How to Build Consumer Segmentation for Targeted Support: A Step-by-Step Guide

Building Segments That Survive Regulatory Scrutiny

The targeted support regime under PS25/22 requires firms to deliver ready-made suggestions to consumers who do not receive personal advice. The foundation of any targeted support proposition is segmentation — grouping consumers by shared characteristics so that each segment receives appropriately tailored suggestions.

This guide walks through the practical steps of building consumer segments that meet COBS 9B requirements, from initial hypothesis through to ongoing validation.

Step 1: Define Your Segmentation Objectives

Before touching any data, be clear about what your segments need to achieve:

• Regulatory purpose: Each segment must receive a ready-made suggestion that puts the consumers in that segment in a “better position” than they would be without it
• Practical scope: Define which products or decisions your targeted support will cover (e.g., drawdown strategy, fund selection, contribution levels)
• Segment count: Too few segments means suggestions are too generic; too many creates operational complexity without proportionate benefit. Most firms find 4–8 segments per product area is a practical range

Step 2: Identify Candidate Characteristics

COBS 9B requires segments to be defined by common characteristics (what consumers in the segment share) and excluding characteristics (what distinguishes them from other segments). Candidate characteristics typically include:

The last category — cross-provider characteristics — is where external benchmarking becomes essential. Your internal data can inform demographic, product, and risk characteristics, but cannot capture how consumers behave across the wider market.

Step 3: Analyse Internal Data

Start with your own client data to build initial segment hypotheses:

1. Cluster analysis: Use statistical clustering (k-means, hierarchical, or latent class) on your client characteristics to identify natural groupings
2. Behavioural profiling: Examine transaction patterns, product usage, and engagement metrics within each preliminary cluster
3. Gap identification: Note which characteristics you can observe internally and which require external data

At this stage, your segments are hypotheses — they reflect patterns in your own data but have not been validated against the broader market.

Step 4: Validate with External Benchmarks

This is the critical step that transforms internal hypotheses into defensible segments. Using cross-provider behavioural data:

• Test common characteristics: Do consumers who share these characteristics in your book also cluster together in market-wide data? If your internal “conservative drawdown” segment looks similar in cross-provider data, the common characteristics are validated
• Test excluding characteristics: Do your segment boundaries actually separate distinct behavioural groups in the broader market, or are the boundaries artefacts of your own product design?
• Identify missing segments: Does the external data reveal consumer groups that exist in the market but are absent from your book? These may represent underserved needs
• Calibrate boundaries: Adjust segment thresholds (e.g., pot size breakpoints, age boundaries) using market-wide distributions rather than your own client distribution

Step 5: Design Ready-Made Suggestions

For each validated segment, design a ready-made suggestion that is:

• Appropriate: Suited to the common characteristics of the segment
• Differentiated: Distinct from suggestions offered to other segments, reflecting the excluding characteristics
• Measurable: Structured so you can track whether consumers who follow the suggestion end up in a better position

Step 6: Establish Outcome Monitoring

COBS 9B requires ongoing evidence that targeted support delivers better outcomes. Your monitoring framework should include:

• Internal outcome tracking: Measure consumer outcomes within each segment (e.g., fund performance, drawdown sustainability, protection coverage)
• External benchmarking: Compare outcomes against market-wide benchmarks to demonstrate “better position” relative to consumers who did not receive targeted support
• Segment stability monitoring: Track whether segment compositions remain stable or are drifting, which would indicate the need to recalibrate

Common Pitfalls to Avoid

• Product-led segmentation: Building segments around your products rather than consumer needs creates a circular argument the FCA will challenge
• Static segments: Segments defined once and never revisited will drift from reality. Build in regular review cycles
• Ignoring cross-provider behaviour: Segments that only reflect behaviour within your firm cannot demonstrate market-wide validity
• Insufficient documentation: Every methodological decision should be documented and defensible. If you cannot explain why a segment boundary is where it is, the FCA will question it