Outcomes monitoring: good practice and areas for improvement

Firms with structured monitoring frameworks were able to be clear about how to produce better outcomes for their customers. The most effective approaches had clear definitions of what good customer outcomes look like in practice.

Using data and MI effectively means firms can identify risks and make the right decisions to address them. The stronger approaches produced clear evidence of how firms’ analysis helped them to improve customer outcomes.

Good practice involves looking at specific details of different parts of the journey, rather than simply having high-level indicators of overall outcomes. This helps firms build a clearer picture of how outcomes develop, and identify where customers may be at risk of harm.

Firms should have clear governance arrangements to oversee and challenge outcomes, including regular reviews of information and effective escalation routes. Stronger approaches show how these elements lead to improvements over time.

Firms also need to consider whether poor outcomes arise elsewhere in the distribution chain and work with relevant firms where appropriate.

Clear and practical definitions

One firm defined good outcomes for each of its key products, rather than having a single broad statement covering all the products. It described what customers should experience at key stages, such as joining, using the service, and leaving. The firm linked these statements directly to its customer journeys, so it could see where poor outcomes could happen.

This approach means firms can specifically check outcomes at each stage of a customer’s journey, rather than relying on high-level indicators.

Some firms set out the key risks of customer harm in their products and services, and used these to set clear outcome expectations. They translated these risks into specific, practical definitions for each product, with indicators such as complaints, attrition and compliance monitoring results. They used these indicators to check if they were delivering good outcomes, and focused their monitoring and testing where harm was more likely.

This approach gives firms measurable definitions of outcomes, so they can clearly see where they need to make changes.

Metrics, thresholds and ongoing reviews

One firm used its past performance to set targets for tracking customer outcomes. These targets were not fixed; the firm reviewed and challenged them through governance to make sure they were not too easy to meet. It monitored a defined set of measures across areas such as customer experience, complaints and operational performance. It reviewed its targets at least once a year, adjusting them where performance improved or a measure no longer reflected the most relevant risks. In some cases, it set targets above historic performance to encourage improvement.

This approach helps monitoring to remain relevant. By reviewing and updating targets, firms can better judge whether:

• Outcomes are genuinely improving.
• Their measures still reflect where customers may be at risk of harm.

One firm did not rely on a single measure but used a combination of information to assess customer outcomes. This included customer feedback, complaints and performance information, alongside outcomes-based testing. It reviewed these sources regularly.

This gave the firm a more complete picture of customer outcomes, so it could see where it might need to do more analysis or make changes. This approach helps firms spot issues and address them earlier.

Structured and joined-up frameworks

Some firms operated their different monitoring activities as part of a single framework. They had regular governance discussions about different areas such as complaints handling, service reviews and testing. When they identified issues, the firms followed a structured process to understand what had gone wrong and agree actions. They tracked these actions and had governance updates to check that changes had been implemented and were working.

This approach reduces the risk of firms identifying issues but not resolving them, leading to a consistent and coordinated response.

In another example, firms focused on how they share responsibility for monitoring and oversight across the business and its partners. They set out who is responsible for reviewing information, making decisions and taking action, aligned this with their product and risk governance.

They also considered information from third parties, including brokers and service providers involved in delivering the product. This is so they could identify risks effecting customer outcomes wherever they arose. Where issues were linked to external partners, firms had processes in place to engage with them and follow up on improvements.

This approach helps ensure that monitoring is complete and consistent across internal teams and external partners. Firms reduce the risk of gaps in oversight and are better able to manage outcomes across the full customer journey.

In contrast, some firms’ monitoring frameworks were not clearly focused on customer outcomes or the risks of harm. While firms often described monitoring at a high level, they did not have a clear structure for identifying poor outcomes or understanding why they happened.

These weaknesses limit firms’ ability to identify risks of harm and assess whether they are delivering good outcomes (PRIN 2A.9.2). Frameworks that are not aligned to customer journeys or clearly defined outcomes are less effective at identifying emerging issues or supporting timely action. This falls short of our expectation that firms use monitoring to identify and address poor outcomes in practice. This can increase the risk that firms do not identify poor outcomes for customers, or address them promptly. (PRIN 2A.9.11R; FG22/5 Chapter 11).

High-level frameworks without clear outcome definitions

One firm described a broad outcomes framework but did not clearly define good and poor outcomes for key customer journeys, or have evidence for the thresholds it used to assess them. This can make it harder to make consistent, outcome-based decisions.

Some firms had a broad definition of good customer outcomes and a list of performance indicators. However, they did not clearly explain what good outcomes look like at each stage of the customer journey, or how each metric demonstrated those outcomes. This can make it difficult to show whether outcomes are good or bad.

