi-product a lot of useful insight into customer behaviour
