Untap Rail Index · Issue 01 · April 2026
Late Britain.
When trains fail, how often does the passenger get paid?

About this study
The question, in plain words.
Every UK train passenger has the right to a refund when their journey is delayed by 15 minutes or more, or cancelled. The rules are public, the thresholds are simple, and the form lives on the operator’s website. And yet most of the money owed to passengers is never claimed.
This report asks one question: when a train fails, how often does the passenger actually get paid? It answers using only public data: performance from the Office of Rail and Road, compensation claims from the same regulator, and cash totals from the Department for Transport. Every figure has a citation next to it. The full dataset is downloadable. Nothing is estimated; where the published data has gaps, the report leaves them.
The gap, to scale · FY 2024 to 2025
Trains failed at 5.75%. Compensation reached 0.39% of journeys. The space between is what this report measures.
Sources: ORR Table 3103c · ORR Tables 4410 ÷ 1223
£150.6m
paid in compensation by the 15 disclosed operators.
45%
of eligible passengers claimed.
≈£80m
of Delay Repay goes unclaimed each year.
Five things to know.
Five findings, all from public ORR and DfT data, all checkable against the citations beside them.
Claim rates stopped rising in 2025.
Across five waves since 2016, the share of eligible passengers claiming Delay Repay climbed from 35% to 47%. In 2025 it fell to 45%. The 2-point dip is small enough to be noise; specific groups inside the average fell by more, and DfT marks those drops as real.
DfT and Transport Focus, 2025, Table 4.1
Contactless pay-as-you-go users gave up the most.
Contactless tap-in-tap-out is the fastest-growing way to pay for rail in Britain. It also has the lowest claim rate of any modern ticket type. Paper-barcode users still claim at 70%.
DfT 2025, Table 4.9
Long-distance routes get most of the compensation.
Five operators (Avanti, LNER, GWR, EMR, TPE) carry 12% of journeys but received 71% of the £150.6m paid by the 15 disclosed operators. The compensation system pays out best at the fares where claiming is most worth it.
71% = £107.8m of £150.6m (DfT transparency table). 12% = 209.5m of 1,728.7m journeys (ORR Table 1223).
Britain’s biggest operator is getting worse.
Britain’s biggest passenger operator saw its CaSL deteriorate while every long-distance operator improved. Of those large enough for a full read, four of five long-distance operators got better, and East Midlands Railway was a hair worse.
ORR Table 3103c; ORR Table 1223
Automatic compensation is going backward, not forward.
c2c remains the only UK operator with a fully automatic Delay Repay scheme. While operators talk about automation, the industry-wide automatic share has actually fallen, even as overall performance held roughly steady.
DfT and Transport Focus, Rail Delays and Compensation 2025, Table 5.1
After nine years of going up, claim rates stopped rising.
For nine years, the share of UK rail passengers actually claiming the compensation they were owed climbed steadily, from 35% in 2016 to a peak of 47% in 2023. In 2025 it stopped climbing. The latest wave of the joint DfT and Transport Focus tracker put the figure at 45% (Source: DfT and Transport Focus, Rail Delays and Compensation 2025, Table 4.1.), the first non-rising reading in the series.
The 2-point dip itself sits at the edge of survey noise. The breakdown by cohort does not.
After nine years of climbing, the share of passengers actually claiming Delay Repay has stopped rising.
Share of eligible passengers claiming Delay Repay, by wave, 2016 to 2025 (%)
Source: DfT and Transport Focus, Rail Delays and Compensation 2025, Table 4.1; n≈11,000 per wave.
UntapLondon and South East commuters' claim rate fell from 47% to 43%, a four-point decline DfT records as significant (Source: DfT and Transport Focus, Rail Delays and Compensation 2025, Table 4.12.). The pay-as-you-go contactless cohort fell from 35% to 23% (Source: DfT and Transport Focus, Rail Delays and Compensation 2025, Table 4.9.). The ticket format with the most growth in passenger volume now has the lowest claim activity of any modern format. Among aware non-claimants, 41% still cite the same reason: "not worth bothering for the amount I'd get back" (Source: DfT and Transport Focus, Rail Delays and Compensation 2025, Table 8.1.).
