The Outcome Dashboard

The Problem

The FCA requires firms to evidence good outcomes with data. Most retail financial services firms responded by assembling management information packs that report activity — complaint volumes, call handling times, application turnaround, claims ratios, retention rates — and presenting them to the board annually. The problem is that activity metrics can look healthy while customers are being systematically harmed. A bank can report excellent service levels while its inert savers earn materially less than its active ones. A wealth manager can report a high suitability review completion rate while many of those reviews never actually happen. An insurer can report low complaint volumes while dissatisfied customers simply lapse. The FCA's Year 2 Board Reports review found that many firms continue to present activity metrics dressed up as outcome evidence — and that the gap is what regulators now look for.

The Pattern

The structural move is to distinguish rigorously between what the firm did and what happened to the customer — and to organise the board's attention around the second question:

The structural moves

Outcome translation

For each of the four Consumer Duty outcomes, define the customer-facing question the metric must answer. Not "how many letters did we send?" but "can customers identify the key terms that affect them?" Not "what is the retention rate?" but "is the total cost to customer defensible relative to the value they receive?" Not "how many reviews did we complete?" but "did the review change anything for the customer?" The discipline is in refusing to accept activity as a proxy for outcome

Segmentation that reveals

Aggregate metrics mask the outcomes the Duty is designed to surface. Segment by vulnerability status, distribution channel, product, tenure, and — where relevant — legacy versus new-business cohorts. If long-tenured customers receive materially worse pricing than new ones, that is a signal. If vulnerable customers have equivalent acceptance rates as everyone else, that is evidence. The FCA's cash savings update made the point bluntly: aggregate easy-access rates looked fine; segmented analysis showed inert savers earning materially less than active ones. The value of segmentation is not in the data — it is in what the differences reveal

Leading indicators

The dashboard should surface emerging problems before they become harm. Cancellation within ninety days may indicate mis-selling. A widening gap between expected and actual product usage may indicate value erosion. Complaint themes clustering around a specific feature may indicate comprehension failure. A drop in suitability review delivery may indicate ongoing-service decay. The Duty requires firms to be forward-looking; the dashboard should make that possible

Sensing infrastructure

Outcome data does not arrive passively. The firm must design where insight comes from: post-interaction comprehension checks, claims and outcome experience surveys, exit interview analysis, mystery shopping of advice and onboarding journeys, sampling of suitability review delivery against records. Each of the four outcomes needs a defined sensing mechanism — a deliberate way to detect whether customers are experiencing what the firm intends. Without this, the dashboard reports what the firm already knows rather than what it needs to discover

Indicators It's Working

✓The board receives metrics that answer customer-outcome questions, not operational-activity questions — and can articulate the difference

✓Data is segmented by vulnerability status, channel, product, tenure, and cohort — with narrative explaining what the differences reveal

✓Leading indicators are defined, with documented examples of early intervention triggered by dashboard signals

✓Each of the four outcomes has a defined sensing mechanism — the firm can explain where the data comes from, not just what it shows

✓The annual Consumer Duty board report draws from the dashboard and includes actions taken in response — not just a restatement of the data

In Practice

Worked example 1

A retail bank rebuilt its Consumer Duty board pack after the FCA's Cash Savings Market Review. The original pack reported aggregate easy-access savings rates and showed the firm broadly in line with peers. The new outcome view segmented the book by tenure and engagement: long-tenured, inert savers were earning materially less than recently acquired customers on equivalent products. The same data also surfaced that customers who had moved into financial difficulty were still being targeted by the credit acquisition engine. The board, presented with two segmented charts rather than forty pages of aggregate MI, took two decisions in a single meeting: a tenure-based rate uplift programme, and a suppression rule linking financial difficulty status to marketing eligibility. The activity metric had shown no problem. The segmented outcome metric revealed two.

96%

Worked example 2

A wealth manager, prompted by the FCA's ongoing advice review, redesigned its board MI to separate delivery from outcome. The previous pack reported a 96% "review completion" rate; the new pack reported review delivery (was the review actually held with the client?), review impact (did anything change as a result?), and charge transparency (could clients articulate what they were paying and why?). Sampling against records found that around a fifth of recorded reviews had no evidence of client contact in the period, and that for clients in long-term drawdown the review rarely produced a documented change despite ongoing fees being charged. The board commissioned a remediation review and tied a portion of adviser variable pay to evidenced review impact rather than completion volumes. The activity number had been comfortable. The outcome data was not.

Watch-Outs

The primary failure mode is building the dashboard to satisfy a regulatory requirement rather than to inform decisions. If the board receives it, notes it, and moves on, the dashboard is compliance theatre. A second is data without narrative — presenting charts without explaining what they mean or what the firm should do about them. A third is measuring only what is easy to measure: volumes, turnaround times, and complaint counts are readily available; comprehension quality, value delivery, and outcome adequacy require effort to capture but are what the FCA actually cares about. A fourth is treating internal audit as a tick-box on the dashboard's existence rather than a challenge function on whether the MI is decision-useful — the Chartered IIA's Consumer Duty guidance makes this distinction explicit.

Evidence & AI Lens

E1The FCA has stated that firms must put data at the heart of their response to the Consumer Duty
E2The FCA's Board Report Review (December 2024) found that leading firms presented segmented outcome data — by vulnerability, channel, tenure, and cohort — while laggards relied on aggregate activity metrics that masked poor outcomes for specific segments
E3The FCA's multi-firm work emphasised monitoring comparative outcomes of different customer groups, specifically whether vulnerable and inert customer cohorts receive equivalent outcomes to engaged ones
E4The FCA has made clear that firms unable to evidence good outcomes through data should expect searching questions at best and enforcement at worst

AI & Automation Lens

Machine learning can identify non-obvious correlations that surface outcome risks — connecting early cancellation, lapse, or withdrawal patterns with specific channels, segments, or demographics to flag potential mis-selling or value erosion before complaint volumes rise. Automated anomaly detection can monitor metrics continuously, alerting governance teams when indicators cross thresholds rather than waiting for the next quarterly review. Natural language processing applied to complaint free-text, call transcripts, and adviser file notes can surface emerging themes — comprehension failure, value concerns, undelivered service — long before they show up in structured complaint categories.

The problem, in one line

The move

First signal it's working