Weak links between outcomes, metrics and customer journeys

Some firms used operational activity metrics, like conversion rates or review completion, as a proxy for customer outcomes – but did not define good or poor outcomes for customers at different stages of the journey.

Limited coverage of customer journeys and customer groups

Some firms defined good customer outcomes at a broad level but did not clearly link them to key stages of the customer journey. There was also limited evidence of how outcomes differ across customer groups, including people in vulnerable circumstances. This can make it harder to know where customers may have worse outcomes or need more support.

Some firms had a range of monitoring measures, but did not know if different customer groups had good outcomes at different stages because they had not defined those outcomes.

Indicators and thresholds for foreseeable harm and risk

One firm tested rejected-applicant data to see if its distribution channels were reaching its target market. It found that some channels were producing high levels of unsuitable applicants who did not have enough income or savings. The firm ended 2 paid affiliate relationships as a result. This was a proactive approach, identifying an issue and taking prompt action to reduce the risk of customers being offered unsuitable high-risk products.

Another firm offering high-risk products used a financial vulnerability indicator to identify customers whose net deposits were high compared to their declared income. It reviewed customers above its internal threshold individually, considering their wider financial profile, declared wealth and trading history. It tailored its response for each customer, including:

• Closing accounts where customers borrowed to fund their trading.
• Doing further reviews where the wider circumstances suggested lower risk.
• Checking customers’ welfare where frequent deposits raised concerns.

This helped the firm identify customers at greater risk of harm and respond effectively and proportionately.

MI showing customer friction and improvements in customer journeys

One firm’s customer feedback and operational MI showed that customers wanted a more direct way to raise queries outside the formal complaints process. The firm introduced in-app chat so customers could speak to someone and get an immediate response. It then expanded the chat functionality using AI to route customers to the appropriate support team. The AI chatbot also used keyword recognition to help identify indicators of potential vulnerability. When it found these indicators, it escalated cases to the right team for prioritised handling.

The firm monitored the impact through customer support MI: average first response time reduced from 22 hours to under 2 minutes over a 6-month period, and average resolution time reduced from 4 days to under 3 hours. This shows how firms can use customer contact data to improve access to support, route queries more effectively, and identify customers who may need additional support earlier.

One firm’s complaints MI showed that some customers experienced delays when they provided original or certified identity documents by post to complete withdrawals. It piloted a multi-bureau verification approach before implementing it fully. The firm monitored the impact through committee and Board reporting, and found a 5% improvement in anti-money laundering pass rates and a 20% improvement in bank verification pass rates. This shows how firms can use complaints MI to identify friction in a key customer journey, support a targeted process change and monitor outcomes to make sure that change has worked.

Improved data, testing and AI

One firm piloted synthetic testing to identify where communications may confuse or mislead different customer groups. The AI-based tool assessed communications against themes including vulnerability, fair treatment, root cause indicators and the post-advert customer experience. Outputs included a comprehension score, engagement heatmap, risk matrix and recommendations for improvement. The firm tested the tool on a handful of communications that had already been reviewed through human testing, covering awareness, purchase and onboarding stages. It reported a high average match rate, with the AI tool’s outputs largely aligned with the human-testing findings.

One firm used transaction-level MI and customer testing to identify that some payments were being categorised incorrectly. This included HMRC payments being placed in the wrong category and gambling-related transactions not always being recognised consistently. The firm tested sample transaction descriptions against external reference data, refined its categorisation rules and removed false positives. After making these changes, it recorded a 12.8% uplift in categorisation accuracy based on transaction descriptions. This helped improve the quality of the MI the firm used to monitor customer outcomes, by giving it a more reliable view of the types of transactions customers were making and where further review or action may be needed.

Some firms had a narrow or reactive set of indicators, making it harder to identify risks or assess customer outcomes. They often collected a range of data, but did not consistently show how they used it to anticipate issues, track outcomes or assess changes they made.

In particular, firms sometimes relied on lagging indicators, lacked clear thresholds or forward‑looking metrics, or did not have complete audit trails from identifying an issue through to action and outcome. In some cases, inconsistencies or gaps in key data limited firms' ability to reliably evidence customer outcomes. This falls short of our expectation for firms to use appropriate, reliable and actionable information to understand and improve customer outcomes (FG22/5 Chapter 11).

No evidence of how MI leads to decisions and improved outcomes

Some firms collected relevant MI but could not show how it helped them make decisions or improve customer outcomes. For example, one firm said it monitored onboarding redesign, payment instruction clarity and high-cash or orphaned clients. However, it did not consistently show what MI it considered, where it discussed or challenged the MI, what decisions it took, when it took action, or how it assessed the impact. The firm’s own reporting also noted that it should keep better records of MI trends, actions and links to customer outcomes.

