The Annual Reckoning

Demonstrating Good Outcomes

Operating Model

All Sectors

FG22/5 mandates an annual board assessment of whether the firm is delivering good outcomes for retail customers. The FCA's December 2024 review of 180 first-year board reports identified pervasive shortcomings: data quality insufficient to justify conclusions, MI without explanations, vulnerability treated as a catch-all, third-party outcomes opaque, and challenge from the board rarely visible in minutes. The Year 2 blog (April 2026) confirmed progress but flagged the same documentation gap — boards approved reports without recording the questions they had asked. The challenge is to design the assessment so it tests evidence rather than ratifies it.
The structural move is to design the annual reckoning as a working governance instrument rather than a calendar event — with prepared evidence framed by outcome, structured cross-line challenge, recorded disagreement, decisions with named owners and timescales, and a documented audit trail of what changed as a result.
Multi-line evidence and prepared challenge
The reckoning is structured around evidence prepared and challenged by all three lines before it reaches the board, not a single function's draft passed up for sign-off. The first line — product, pricing, distribution, customer support — owns the outcome data per outcome and presents it. The second line tests the framework: are thresholds defensible, are vulnerability cohorts properly disaggregated, are distribution chain outcomes evidenced, are conclusions consistent with the MI? The third line — internal audit — provides an independent view on the assessment process itself, consistent with the Chartered IIA's framing that audit's role is to test whether the firm understands, monitors, and responds to outcomes, not to sign off the report. The pre-meeting packs include each line's position; disagreements are surfaced in writing, not absorbed in drafting. The design test: can the firm point to a documented second- or third-line challenge that altered a conclusion, a threshold, or an action in the report before it reached the board?
Outcome-anchored agenda and recorded challenge
The agenda is anchored in the four outcomes — products and services, price and value, consumer understanding, consumer support — with vulnerability and distribution chain treated as cross-cutting lenses rather than separate items. Each outcome is presented with a definition of what good looks like for the cohorts the firm serves, the MI used to evidence it, the threshold rationale, and the cohorts where outcomes diverge from the firm-wide picture. Challenge is structured rather than spontaneous: named non-executives, the Duty Champion (Pattern 29), and the second line carry specific challenge briefs into the meeting. Critically, the minutes record the questions asked, the evidence requested, the disagreements raised, and the follow-up actions agreed — the documentation gap the Year 2 blog flagged as the most common Year 1 weakness. The design test: from the minutes alone, could a successor board, internal audit, or the FCA reconstruct what was actually contested and what was decided?
Decisions, owners, and inter-cycle cadence
Each issue surfaced produces a decision with a named accountable executive, a timescale, and the data that will evidence whether the action delivered the intended outcome. Action plans without owners or timescales were named in both the December 2024 review and the April 2026 blog as a recurring weakness; the discipline is to refuse to close an item without those three elements recorded. Between annual reckonings, a defined cadence — typically quarterly outcome reviews at executive committee or risk committee — tracks delivery, surfaces signals that did not wait for the cycle, and feeds the next reckoning with verified outcomes rather than restated intentions. The annual event then tests not only the current period's MI but the integrity of the previous period's commitments: which decisions were delivered, which were not, what changed for customers as a result, and what new harms or value erosions have entered the picture. The design test: can the firm produce, for any decision in last year's reckoning, the evidence that it was acted on, measured, and either closed or carried forward with a current owner?
- Board minutes for the annual reckoning record the specific challenges raised, the evidence requested, the disagreements surfaced and resolved, and the follow-up actions agreed — not generic statements that the report was reviewed and approved.
- The pack reaching the board includes positions from first, second, and third lines, with disagreements surfaced rather than absorbed in drafting, and at least one documented case in the past two years where second- or third-line challenge altered a conclusion before approval.
- Every action arising from the reckoning carries a named accountable executive, a timescale, and the data that will evidence whether the action delivered its intended customer outcome — the Dec 2024 and April 2026 weaknesses explicitly closed.
- An inter-cycle cadence — typically quarterly outcome reviews — tracks delivery of prior reckoning decisions and surfaces signals between annual events, so the annual reckoning tests verified outcomes rather than restated intentions.
- A retail bank redesigned its second annual Consumer Duty reckoning after an internal review against the FCA's December 2024 publication identified its first-year report as carrying most of the named weaknesses — vulnerability treated as a catch-all, RAG ratings without threshold rationale, action plans without timescales. The redesign moved production out of the central Duty team and into the four outcome owners, with the Duty team coordinating rather than drafting. Each outcome owner prepared a paper anchored on cohort-level data: cash savings outcomes split between front-book and back-book customers (responding directly to the FCA's September 2024 cash savings update), arrears outcomes by vulnerability characteristic, packaged account utilisation by cohort, mortgage retention exits by reason. Internal audit submitted a concurrent paper testing the integrity of the assessment process itself. The board meeting ran on an outcome-by-outcome agenda with the Duty Champion holding a structured challenge brief for each. Two outcomes were sent back for re-work — distribution chain evidence on broker-introduced mortgages and the threshold rationale for the consumer support RAG ratings — and re-tabled at a follow-up session before the report was approved. The minutes recorded eleven specific challenges with named owners and forty-day delivery commitments. The Year 2 board report, when published, included a challenge log as an annex.
- A wealth manager reviewed its first Consumer Duty reckoning and concluded that the central failure was sequencing rather than content. The board had received a comprehensive 120-page pack one week before the meeting, the Duty Champion had submitted written commentary, and the meeting had run for ninety minutes — not enough to interrogate four outcomes, vulnerability, distribution chain, and closed products. The redesign split the assessment across three sittings. A pre-reckoning quarterly meeting in month nine reviewed leading-indicator MI from the dashboard (Pattern 11) and identified the issues likely to demand depth at the annual event. The annual reckoning itself ran across two sessions: an evidence session with second-line and third-line presentation alongside the executive narrative, and a decisions session two weeks later where the board agreed the action plan, owners, and timescales. The change surfaced a fair-value issue the first-year cycle had missed — a cohort of long-tenure clients paying ongoing-advice charges against suitability reviews not delivered in the previous twenty-four months, the same issue the FCA's October 2024 ongoing advice review had flagged at industry level. The board agreed a remediation programme with a defined fee-suspension trigger and outreach plan; the minutes recorded the challenge, the dissent from one non-executive on the proposed trigger threshold, and the resolution. The Year 2 reckoning then tracked delivery against that decision as a standing agenda item.
Common failure modes
The most common failure mode is the report-as-deliverable: the document becomes the work product, the meeting becomes the approval. The Year 2 blog (April 2026) named this directly — most boards reviewed and approved reports, but many did not adequately document the challenge they had provided, leaving no trace of what was actually tested. A second is single-source production. The December 2024 review found reports produced almost solely by Compliance or a dedicated Duty function, missing the scrutiny of business areas with subject matter expertise; the assessment then carries the worldview of the function that drafted it. A third is data without explanation: MI dashboards presented to the board with conclusions claimed but not connected to the underlying numbers — RAG ratings without threshold rationale, vulnerability data treated as a catch-all rather than analysed by specific need. A fourth is action plans without owners: the Dec 2024 review and the Year 2 blog both flagged improvements proposed without timescales, accountable executives, or evidence-of-effectiveness measures, allowing issues to drift between cycles. A fifth is annual-only cadence. The Duty does not mandate quarterly review, but a once-a-year reckoning will detect harm only at the cadence of the report, by which point complaint volumes, supervisory communications, or peer issues have usually moved ahead of the firm. A sixth, particular to year three onward, is the trajectory illusion: the firm reports year-on-year improvement against its own baseline without testing whether the bar itself has risen.

