Background 

1.1    Every train and station operator must, as part of an Accessible Travel Policy (ATP) approved by ORR, commit to providing assistance to older and disabled rail passengers that request it. Passengers may book assistance in advance or request it at the station (known as turn up and go, or TUAG). This is an obligation of operators’ licences to operate passenger services and manage stations. 

1.2    The demand for passenger assistance is increasing: between April 2024 and March 2025 there were 1,620,880 pre-booked passenger assists requested, an increase of 18% on the previous year and more than in any previous year. An assist is all the help provided at one station. A journey leg consists of two assists, one at the boarding station and one at the alighting station.

1.3    Since 2017 ORR has conducted ongoing passenger research to assess the extent to which the needs and expectations of passengers who have booked assistance in advance are being met. Our most recent report of survey findings, which covers the period from April 2024 to March 2025, suggests that satisfaction with assistance where it is delivered is high, but that service reliability is not good enough. In that period, 11% of respondents to the survey reported that they received none of the assistance they had booked. 

1.4    We know that failures of this sort can leave passengers feeling angry, frustrated and helpless. To help us drive improvements, we sought feedback on a new framework to benchmark the performance of operators’ delivery of assistance. Our Rail Passenger Assistance: Benchmarking Train Operators’ Performance consultation ran from 11 December 2024 to 14 February 2025. Respondents were in favour of ORR publishing a report that benchmarks operators’ performance in delivering assistance, highlighting how this would make it easier to identify issues and areas for improvement as well as celebrate success. Respondents also supported a broad focus on two key areas of assessment: delivery of assistance and capability to improve. Using this feedback, we developed the benchmarking methodology set out below, focusing our assessment of capability to improve on the poorer performing operators, while benchmarking delivery across all operators.

Benchmarking methodology

1.5    Our benchmarking assessment comprises two key components: a quantitative evaluation of Delivery performance (Chapter 2) and a Qualitative review of two poorer performing operators’ Capability to Improve (Chapters 3 and 4).

Delivery assessment

1.6    The Delivery assessment is based on data collected through our Experiences of Passenger Assist survey between April 2022 and March 2025, focusing on three core metrics: reliability, satisfaction, and staff knowledge and proficiency. 

• Reliability was measured by the percentage of passengers who reported receiving none of the assistance they had booked;
• Satisfaction captured the proportion of passengers who were ‘satisfied’ or ‘very satisfied’ with the assistance they received, when met; and
• Staff knowledge and proficiency reflected passengers’ views on the competence of staff delivering that assistance, when met.

1.7    Across the three-year period, 26,306 responses have been analysed. To ensure a robust and representative analysis, we applied a three-year scoring system. We gave greater emphasis to the most recent year (weighted 1:2:3). An overall delivery score was then calculated using the following weightings: 50% for reliability, 25% for satisfaction, and 25% for staff knowledge and proficiency. As the scores for the latter two metrics placed all operators in a narrow performance band, for 2025 we decided to rank operators based on reliability as the key metric for Delivery.

1.8    Performance has been attributed to the operator with overall responsibility for assistance at a station. In most cases this is the station operator, however at some Network Rail stations, one or more train operators are responsible for delivering assistance instead and our analysis reflects this.

1.9    In 2025, the following operators met the threshold of at least 50 survey responses per year and were included in the benchmarking assessment (total sample sizes over three years are provided in brackets):

• Avanti West Coast (1,936 survey responses over three years)
• Chiltern Railways (288)
• East Midlands Railway (738)
• Govia Thameslink Railway (836)
• Great Western Railway (3,830)
• Greater Anglia (839)
• London North Eastern Railway (5,338)
• Network Rail (4,442)
• Northern Trains (1,501)
• ScotRail (1,347)
• South Western Railway (1,163)
• Southeastern (857)
• TransPennine Express (845)
• Transport for Wales (1,185)
• West Midlands Trains (831)

1.10    Operators excluded from the assessment, due to low response rates or because they do not manage stations, included Arriva Rail London (London Overground), c2c, CrossCountry, Grand Central, GTS Rail (Elizabeth line), Hull Trains, London Underground and Lumo. We are actively working to expand future coverage by increasing survey sample sizes and exploring alternative data sources, such as feedback from the Passenger Assistance mobile app. This is explored further later in this chapter.

Capability to Improve assessment

1.11    We used the rankings from the Delivery assessment to target Capability to Improve assessments at two operators. These were the operators whose delivery performance raised the greatest concern, and where previous engagement, following on from letters we wrote to operators with the lowest reliability scores in our 2023 to 2024 passenger assistance survey had not led to meaningful change: South Western Railway (SWR) and West Midlands Trains (WMT).

1.12    An overview of the findings of our Capability to Improve assessments and proposed next steps for the two operators are provided in Chapters 3 (SWR) and 4 (WMT).

1.13    The Capability to Improve assessment examined four key areas:

1.14    Operators were assessed against the descriptors in Annex A. The process began with an information request, followed by iterative engagement where necessary to clarify and expand upon the evidence provided. In order to supplement this written evidence, we also undertook the following activities:

• Visited a sample of seven stations on each operator’s network to speak to staff and see assistance delivery in practice;
• Held a series of focus groups and interviews with staff that deliver assistance (the summaries of which are published separately on ORR's website); we aimed to recruit larger samples but were limited by recruitment and other challenges which limited volunteers to seven staff members for WMT and ten for SWR;
• Observed each operator’s accessibility panels of disabled passengers and stakeholders, and a WMT disability awareness induction training session; and
• Held meetings with key management staff and senior leaders at the two operators.

1.15    The findings and recommendations in this report are based on the evidence supplied to us during the course of our engagement. In some cases, despite a number of attempts to secure additional information and provide operators with opportunities for further disclosure, this evidence was limited, lacked coherence and structure, or presented only partial snapshots of documents with little context. This was particularly the case with WMT. Where the evidence is inconclusive as a result or did not provide us with sufficient confidence of capability, this has informed our overall assessment.

Future framework evolution 

1.16    We are actively working to evolve the benchmarking framework to include further sources of data, both to help broaden its scope to other operators and to develop other measures of operator performance.

1.17    To begin with, we have now incorporated an additional question into the passenger assistance survey to assess passengers’ confidence to travel with an operator following their experience of the specific assistance in question. From 2026, we intend to report on the percentage of passengers that report they are ‘more confident’ or ‘much more confident’ about doing a similar journey in future. 

1.18    From 2026, we also intend to report on the percentage of eligible staff who have received either induction training or refresher training within the past two years. 

1.19    Operators have been reporting to us against new passenger assistance outcome categories since April 2025, based on data captured by staff at stations. We will consider whether and how to build this data into the benchmarking assessment once we are content with data quality. 

1.20    We are also working with Rail Delivery Group (RDG) to explore how we might capture the reliability of turn up and go (TUAG) passenger assistance. 

1.21    For satisfaction, we are exploring with RDG how we might use feedback provided by users of the Passenger Assistance mobile app. We will consider whether and how to build this data into the benchmarking assessment.

Good practice

1.22    We asked all train operators to provide examples of good practice of the benchmarking descriptors set out in Annex A. Four operators responded: East Midlands Railway, Govia Thameslink Railway, Greater Anglia and Southeastern. In Chapter 6 of this report, we highlight a selection of initiatives from these operators for the benefit of other operators wishing to improve the reliable delivery of good quality assistance.

1.23    In 2026 we will consider the opportunities for all operators to learn lessons from any further good practice shared by higher-ranking operators.