What we found

Common Platform changes

2.1 HMCTS’s and MoJ’s investigation identified a range of causes for the Crown Court data inaccuracies, including human error, technical issues and data coding issues.

2.2 One issue was how users of Common Platform were recording the closure of cases on the system: court staff were recording the final result for an offence in a free text field instead of using defined result codes. A free text field is a necessary part of Common Platform, allowing frontline staff to record more-complex results that cannot be easily codified. But the free text field should not have been used to record case closures. HMCTS informed us that the design of Common Platform inadvertently led to this operational error, despite guidance asking staff not to use the system in this way. This issue was fixed in August 2024, ensuring that free text fields can no longer be used to record the final result. HMCTS also issued updated guidance to frontline staff.

Data validation

2.3 To assure itself that the data quality issues have been resolved and that the management information accurately represents the source data, HMCTS conducted an extensive data validation exercise with operational and service teams between July and October 2024.

2.4 HMCTS carried out two types of data check:

• Data quality reporting – identifying cases with unusual characteristics that might indicate an error that affects the case’s open or closed status.
• Random sampling – taking a representative sample of 7,000 cases and testing whether the open or closed status in the management information matches the information on the case management systems (Common Platform and Xhibit).

2.5 HMCTS gathered operational colleagues and data experts across HMCTS and MoJ to discuss and agree which Common Platform cases should be prioritised for data quality reporting. In total, around 3,700 cases were identified by focusing on cases affected by the free text field issue, older cases where there had been no recent activity and cases with a hearing listed on a weekend or bank holiday. Operational colleagues then used those reports to check the accuracy, completeness and timeliness of the information on Common Platform and Xhibit and amend any records requiring correction. All cases affected by the free text field issue required correction.

2.6 HMCTS also checked a total of 7,000 randomly sampled cases, split equally between open and closed cases. HMCTS set a pass threshold of 95%, meaning that at least 95% of cases checked should have the same status in the management information and at source. All regions passed the check, with around 300 cases needing investigation. In approximately 70% of these cases, the mismatch was the result of the time lag between the data being extracted and the validation review, during which the case status had changed – a common occurrence in an administrative data system. Other reasons for a mismatch included data entry errors and duplicate entries.

2.7 We consider that HMCTS took a systematic and rigorous approach to data validation by using two complementary data checks to investigate the data quality issues.

2.8 MoJ carried out its own analysis to check the cases that HMCTS had identified for investigation against the published official statistics. This enabled MoJ to quantify the scale of the data inaccuracies and assess their impact on the official statistics. MoJ found that around 2,500 Crown Court cases recorded as open in the MoJ statistics for the period ending June 2024 should have been recorded as closed. This equated to a 3% reduction in the number of open Crown Court cases.

2.9 As part of the data validation exercise, MoJ found that some cases were not being accurately identified as entering the Crown Court from the Common Platform data pipeline. MoJ made improvements to address this issue which resulted in an increase of around 2,500 cases to the baseline Crown Court caseload for the period ending June 2024. Taken together, MoJ determined that the overall impact of the two quality improvements on the trends in the Crown Court caseload statistics was minimal.

2.10 MoJ has been transparent with users about the approach and outcomes of this data validation work. The Crown Court Data Quality update published alongside the December 2024 statistics outlines the nature, scale and impact of the data quality issues identified and describes how HMCTS and MoJ had rectified these issues. We consider that MoJ’s quality assurance arrangements are robust.

2.11 This quality assurance work has given HMCTS and MoJ confidence in the accuracy of Crown Court data from Common Platform and Xhibit. As a result, MoJ restarted the publication of the criminal court statistics in December 2024.

Quality assurance processes

2.12 To support ongoing monitoring and validation of Crown Court cases, HMCTS worked with operational, policy and analytical experts to develop routine data quality reporting. The new reporting process is currently being rolled out. It includes both the checks that were developed for the data validation exercise and some additional checks, such as whether cases are missing a marker for the outcome of a trial.

2.13 As part of this work, HMCTS is considering adding new functionality to Common Platform to improve underlying data quality. For example, it is exploring adding in-system validation checks to Common Platform. These automated checks would notify users when they have entered an unexpected value and ask them to confirm if the details are correct or stop them from taking further action.

2.14 We welcome that HMCTS is building on the success of the data validation exercise by introducing routine data quality reporting. This has enhanced quality assurance and will help ensure that unusual cases are identified early and errors are corrected quickly, which will improve the quality of Crown Court data in the long term.

Identifying issues

2.15 We heard that the unclosed cases issue linked to the free text field was only identified through analysis for a Freedom of Information request for the longest open cases on the Crown Court caseload, not as part of routine data checking and validation. In our view, applying good practice would have meant that HMCTS was routinely monitoring the longest open cases as part of its quality assurance activities. HMCTS’s improvements to its quality assurance arrangements should prevent a repeat of this scenario, where closed cases are left unclosed on Common Platform.

