What Targeted Support Segmentation Logic Looks Like: Internal Only vs. External Benchmarking

Two Approaches to Segmentation — One Regulatory Standard

Every firm approaching targeted support under PS25/22 faces a fundamental design question: how should we define our consumer segments? The answer determines not just the quality of your targeted support proposition, but its regulatory defensibility.

In this article, we compare two approaches side by side using a pension decumulation example — one using only internal data, the other validated with external cross-provider benchmarks — to illustrate the practical difference.

Scenario: Pension Decumulation Segmentation

A mid-size workplace pension provider with 120,000 members approaching or in decumulation wants to build targeted support segments under COBS 9B. The provider needs to deliver ready-made drawdown suggestions to members who do not take personal advice.

Approach A: Internal Data Only

Using only its own member data, the provider identifies three segments:

The problems with this approach:

• Segments are defined by a single internal variable (pot size) that reflects the provider’s own member mix
• No behavioural differentiation — a cautious 62-year-old with £80,000 gets the same suggestion as an aggressive 55-year-old with £80,000
• The £30,000 and £150,000 thresholds are arbitrary product boundaries, not evidence-based consumer groupings
• No way to demonstrate that consumers in each segment share genuinely “common characteristics” beyond pot size
• No independent evidence to defend the excluding characteristics

Approach B: Internal Data + External Benchmarking

Using internal data validated against cross-provider behavioural benchmarks, the same provider builds a richer segmentation:

Why this approach is stronger:

• Segments are defined by behavioural characteristics, not product thresholds
• Common characteristics are validated against market-wide data, not just the provider’s own book
• Excluding characteristics are evidence-based — each segment is behaviourally distinct in both internal and external data
• Ready-made suggestions can be genuinely tailored to distinct needs rather than one-size-fits-pot-size
• Better position can be measured against market benchmarks for each segment

The Regulatory Difference

When the FCA reviews your targeted support proposition, the quality of your segmentation evidence will determine whether your approach is challenged or accepted. Consider the supervisory dialogue:

FCA question: “How did you determine that consumers in Segment 2 share common characteristics?”

Internal-only answer: “They all have pot sizes between £30,000 and £150,000 in our scheme.”

Externally validated answer: “Cross-provider behavioural data shows that consumers actively modelling drawdown scenarios across multiple providers share a distinct set of characteristics: moderate risk appetite, multi-provider engagement, and active comparison behaviour. Our internal data confirms this pattern within our member base, and external benchmarks validate that these characteristics cluster together across the broader market.”

The difference in defensibility is substantial.

Implementation Considerations

Moving from internal-only to externally validated segmentation requires:

• Data procurement: Sourcing cross-provider behavioural data from an independent provider
• Methodology documentation: Recording how internal and external data were combined to define each segment
• Governance framework: Establishing review cycles to ensure segments remain valid as market behaviour evolves
• Outcome measurement: Building the capability to track consumer outcomes within each segment against external benchmarks