Related Patterns

Evidence & Supporting Facts

FG22/5 chapter 11 and PRIN 2A.8.3-5R require a firm to prepare an annual report on its monitoring under PRIN 2A.9, and oblige the governing body to review and approve the report, confirm compliance, assess strategy consistency, and agree required actions — 'it should be clear that these steps have been followed for each report'.

FCA Consumer Duty Board Reports Review (Dec 2024, 180-firm sample) named five recurring weaknesses: insufficient data quality to justify conclusions, weak distribution chain evidence, inadequate analysis of vulnerable cohorts, board challenge not evident in minutes, and action plans missing timescales, owners, or evidence-of-effectiveness measures.

FCA Year 2 Consumer Duty Board Reports blog (April 2026) confirmed progress on action plans and ownership but identified that 'while most Boards reviewed and approved the reports, many did not adequately document the challenge they had provided' — making it impossible to see how senior leaders had tested the evidence given to them.

Chartered IIA Consumer Duty for Internal Auditors guidance positions third-line challenge as outcome-testing rather than control-testing — auditors should walk in the customer's shoes, use data to understand what actually happens, and assess whether management understands, monitors, and responds to outcomes, providing the assurance backbone for the reckoning's third-line input.

AI Automation Lens

Large language models can be deployed as pre-meeting assurance agents — ingesting the draft board pack and surfacing weaknesses against the FCA's named criteria: claims unsupported by the cited MI, RAG ratings without threshold rationale, vulnerability treated as a single segment rather than analysed by specific need, action plans without owners or timescales, and outcomes asserted but not connected to evidence. NLP applied to the pack against the December 2024 and April 2026 publications can flag any of the named poor-practice patterns and route them back for re-work before the meeting. Machine learning trained on the firm's complaint, claims, and outcome data can produce an independent challenge view of the executive's narrative — surfacing cohorts whose outcomes diverge from the firm-wide picture and presenting them to the Duty Champion (Pattern 29) or non-executive directors before the executive paper does, restoring the time-asymmetry the challenge function needs. Sentiment analysis of the meeting transcript and minutes can test whether challenge raised was actually closed out in the next cycle, surfacing the unresolved-issues backlog the Year 2 blog said most firms could not document. Critically, AI does not replace board judgement, dissent, or consequence — those are governance acts, not technical ones. AI provides the parallel evidence base that makes challenge defensible, and the linguistic discipline that catches the document weaknesses the FCA has now twice put on record.