Free tool · Calculator
How many vulnerable customers are you not seeing?
A low vulnerability rate is usually read as good news about the customer base. It is more often a finding about the process, and this is the arithmetic that separates the two.
The short answer
The FCA reports that about four in ten consumers with vulnerability characteristics disclose them to their provider, and only 19% feel providers encourage disclosure. Enter your customer count and how many carry a flag, and this returns an estimated true number, the customers you are likely not seeing, and how far your flagged rate sits from a plausible one. It uses your own disclosure assumption rather than a fixed multiplier.
On this page · 5 sections
This is deliberately a crude instrument. It multiplies what you can see by an assumption about what you cannot, and the assumption is the whole model — which is why it is a field you set rather than a constant we chose.
Use it for one purpose: deciding whether your flagged rate is plausible. If the estimate lands far above what you record, the finding is about how you ask, not about who your customers are.
Your numbers
Opens on a worked example, not an industry average. Replace every figure with your own — there is no such thing as a default here.
—
Show the working
The number this is really about
The FCA found that only 29% of firms tested product impacts on vulnerable customers, and that of firms training non-frontline staff, 54% said the training covered how vulnerability related to each person’s role. Those are decisions rather than demographics, and both are things a firm can change this quarter.
The disclosure rate is the same kind of number. It describes your process, not your book.
What this cannot tell you
It is an estimate built on one assumption you supply. Vulnerability is also a spectrum of circumstances rather than a category of person — somebody can be vulnerable for a fortnight — so a flag set once and never reviewed creates a different problem from the one it solves. Any flag worth setting carries a date.
Questions people actually ask
What proportion of customers disclose vulnerability?
The FCA’s review published on 7 March 2025 found only about four in ten consumers with vulnerability characteristics disclosed them to their provider, and that 19% felt providers encouraged disclosure. A firm that waits to be told will therefore identify a minority of those it should.
Is a low vulnerability rate good news?
Usually not. A very low recorded rate is more often a finding about how a firm asks than about who its customers are, and it is the kind of number a supervisor treats as a question rather than an achievement.
How accurate is this estimate?
It is deliberately crude: it divides what you record by a disclosure rate you supply. The assumption is the whole model, which is why it is an editable field rather than a constant. Use it to judge whether your recorded rate is plausible, not to report a number.
What should a firm do after running this?
Cut one outcome measure you already report — complaint rate, time to resolution — by the vulnerability flag and look at whether the two groups differ. That converts the question from an estimate into evidence, which is what the FCA found most firms in its outcomes work could not show.
Sources
- Financial Conduct Authority, “Delivering good outcomes for customers in vulnerable circumstances – good practice and areas for improvement”, 7 March 2025: 725 firms surveyed; about 4 in 10 consumers with vulnerability characteristics disclosed to providers; 19% felt providers encouraged disclosure; of firms with training for non-frontline staff, 54% said it covered how vulnerability related to each person’s role; 29% tested product impacts on vulnerable customers; most firms in the outcomes work “were unable to show how they effectively monitor and take action on outcomes”. fca.org.uk ↗ — primary; the regulator’s own review. Firm-behaviour figures are self-reported re-checked every 6 months
- The estimate is flagged ÷ disclosure rate, which assumes disclosure is independent of everything else about a customer. It is not, so treat the output as an order of magnitude. — our own model, with its central assumption stated as a weakness
- Nothing entered here leaves the page. No network call, no identifying field. — first-party, and enforced by the build
Checked 23 September 2026. Next scheduled check 22 March 2027. Numbers that move — leaderboards, live indices — are re-checked every 30 days; annual datasets and rules in force every six months; dated research once a year. If something here has gone stale before we got to it, tell us and we will correct it and say what changed.
Cut one measure you already report by the flag.
That single cut turns an estimate into evidence, and it is the thing 71% of firms in the FCA’s work could not produce. We build the capture so the flag exists, carries a date, is visible to whoever needs it, and can slice any measure you already publish.
Start the conversation