Behind those passenger numbers sits a £150.6 million industry compensation bill paid by the 15 operators DfT publishes figures for (Source: DfT, Train operating companies: Passenger’s Charter compensation, 26 February 2026.), on what the regulator records as around 5.75% of trains running 30 minutes late or being cancelled altogether (Source: ORR, Table 3103c, Cancellations and Significant Lateness, FY 2024-25.). Two private-sector estimates put unclaimed Delay Repay at £80 million to £90 million a year.
How we put the numbers together.
Performance. ORR’s official measure of Cancellations and Significant Lateness (CaSL): the percentage of passenger trains cancelled or arriving at their final destination 30 minutes or more after the booked time. Cancellations include trains that ran less than half their booked mileage. ORR Table 3103c.
Compensation activity. ORR Table 4410. "Approval rate" means approved divided by closed (the ORR convention). "Claims paid as a share of journeys" means approved claims divided by passenger journeys (ORR Table 1223).
Compensation paid in £. The DfT publishes a transparency table covering total Delay Repay value paid by 15 operators (the 14 DfT-contracted English operators plus Transport for Wales). DfT transparency table. FY 2024 to 2025 total: £150.6 million, an 8.6% increase on FY 2023 to 2024.
Worse trains paid out more. Bigger operators paid out less.
The 15 operators whose compensation totals DfT publishes. Sorted by claim payout per journey. Click any column to re-sort, or open the appendix for the nine operators outside the main league.
Click any column to sort.
Avanti West Coast Long-distance | 13.75% -0.20pp YoY | 3.45% | £40.5m | 86.5% | 34.9 |
LNER Long-distance | 9.12% -1.20pp YoY | 1.98% | £28.0m | 83.0% | 26.6 |
Great Western Railway Long-distance | 6.33% | 1.30% | £29.7m | 84.2% | 89.0 |
East Midlands Railway Long-distance | 5.88% +0.28pp YoY | 1.02% | £5.4m | 81.9% | 31.5 |
TransPennine Express Long-distance | 6.90% -0.90pp YoY | 0.89% | £4.2m | 82.1% | 27.5 |
CrossCountry Long-distance | 11.50% -2.08pp YoY | 0.65% | £10.9m | 67.5% | 37.8 |
West Midlands Trains Commuter | 6.45% | 0.50% | £3.7m | 78.2% | 67.7 |
Northern Trains Regional | 7.43% | 0.41% | £3.3m | 76.2% | 91.8 |
Transport for Wales Regional | 7.88% | 0.37% | £2.3m | 75.2% | 31.3 |
South Western Railway Commuter | 5.38% +0.50pp YoY | 0.34% | £8.2m | 84.0% | 165.6 |
Southeastern Commuter | 3.80% +0.42pp YoY | 0.33% | £4.3m | 81.2% | 137.9 |
Chiltern Railways Commuter | 3.20% | 0.32% | £1.2m | 87.9% | 22.9 |
Govia Thameslink Railway Commuter | 7.32% +1.32pp YoY | 0.27% | £5.8m | 81.6% | 298.0 |
Greater Anglia Commuter | 2.38% | 0.25% | £2.9m | 79.7% | 81.8 |
c2c Commuter | 2.50% | 0.06% | £0.1m | 83.6% | 37.3 |
The operators with the worst service paid the most in compensation. The operators with the largest passenger volumes paid the least per journey.
UK rail operators, FY 2024 to 2025: services 30+ min late or cancelled (x-axis) vs approved claims per 100 journeys (y-axis)
Source: ORR Table 3103c (x-axis); ORR Tables 4410 ÷ 1223 (y-axis); Untap analysis.
UntapFive patterns in the data.
The long-distance premium
Among the 15 main-league operators, the five with the highest claim-payout rate are all long-distance. Long journeys, high fares, account-based ticket types, all converge on the same shape: claiming is worth the friction.
DfT 2025, Table 4.7. n=4,775 eligible-delay respondents.
