Independent evidence that customers understand you, built to stand up to FCA scrutiny

Marking your own homework is the weak point

The firm wrote the letter. The firm reviewed the letter. The firm reports that customers understand the letter. A board can sign that, but it cannot test it.

In March 2026 the Financial Conduct Authority (FCA) published its review of consumer understanding (opens in a new tab) at 38 firms. It found firms that did not test communications with customers at all, or “relied on very small, unrepresentative groups”, and firms that assumed no complaints meant the letter was clear. An independent test changes what the report rests on. The people who read the letter do not work for you. The recordings exist, and the pack says who ran the test and how.

Readability software does not fix this. It counts syllables and sentence length and gives you a score. The score can tell you a paragraph is short. It cannot tell you that a reader took “tax-free” to mean “tax-free every year”, closed the page and phoned a friend instead.

What the Duty asks you to show

Your board must review and approve a report at least once a year on whether customers are getting good outcomes, with the evidence behind it. For consumer understanding, the FCA expects firms to test communications where appropriate and to show how they know customers understand them. By July 2026 the FCA had 11 open investigations into possible breaches of the Duty, and had said that three years in, firms have had long enough to embed it.

What a test looks like

We sit with eight people, one at a time, on a video call, and give each of them the real thing: the letter, the statement, the screen, or a journey of up to five touchpoints. We ask them to do what they would do at home and to say out loud what they think it means. Each session is recorded. Afterwards we mark every point where what they understood is not what you meant.

Why moderated sessions, not a survey

• A panel survey of around ninety people can tell you that customers misunderstand a letter. A moderated session shows why, and which sentence to change.
• Online surveys under-represent the people the Duty cares most about. People with low literacy, cognitive impairment or sight loss drop out of a web survey and never appear in the data. The FCA’s review found firms that had not tested with people with accessibility needs, language requirements or lower financial capability. In a moderated session they take part, with a researcher there to support them, and you hear them. For the vulnerability requirement this is the better instrument, not the weaker one.
• Survey panels recruit people familiar with a product type, not your customers. So do we. The difference is not who takes part; it is what you learn from them.

Why eight people

Eight one-to-one sessions produce eight hours of recorded reading, with every misunderstanding tied to the sentence that caused it, and a sample that covers the FCA’s four drivers of vulnerability by design rather than by chance. Eight is enough to say “six of eight” and show the clips, and small enough that every participant is someone the researcher sat with. That is evidence a board can act on and a regulator can replay.

You send the communication. Finding the right people is our job.

We recruit participants matched on what actually drives comprehension: the product they hold, where they are in the decision, age band, employment status and social grade, financial capability and confidence, and characteristics of vulnerability.

You supply the communication and nothing else. No customer list, no data protection review of our access to your customers, no internal argument about showing your own letters to your own customers.

For example, a pensions client

With a pensions client we would be looking to match profiles like these four. The exact profile is agreed with you for each study.

The same approach applies in other sectors at the equivalent decision moments: arrears and overdraft notices in banking, fixed-rate expiry in mortgages, claims decisions in insurance.

Vulnerability is a quota, not a separate group

Every study includes participants with characteristics across the four drivers in the FCA’s guidance FG21/1: health, life events, resilience and capability. Around half of UK adults show at least one, so a sample without them is not a sample of your customers.

We never ask anyone whether they are vulnerable. Screening uses neutral, factual questions, and we classify afterwards.

Every report includes a full sample composition table, with employment status, social grade and vulnerability drivers, so you can see exactly who was spoken to.

What you get every month

One study of one communication, or of one journey of up to five touchpoints: a letter, the email reminder, the web page, the form, the confirmation. We build the test environment from screenshots you supply, so we never need access to your live systems.

Testing the journey rather than the screen is what surfaces sludge: the friction between steps that stops people finishing. A single-screen test or a survey structurally cannot see it. Each study produces:

Three steps, three to four weeks

Studies run one a month, so the next one starts while the last pack is being written. It is a monitoring programme, not a periodic audit.

One price

Why monthly

A monthly invoice is a number most directors can approve without a new budget line. The FCA expects monitoring, not a one-off exercise, and a study every month means each quarter’s report rests on fresh evidence rather than on a test done last year. You stop when the evidence stops being useful.

What it compares with

• A fine. The FCA fined TSB £10.9 million for how it treated customers in arrears, and TSB paid £99.9 million in redress on top. A year of testing is about half of one per cent of that fine.
• A readability tool. Cheaper, and it tests nothing with anyone.
• A survey panel. Ninety people telling you that something is wrong, and no recording of why.
• A researcher on staff. Similar money, but then your firm is marking its own homework.

Who this is for, and who it isn’t

Run by someone who has watched a thousand people fail to understand things

Twenty years in fintech, insurance and pensions, most of it spent sitting next to customers while they try to use something a committee signed off. Creator of the Build Loop, the decision method behind the way these studies are run.

Get a free teardown of one of your communications

Every engagement starts the same way. Send us one live customer communication: a letter, a statement, a screen or a journey. Within five working days you get a written teardown, independent and specific: where a reader is likely to misunderstand it, why, and what we would test first. No call, no obligation.

What happens next

1. You send the form below.
2. Within five working days we email you the teardown.
3. If it is useful, a first study starts at £4,500 a month, with a three-month minimum.
4. If it isn’t, nothing else happens.