Common Platform issues

2.16 Common Platform has had a range of technical issues since its rollout. These include difficulties progressing some cases through the system, requiring a case to be created on the legacy system instead; and cases not transferring from Common Platform to HMCTS’s Strategic Data Platform (its core analytical platform). HMCTS told us that its focus in 2024 was on improving testing and increasing its capacity to fix remaining issues. It has made substantial progress on reducing the number of issues and expects to be in a better position by 2025/26. HMCTS said that it has prioritised fixing issues over adding new functionality to Common Platform, which it will roll out in 2025.

2.17 Importantly, most of the remaining issues do not affect the management information and official statistics because they are not relevant to the opening and closing of cases (the key data field used to produce the caseload statistics). We heard that fewer than 0.1% of cases in the management information are affected, and not all of these are linked to the open/closed status. We therefore agree with HMCTS’s assessment that Common Platform technical issues have a negligible impact on data quality.

External review

2.18 The KPMG review, commissioned by the Lord Chancellor, had a different scope to our review. KPMG’s review concentrated on HMCTS’s and MoJ’s data pipelines, data processes and internal documentation of Common Platform and Xhibit, whereas we mostly focused on MoJ’s published statistics. The main areas of overlap were the quality assurance of source data and the methods used in the ‘One Crown’ work.

2.19 The KPMG review’s headline finding is that MoJ can have a significant level of confidence in its Crown Court caseload statistics. This echoes the findings of our review and provides further evidence of the quality of MoJ’s statistics. KPMG identified a few minor areas for improvement related to the data pipeline and quality assurance processes that MoJ should consider addressing to follow best practice.

2.20 MoJ published an executive summary of KPMG’s findings (PDF) alongside the December 2024 statistics and published the full report (PDF) in March 2025. We are pleased that MoJ published KPMG’s findings in full. This transparency helps users understand potential risks to the quality of the statistics and the evidence that underpinned KPMG’s recommendations.

Communicating changes

2.21 MoJ notified users about the cancellation of the June 2024 statistics via a statement on the statistics landing page that explained the reason for the change. MoJ later expanded the statement to include further details about its quality assurance work and the likely scale and impact of the data quality issues. This helped to inform users about MoJ’s approach to investigating and resolving the issues.

2.22 The cancellation of the June 2024 publication was announced less than 24 hours before the statistics were due to be published. We understand this late notice was because MoJ was only made aware of the potential scale of the data quality issue by HMCTS the day prior to publication. At that point, MoJ did not have a clear picture of the scale of the data quality issue. HMCTS’s initial investigation had suggested that the scale of the issue was larger than expected, and it therefore needed to undertake further work to quantify the issue. By cancelling the statistics, MoJ avoided the risk of misleading users about trends in the published criminal courts statistics; it was a proportionate response. It is good that MoJ immediately and proactively raised its data quality concerns with us.

2.23 MoJ cancelled the September 2024 statistics a week before they were due to be published. It told us that it had considered publishing partial statistics but decided against it due to the ongoing ‘One Crown’ work, which may have affected trends in the statistics. While it would have been good practice to give earlier notice about the cancellation, we consider that this was a sensible decision given the uncertainties around the quality of the data.

The ‘One Crown’ project

2.24 Until recently, MoJ and HMCTS maintained separate versions of the data used for reporting, analysing and modelling the Crown Court caseload. This practice at times has resulted in differences between the figures in the management information produced by HMCTS and the official statistics published by MoJ. Although MoJ and HMCTS did try to correct and reconcile the data regularly, the gap between the management information and the official statistics grew over time. As a result, confusion arose about which numbers were correct, particularly regarding open caseload figures.

2.25 To address this lack of coherence in the Crown Court data, HMCTS and MoJ established the One Crown data project (One Crown). The aim of One Crown was to align HMCTS’s and MoJ’s methodologies to create a single dataset for both HMCTS management information reporting and MoJ official statistics, and thereby increase the coherence, transparency and clarity of the data for users. We heard that One Crown has been on HMCTS’s and MoJ’s workplan for a while. It is good that HMCTS and MoJ used the pause in the official statistics as an opportunity to carry out this improvement work.

2.26 MoJ and HMCTS agreed on methodologies and definitions for key metrics, focusing initially on the Crown Court caseload, including receipts, disposals and open cases. All decisions were discussed and approved by the One Crown Steering Group, which is made up of MoJ and HMCTS staff.

2.27 MoJ and HMCTS used a mix of approaches to align the metrics. For some metrics, such as Most Serious Offence, the MoJ methodology was adopted. For other metrics, like the Reference data source, the HMCTS methodology was adopted. And for several metrics, including Categorising Indictable Only and Triable Either Way, a new methodology was jointly developed.​ When deciding on the best approach for each metric, MoJ and HMCTS examined the impact on the headline metrics, such as whether the decision would inflate caseload numbers. They also considered the number of changes that would need to be made to Xhibit and Common Platform to implement the decision. Having a single agreed and validated methodology used by both HMCTS and MoJ reduces confusion for data users and increases trust in the data. It also minimises the risk of future divergence between the management information and official statistics.