Stronger examples included a clear audit trail showing how MI translated into decisions, actions and improved outcomes.

Unclear reasons for indicators, thresholds and tolerances

Many firms set specific thresholds but did not clearly explain how they indicated good or poor outcomes. For example, a firm set thresholds for complaints, file review pass rates, client retention and mortgage review engagement. However, it could not consistently explain what it had based these thresholds on. The stronger examples explained why each threshold was meaningful for identifying potential harm.

Lack of testing for remedies

Some firms found friction in their customer support and agreed remedial action, but some issues persisted. For example, one firm identified unclear LiveChat interactions, verification delays, repeat contact and inconsistent first-contact resolution. It agreed a set of actions including clearer expectation-setting, stronger ownership, improved escalation handling and staff training. However, later evidence showed the firm was still passing customers between customer service agents, and did not always resolve complex issues first time.

Evidence was stronger where firms tested whether interventions reduced repeat contact, avoidable customer effort and unresolved journeys, rather than relying only on the fact that they had agreed some actions.

Quality, reliability and granularity of outcomes MI

Some firms monitored outcomes for customers in vulnerable circumstances separately to other customers, but did not segment those outcomes by vulnerability drivers. For example, one firm aggregated its vulnerability MI rather than splitting it by drivers such as health, financial resilience or life events. It said low volumes made further segmentation challenging, but it had identified this as an improvement area.

Aggregated MI can make it harder to identify whether different groups of customers with characteristics of vulnerability have different needs or experience different barriers and outcomes. Stronger evidence showed how firms tested whether different customer groups received appropriate support.

Response to poor outcomes, and supporting staff capability

Some firms used governance and oversight arrangements to identify gaps in frontline staff performance. They used quality assurance (QA) reviews, complaints analysis and service-level MI to see where customers were not receiving the expected outcomes. This included QA breach categories and QA scoring approaches that highlighted incorrect customer outcomes, providing a clear trigger for further review and action. Where they found issues, firms took corrective action such as reopening and reassessing cases, contacting affected customers, correcting errors, and making changes to processes or service delivery.

Firms also took action to address staff capability where this was a cause of poor outcomes. This included removing underperforming agents, increasing call monitoring, introducing escalation processes, and providing targeted training, clearer expectations and ongoing supervision.

Firms used the same reporting processes to evaluate whether their changes were improving frontline staff performance and customer outcomes.

Governance, oversight and cultural practices were not consistent across the firms we reviewed. In particular, it was not always clear how arrangements operated end-to-end, from identifying issues through to testing whether actions have improved outcomes.

Under the Duty, firms are expected to monitor outcomes, identify risks or instances of harm, understand their causes and take appropriate steps to address them (PRIN 2A.9). Governing bodies are expected to review this information, provide appropriate challenge and agree actions where they identify risks to customers (PRIN 2A.8.4R; PRIN 2A.8.5R)

If firms cannot clearly show how they identify, escalate and act on issues, it is hard for them to demonstrate that governance arrangements support good customer outcomes.

The following areas highlight where firms typically need to strengthen their approach.

Unclear governance frameworks

Many firms describe clear governance structures with committees, reporting lines and escalation routes. This shows that they have spent time putting the right frameworks in place.

However, it is often less clear how these arrangements work day-to-day. Firms tend to describe what exists, rather than how they use it. It is not always clear how issues move through governance, how the firm makes decisions, or how they lead to better outcomes for customers.

A similar issue comes up in how firms oversee third parties. Firms often refer to regular engagement, but it is not always clear how they track outcomes relevant to their activities across these relationships, or how this leads to appropriate action.  

Limited challenge and direction from the board

Boards are usually given regular updates on customer outcomes and are described as central to oversight – but it is not always clear how they use this information. In many cases, they focus on reviewing and approving reports rather than challenging them or pushing for further action.

Because of this, it can be hard to see how boards shape decisions or drive improvements.  

Actions are taken, but firms do not explain why or whether they worked  

Firms often show that they spot problems and take action. This is a positive sign that monitoring is leading to change.

At the same time, there is often less detail on what sits behind those actions. It is not always clear what caused the issue in the first place or whether the actions actually fixed it.

This makes it harder to see if firms have clearly identified a problem, understood it, and checked that the solution has worked.  

Culture described and supported by training, but not reflected in behaviour or incentives

Many firms place a strong emphasis on culture. They usually support this with training and internal communication about the Duty. But this often amounts to describing a culture at a high level, rather it being how the organisation operates in practice.  

There is often no evidence of how firms’ culture actually affects their decisions, how they hold people accountable, or how they check if their approach is making a difference.