Our data and methodology

Our insurer figures come from the National Financial Ombud Scheme, Annual Report 2025, covering 2025, last checked 26 September 2026.

The Ombud is free for consumers and independent of the insurers it investigates. It publishes, per insurer, how many formal complaints were opened and what share were closed with some benefit to the complainant. That is the only public, comparable record of how insurers behave when somebody pushes back.

• Complaints opened, as published. Shown as a bar so you can compare insurers to each other, because the raw count on its own tells you little.
• How many in every hundred complaints ended with the customer getting something back. Published per insurer and already a ratio, so it needs no adjustment from us.
• No ranking. We do not sort insurers into a league table on one ambiguous number, and the homepage lists them without ordering them by score.

For context, across all non-life insurers 16.5 in 100 complaints ended in the customer’s favour.

This is the honest limit, and it is why we do not turn the number into a score.

A low share of complaints upheld can mean an insurer handled things well the first time, so the complaints against it were not justified. It can equally mean the insurer held its ground and people gave up. One figure cannot separate those two, and anyone presenting it as a quality score is over-reading it.

Bigger insurers also appear more often simply because they have more customers. A complaint count without that context is close to meaningless.

Our first version divided each insurer’s complaints by its market share, to correct for size. The arithmetic was right and the result was unfair.

Complaints to the Ombud are overwhelmingly from household customers, while market share counts commercial business too. So an insurer selling mainly to households scored worse than one with a large commercial book, before anything about its actual conduct was considered. We removed it.

The share of complaints upheld is already a ratio, so it does not need that correction, which is why it leads now.

• Licence standing, from the FSCA register of authorised financial services providers.
• Whether an insurer can afford to pay claims, from the solvency figures every insurer files with the Prudential Authority.

Both are public. Neither appears on the site until we have checked it properly and can explain it in the same plain terms as the rest.

When we get a figure wrong we correct it and say what was wrong. Our editorial policy lists the corrections we have already made, including two data errors we caught before anyone else did.