Commuter compensation is falling.
London and South East passengers’ claim rate fell four points. GTR’s CaSL got worse year on year while every long-distance operator improved. Part of the commuter picture is real deterioration, part is the season-ticket accounting gap.
DfT 2025, Table 4.12; ORR Table 3103c
One in three CrossCountry claims is rejected.
CrossCountry rejects roughly a third of closed claims. Approval rates fell year on year at c2c (-12.1pp), CrossCountry (-7.5pp) and TfW (-8.3pp) over the rail periods ORR has reported. The published data does not say why.
ORR Table 4410. National approval rate 81.6%.
Contactless users almost stopped claiming.
The single steepest decline among ticket-format cohorts. The data does not isolate cause. What it does say is that the format with the most growth in passenger volume has the lowest claim activity.
DfT 2025, Table 4.9
At fares of £50 or more, 55% of eligible passengers claim. At fares under £10, only 32%.
Claim rate by fare paid, eligible-delay passengers, 2025 wave (%)
Source: DfT and Transport Focus, Rail Delays and Compensation 2025, Table 4.7; n=4,775 eligible-delay respondents.
UntapFive operators, five kinds of story.
The gap is not one shape. Five operators each show a different way the system fails: bad service paying out most, biggest operator getting worse, claims rejected rather than paid, automation that works but only for some, and regional passengers who barely claim at all.
When bad service meets high fares.
The pattern: the worst service record in Britain (13.75% CaSL, 39.9% On Time) running on the country’s highest single fares. Result: the biggest compensation bill of any operator, paid to a small share of total passengers. Long-distance fares mean each delay is expensive to compensate.
DfT transparency table; ORR Tables 1223, 3103c
When the operator is huge — and getting worse.
The pattern: the largest passenger operator in Britain, and the operator whose service deteriorated most in FY 2024-25 of any in the main league. Smartcard auto-DR covers a narrow ticket range; season-ticket compensation flows through separate routes outside Table 4410. The gap looks worse than it is, and is also widening.
ORR Tables 1223, 3103c, 4410
When claims are rejected, not paid.
The pattern: claims received grew 32% year on year, but a third of closed claims were turned down — the lowest approval rate of any DfT-contracted major, and falling. CrossCountry’s approval rate dropped 7.5 percentage points in the rail periods ORR has reported. The published data does not say why.
ORR Table 4410; ORR factsheet rail periods 5 to 7
When automation works, but only for some.
The pattern: c2c is the only UK operator paying Delay Repay automatically — and it works. CaSL 2.50%, claim payout 0.06% (very low because the service is reliable). But the auto scheme covers smartcard users only. Most of c2c’s customer base, and 23 other operators, get nothing automatic.
DfT and Transport Focus, Rail Delays and Compensation 2025, Table 1.1
When passengers don’t even claim.
The pattern: Britain’s largest regional operator. CaSL is 7.43% (worse than the GTR figure that gets all the attention), but the claim payout rate is 0.41%. The cause here is not a scheme-design quirk and not a rejection problem. It is awareness: in the DfT survey, regional passengers claim at 35%, the lowest of the three sector cohorts. The compensation rules work; the people who would benefit do not know they exist.
DfT 2025; ORR Tables 4410, 1223, 3103c
Why most people don’t claim.
The DfT and Transport Focus tracker (May 2025, fieldwork March 2025, 11,007 rail passengers screened, 4,775 with an eligible delay) (Source: DfT and Transport Focus, Rail Delays and Compensation 2025.) is the strongest authoritative source on why eligible passengers do not claim. Four numbers shape the picture.
Awareness and eligibility uncertainty
Roughly half of non-claimants either did not think to claim or did not think they could. (DfT does not say whether respondents could pick both, so the true total may be lower than 47%.)
DfT 2025
Not worth the bother
At small fares, this dominates. At fares of £100 or more it falls below 15%. The smaller the fare, the louder the reason gets.
DfT 2025, Table 8.1
Time
Trainline and YouGov’s separate December 2025 survey found 58% of recent claimants spent six or more minutes filling in the form.
