Segmentation Engine

The Segmentation Engine

The more granular our view of each customer, the better our understanding of that customer. Our challenge is to balance segmentation with value and complexity.

Typically, organisations use lazy segmentation such as segmenting by age and income. This is a fairly blunt instrument unlikely to result in materially better insights than that obtained by your competitors.

Image

How does it work

Image

Our segmentation engine which provides an automated solution combining machine learning with our experience of conducting segmentation across multiple clients, enables you to utilise the power of machine learning to generate unbiased customer segmentation quickly.

Humans are prone to bias and are generally unable to conceptualise multiple factors influencing what separates groups of customers. All too often this results in simple segmentation based on common heuristics such as age, income and risk score.

By applying machine learning, we can adopt an unbiased approach to segmentation. This enables us to better understand and manage our portfolio's pockets of value and risk.

  • Insurance use case: While working for a digital insurer we discovered a cluster of low-risk high-value customers that also had a low drop-off rate when completing the online application. This cluster could then be targeted for further growth.
  • Banking use case: Applying our tool to a leading retail bank, the resulting clusters provided rich segmentation by risk, value and life stage over and above product holding.

Benefits of Our Segmentation Engine

In addition to providing an initial segmentation of your customer base our segmentation engine creates a score allowing you to allocate customers (existing and future) into segments to enable better management of your business.

Typically we have found the following to be useful:

  • Segment size: By focusing on larger segments the return on management effort has the highest impact.
  • By profiling each segment we better understand and serve each segment improving our value proposition.
  • By measuring risk and value we can:
  • Identify high-value segments and focus on attracting and retaining similar customers.
  • By identifying low value and value destroying segments we can remediate our pricing and service offering to these segments.
  • Identify riskier segments and focus on how these risks can be better mitigated, e.g. increased collateral, lower exposures etc.
  • Drive value generating activities within a segment by identifying and activating lower value customers within a segment

We have extensive experience in helping organisations to create and use customer segments to improve portfolio value. Our tools enable rapid exploration and execution using automated machine learning and best practice workflow – accelerating your path to value.

As existing and new customers can be scored and then allocated to a segment, these segment flags can be used as inputs for various use cases that drive customer value, including:

  • Cross sell (next best offer)
  • Churn
  • Profit optimisation (pricing and portfolio tilt)
  • Risk management (Credit scoring)