Showing the Work: Documenting Fair Outcomes Under Consumer Duty in 2026
Photo: BLM Nevada, Public domain, via Wikimedia Commons
Every regulated firm now has a Consumer Duty framework. Far fewer can produce, on request, a clear account of how a specific group of customers fared on a specific journey in the past quarter, why that was acceptable, and what changed as a result. That distinction — between having a framework and being able to show it working — is where supervisory attention has settled.
For assistance businesses the challenge is distinctive. The product is not a single transaction but a chain of promises delivered through third parties: helplines, repair networks, replacement suppliers, travel and medical support. The outcomes a customer experiences are produced at the boundary between the firm and its suppliers, often at an hour when nobody from the compliance function is present. Documentation has to reach into that boundary, or it documents very little.
From attestation to evidence
A board paper that states customers are receiving good outcomes is an assertion. Evidence is a sample of files, a test result, a dated record of a control operating, and a note of the decision it informed. The shift from one to the other is largely a matter of discipline rather than resource: the artefacts are not elaborate, but they must exist, be current and be traceable.
A documentation spine
Six artefacts, maintained continuously, will carry most of the weight.
1. Outcome statements. For each product and journey, a plain-language description of what a good outcome looks like — not a restatement of the four outcomes, but a translation into this business: the customer reaches a human within a defined time, an engineer attends when promised, the interim solution holds, the complaint is resolved at first contact where possible.
2. Journey maps with risk points marked. Where in the journey customers are most likely to suffer foreseeable harm, and which controls sit at those points. For assistance, the honest map will show several supplier-controlled steps.
3. A control inventory and testing calendar. Each control, its owner, its test method, its frequency and its last result. Controls that have never been tested should be visible as such.
4. An evidence log. Dated records of tests performed, samples drawn, findings and actions. This is the document a supervisor or an internal audit team will ask for first.
5. Vulnerability and accessibility evidence. How customers in vulnerable circumstances are identified, what adjustments are made, and how effective those adjustments prove to be in practice rather than in policy.
6. A committee reporting pack. A consistent quarterly view of outcomes, issues, remediation and change, with a clear narrative connecting them.
Testing customer understanding
Understanding is not the same as disclosure. A customer who has been sent the terms has not necessarily understood what will happen when the boiler fails at midnight.
Testing is most valuable at the moments where comprehension determines behaviour: at sale, when exclusions and response times are set; at first contact, when expectations about attendance are formed; and at renewal, when the customer decides whether the cover still fits. Practical methods include small-sample interviews with customers shortly after these moments, cognitive testing of written materials with representative groups, mystery shopping across channels including text and relay, and systematic analysis of complaints for language that signals misunderstanding.
A useful discipline is the teach-back test: can the customer explain, in their own words, what they are entitled to and what they must do? If a sample cannot, the communication is failing regardless of what it says.
Metrics that demonstrate fairness
Volume metrics flatter. Fairness is visible in distribution and in the tail.
- Decision times reported as percentiles, not averages, so that the slowest decile is not hidden.
- Acceptance and decline rates, reviewed for unexplained variation by channel, region, age band or vulnerability flag.
- Interim support uptake, showing whether customers were made safe while awaiting a permanent solution.
- Repeat contact rates within a single incident, an indicator of unresolved first contact.
- Complaint uphold rates and redress, with root-cause coding that feeds change.
- Channel parity, comparing outcomes for telephone, written and relay users.
- Post-resolution outcomes, where obtainable: was the repair durable, was the replacement accepted, did the customer need to return.
Each metric should have an owner, a threshold and a stated response when breached. A dashboard without those is decoration.
The paper trap
Documentation invites its own failure mode: an accumulating library that satisfies the file and changes no decision. Three questions provide a filter. Would this document let a reader see the customer's experience, not merely the firm's process? Could a capable new joiner run the control from it? Has it changed a decision in the last twelve months?
Where the answer to all three is no, the material can be retired. Proportionate evidence, current and used, is stronger than comprehensive evidence that is neither.
A governance rhythm
A workable cadence runs monthly at operational level, where test results and issues are reviewed and actions assigned; quarterly at executive committee, where themes, remediation and supplier performance are considered together; and annually at board level, where the assessment of outcomes is signed off and the coming year's testing plan approved. A named board champion should be able to describe, without notes, which customer groups are currently least well served and why.
Consumer Duty does not require an assistance business to be perfect. It requires it to know where it is not, to say so, and to show what it did about it. That is a documentary habit as much as a cultural one — and, done properly, it is the same evidence base that supports fair value assessments, supplier oversight and the group's wider conduct reporting.