Press · Updated August 2026

For journalists

We publish figures on money UK consumers are owed and do not claim. Most of it comes from public records that are technically published and practically unusable: locked in appendices, split across releases, or named inconsistently between one dataset and the next. Where we have reconciled those, the tables are here to download.

Everything below is free to use with credit to Untap and a link, under CC BY 4.0. You do not need to ask. If you want the working behind a number, or a comment, contact@untap.money reaches a person, not a press office.

Story lines

Four figures that are either fresher or more granular than what is currently in circulation. Each says what it counts, where it came from, and what is being cited instead.

Transport for London took £164.7 million in maximum fares from passengers who did not complete a journey, in a single year

30.9 million
incomplete journeys in 2023
£164.7 million
charged in maximum fares on them
£8.90
the maximum fare ceiling, 2025 fares

This story gets written every few years and the figures in circulation are from 2016 and 2017. Ours came from a Freedom of Information response covering the 2023 calendar year, and we have not found it published anywhere else. The refund exists, it is rarely automatic for the service delay case, and the claim window is eight weeks.

Source: Transport for London, FOI-4311-2324, response of 25 March 2024

Our working: The Forgotten Tap

Private parking operators drop tens of thousands of appeals rather than contest them, and the rate varies enormously by operator

45.8%
of completed appeals ended in the motorist’s favour
104,349
appeals received in the year
14.4 million
keeper-data lookups bought by operators, 2024-25

Appeal outcomes are published, but only in an annual appendix that names operators without ranking them, and never alongside the keeper-data requests that show how many notices each operator sends. We reconciled the two. The result is an operator by operator table that does not otherwise exist.

Source: POPLA annual report and appendix, and DVLA registers statistics, reconciled by Untap

Our working: The Parking IndexDownload the table: 31 operators, comma separated

Train operators approve most Delay Repay claims, but almost nobody claims: 0.39 approved claims per 100 passenger journeys

81.6%
of closed claims approved, industry-wide
0.39
approved claims per 100 passenger journeys
£150.6 million
paid, across the 15 operators disclosed

The approval rate and the claim rate are usually reported separately, which makes the scheme look like it works. Put side by side they show the opposite: the barrier is not refusal, it is that people do not claim at all. The department discloses what it paid for only 15 of the 24 operators.

Source: Office of Rail and Road tables and Department for Transport transparency data, reconciled by Untap

Our working: The Rail IndexDownload the table: 24 operators, comma separated

More than 730,000 tax refunds went unclaimed in a year, averaging £855 each, because HMRC stopped sending cheques automatically

730,000
tax refunds unclaimed in a year
£855
average value of each
£1.7 billion
in unclaimed Child Trust Funds

HMRC published this in its own Employer Bulletin rather than as a statistical release, so it has had very little pickup. It sits alongside £1.7 billion in matured Child Trust Funds that almost a million young adults have never claimed, roughly half of which were opened for children in low-income families.

Source: HMRC Employer Bulletin, June 2026, and the Committee of Public Accounts on Child Trust Funds

Our working: The numbers

About Untap

What it is
A free UK service that works out what you are owed from train delays, flight compensation, Transport for London overcharges and parking charges
What it does not do
File claims for you, or take a percentage of anything you recover
Who runs it
Andrey, solo, in London
Founded
2026
Contact
contact@untap.money

How we handle corrections

Every figure carries the period it covers and the body that published it, and where a number is our own reconciliation or a company's estimate rather than an official statistic, we label it as such on the numbers page. If you find something wrong we would rather hear it than not, and we will correct it and say that we did.

The full datasets, with documented columns, are on the open data page.