DfT 2025, Table 8.1
Format and friction
The most modern formats correlate with the lowest claim activity. We do not assert this is causal; the formats with the fastest growth in usage are also the formats with the smallest fares and the least account-based receipts.
DfT 2025, Table 4.9
About £80 million is left on the table every year.
FIG. 04 · Industry total estimated at ≈£230 million
One in three pounds owed in Delay Repay was not paid out in FY 2024 to 2025.
Sources: DfT transparency table; Trainline / YouGov 2025; DPS Simulation 2024.
Paid
£150.6m · 65%
what 15 disclosed operators actually paid out, FY 2024 to 2025. DfT transparency table, +8.6% on the year before.
Unclaimed (estimate)
≈£80m · 35%
what private-sector estimates (Trainline / YouGov; DPS Simulation) say is owed but never reaches the passenger. Both estimates fall within roughly 10% of each other.
Two figures circulating in the past year:
- £80 million+ unclaimed annually. Trainline and YouGov, December 2025, n=2,000 UK adults (vendor-commissioned).
- ≈£86 million. DPS Simulation analysis, derived from ORR FY 2023 to 2024 figures.
Both come from similar surveys. Neither has been peer-reviewed, and they are not independent of each other. They sit in roughly the same range as what the published data implies.
Industry total Delay Repay paid in FY 2024 to 2025 (the 15 operators in the DfT table) was £150.6 million. The £80 to £90 million range, if accepted, implies a total entitlement around £230 million across the same 15 operators, and an unclaimed share around 35%.
We do not add a new headline estimate. The point of this report is to show where the gap sits, operator by operator.
Automation closes the gap. Awareness no longer can.
Claim rates climbed twelve points between 2016 and 2023 (Source: DfT and Transport Focus, Rail Delays and Compensation 2025, Table 4.1.); they have now plateaued and dipped. The next twelve points will not come from another awareness campaign. They will come from changing the system.
Automation works where it’s wired in
Where the c2c smartcard scheme applies, compensation is delivered without passenger action. Where it does not apply (most c2c passengers and most other operators), the gap remains. The proof of concept exists; the coverage does not.
DfT 2025, Table 1.1
Automation is regressing nationally
Operators are talking about automation. The share of claims actually paid automatically has gone down (11% to 8%); the wider “automatic plus 1-click” share is also down, from 23% to 19%.
DfT 2025, Table 5.1
Contactless is where the biggest gain hides.
The format with the fastest growing passenger volume has the lowest claim rate. A nationally consistent contactless Delay Repay scheme would, on the DfT survey numbers, close roughly a third of the national gap.
DfT 2025, modelled from Tables 4.9 and 4.12
Awareness has done its work
Ten years of campaigning lifted the headline claim rate twelve points. It has now plateaued and dipped. Further closure depends on changing the system, not on better passenger education.
DfT 2025, Table 4.1
Reproduce every number in this report.
Every figure in this report traces to one of the published sources listed below. If you find an error, email contact@untap.money with the cell reference and we will publish a correction.
- ORR Table 3103c · Historic CaSL by operator, quarterly. Download
- ORR Table 4410 · Delay compensation claims received, closed, approved by operator. Download
- ORR Table 1223 · Passenger journeys by operator. Download
- DfT transparency table · Train operating companies: Passenger’s Charter compensation, 26 February 2026. Open
- DfT and Transport Focus · Rail Delays and Compensation 2025, May 2025. PDF
- ORR mid-year factsheet · Delay compensation claims, rail periods 5 to 7 of FY 2024 to 2025. PDF
- ORR full-year factsheet · Delay compensation claims, rail periods 11 to 13 of FY 2024 to 2025. PDF
- ORR TOC Key Statistics · FY 2024 to 2025 factsheets, 21 August 2025. Index
- Trainline and YouGov · December 2025 press release, n=2,000 UK adults. Read
Late Britain is published by Untap. All underlying data comes from ORR and DfT.
Published 28 April 2026 · Comments to contact@untap.money.