Consumer Duty Analytics

The FCA's Consumer Duty requires firms to prove they deliver good outcomes, not just intend them. That proof lives in data: product performance, pricing fairness, customer understanding, and support quality. Consumer Duty analytics is the practice of assembling that evidence, identifying where outcomes fall short, and acting before the regulator asks. The firms doing this well have turned a compliance obligation into an operational advantage.

What is Consumer Duty analytics?

Consumer Duty analytics is the measurement and monitoring of customer outcomes across the four pillars of the FCA's Consumer Duty: products and services, price and value, consumer understanding, and consumer support. It combines quantitative metrics (complaint rates, claims ratios, service response times) with qualitative analysis (customer feedback sentiment, communication clarity, vulnerability identification) to produce a picture of whether the firm is meeting its obligations.

The Duty, effective from July 2023 for open products and July 2024 for closed books, requires firms to monitor outcomes at product level and act when they identify harm or risk of harm. This is not a one-off assessment. It is continuous monitoring with board-level reporting. The annual board report on Consumer Duty outcomes, now a regulatory expectation, requires data that most firms did not collect before the Duty came into force.

The challenge is aggregation. Outcome data sits across multiple systems: complaints in the CRM, pricing in the product engine, claims data in the claims platform, customer feedback in survey tools, vulnerability flags in the contact centre. No single system holds the complete picture. Consumer Duty analytics is, at its core, a data integration problem dressed in regulatory clothing.

The landscape

The FCA's supervisory approach focuses on outcomes data. In its first year of enforcement, the regulator asked firms to demonstrate, with evidence, how they identified and addressed poor outcomes. Firms that could produce dashboards showing outcome metrics by product, customer segment, and time period fared better than those relying on narrative assertions. The expectation is clear: show the data, not the policy.

Closed book products present a particular challenge. Firms must now assess whether legacy products, some designed decades ago, deliver fair value to customers who remain on them. This requires analysing pricing relative to current market conditions, identifying customers who would benefit from switching, and communicating proactively. The Financial Ombudsman Service upheld rates on closed book complaints provide a useful benchmark for whether outcomes are acceptable. The data archaeology required to assess products that predate modern systems is substantial.

Peer comparison is becoming a supervisory tool. The FCA collects outcome data across the market and can identify outliers. A firm whose complaint rate, FOS upheld rate, or claims rejection rate significantly exceeds the market average will attract attention. Firms need analytics that benchmarks their performance against available market data, not just against their own historical baseline. The connection to regulatory reporting automation is direct: the same data feeds both the internal dashboard and the regulatory return.

How AI changes this

Natural language processing analyses customer communications at scale to measure "consumer understanding," the Duty's third outcome. Are customers understanding the product terms, the charges, the risks? AI analyses customer queries, complaints, and feedback to identify where confusion clusters. A spike in questions about a specific fee structure indicates a communication failure, not a customer failure. This insight is impossible to extract from structured data alone.

Vulnerability detection across channels identifies customers who may need additional support. AI systems analyse call transcripts, chat logs, and interaction patterns to flag indicators of vulnerability: financial distress, cognitive difficulty, bereavement, or health issues. This supports the fourth outcome (consumer support) and creates an audit trail demonstrating the firm's proactive approach. The same capability feeds into complaint analytics and contact centre AI systems.

Outcome prediction models identify products or customer segments at risk of poor outcomes before the harm materialises. A product with declining claims acceptance rates, rising complaints, and deteriorating value relative to market alternatives can be flagged for review before customers experience the poor outcome. This shifts the firm from reactive remediation to proactive product management.

Automated reporting assembles the board-level Consumer Duty report from underlying data sources, reducing the manual effort of a process that currently takes many firms weeks of analyst time per quarter. The report template is standardised, the data pipelines are automated, and the narrative is drafted from the data rather than written from memory.

What to know before you start

Define your outcome metrics before selecting any technology. The FCA does not prescribe specific metrics, which means each firm must decide what "good outcomes" look like for its products and customer base. This is a business and compliance exercise, not a data exercise. A dashboard without clearly defined metrics is just a collection of numbers.

Data integration is the hard part. Customer outcome data spans complaints, claims, pricing, product usage, and communication effectiveness. These typically sit in different systems owned by different teams. The Consumer Duty analytics platform must pull from all of them, which means either building a data warehouse or deploying a federation layer. Budget for the plumbing, not just the dashboards.

The board report is the forcing function. Work backwards from what the board needs to see (outcomes by product, actions taken, emerging risks) and build the data pipeline to support it. Many firms made the mistake of building analytics capabilities bottom-up and then discovering they could not produce the board report the FCA expects. Starting from the report ensures every data investment serves the regulatory purpose.

Start with your highest-risk product line: the one with the most complaints, the highest FOS referral rate, or the most complex pricing structure. Build the outcome monitoring for that product, prove it works, and then extend to the rest of the portfolio. Ensure that the underlying data governance is sound, because outcome metrics drawn from inconsistent or incomplete data will produce misleading results. A firm-wide Consumer Duty analytics platform is a multi-year programme. A single-product proof of concept is a quarter.