2.28 The One Crown project aimed to move away from the methods and definitions used in the legacy system (Xhibit) and, where possible, align them with those used in Common Platform. However, MoJ and HMCTS acknowledge that alignment with Common Platform methods has not always been possible because of the system’s functionality limitations and the large number of cases currently in legacy systems. For example, Xhibit creates separate trial and sentence case numbers, while Common Platform keeps them as one case. Due to Xhibit’s limitations, it is not possible to report these split cases as one, so the decision was made to continue splitting cases into trial and sentence. MoJ told us that there will be fewer workarounds and more alignment with Common Platform as the system continues to develop and more cases are completed and migrated from Xhibit.

2.29 One Crown has deepened HMCTS’s and MoJ’s understanding of Crown Court data and the systems that underpin them. In addition, the work has strengthened relationships within HMCTS and between HMCTS and MoJ. For instance, data engineers described to us how knowledge was pooled in both HMCTS and MoJ to design the single data model, and how operational users of Common Platform were consulted to understand how data are entered and the different operational processes involved. As a result, MoJ and HMCTS have a better shared understanding of Common Platform and data quality. Ongoing governance and oversight will ensure that these relationships continue after One Crown has been completed.

2.30 MoJ published a consultation on the One Crown changes to the Crown Court data processing in Criminal Court Statistics Quarterly as part of the release of the December 2024 statistics. This document explains the reasoning behind decisions and clearly sets out the impact on the statistics. We welcome MoJ’s openness and transparency on the decision making, methodologies and impact, and its engagement with users about the changes. However, the published information does not reflect all the decisions made as part of the One Crown project and their impact. So that users have the full picture, MoJ should consider publishing a complete summary of the changes and their impact on the statistics.

Methods changes

2.31 Although the move to Common Platform has led to improvements to the methods for producing the statistics, MoJ identified that the change in system would create discontinuities in the time series. This is due to differences in the way that data are recorded on Common Platform and legacy systems. MoJ anticipated and mitigated these discontinuities by developing estimates from the legacy systems and Common Platform on a ‘best equivalent’ basis, which ensures that the data can be compared over time.

Data corrections

2.32 So far, we have focused this review on the methods and quality of Crown Court data. MoJ has also improved the methods and quality of magistrates’ courts data, which has led to substantial corrections to the open caseload estimates for magistrates’ courts. In particular, the December 2024 statistical bulletin explains the headline caseload figure was revised down by 20% (around 80,000 cases) after HMCTS identified a variety of issues in the data derived from the Libra case management system.

2.33 As part of work to modernise its data and management information (separate from the quality improvement work described earlier), HMCTS found that the system used to extract management information from Libra did not recognise archived cases as closed. These cases were incorrectly included in the open caseload estimate and should have been excluded. MoJ is investigating additional cohorts of legacy cases and expects to make further corrections to the magistrates’ courts data.

2.34 The archived cases issue suggests that HMCTS and MoJ did not previously have a clear understanding of how Libra dealt with archived cases. It is good that HMCTS now has this understanding and that it is enabling the production of more-accurate estimates of the magistrates’ courts caseload. MoJ was transparent about the scale and nature of the corrections.

Quality information

2.35 MoJ publishes two quality documents alongside the statistics: the quality statement and the guide to criminal court statistics (user guide). The quality statement contains a clear and detailed summary of MoJ’s quality assurance (QA) arrangements. However, the information about HMCTS’s QA arrangements is limited. The statement says that HMCTS carries out validation of data when they are entered onto the case management system, but it does not explain how this is achieved. To assure users that data quality is regularly monitored by HMCTS, MoJ should give an overview of the QA conducted by HMCTS.

2.36 The user guide briefly explains the Common Platform, Xhibit and Libra systems, with a focus on developments to data processes following the rollout of Common Platform. However, it does not fully explain the strengths and limitations of Common Platform in relation to data quality. The guide also does not give users an indication of the proportion of the open caseload on legacy systems and when Xhibit and Libra are expected to be decommissioned. To demonstrate to users the suitability of the administrative data systems for producing the statistics, MoJ should publish a comprehensive overview of all systems, including the legacy systems.

2.37 The quality statement indicates uncertainties in the data. This includes general comments around administrative systems and their limitations and quantitative information on how well the magistrates and Crown Court data are matched. How the statistics are revised between quarterly releases is briefly explained; however, we would expect to see more information on the scale of these revisions so that users can better understand how the estimates change over time.

2.38 Additionally, some quality information needs to be updated. For example, the description of Common Platform in the user guide says that the system is still being rolled out to courts, when the rollout was concluded in summer 2023. MoJ acknowledged that the quality statement and user guide may not accurately reflect current processes and data quality.

Recommendation 1: To assure users about all aspects of data quality, MoJ should expand, and routinely review, its quality information. The published documents should cover:

• HMCTS’s quality assurance arrangements
• HMCTS’s and MoJ’s roles and responsibilities in collecting and processing data
• the strengths and limitations of Common Platform in relation to data quality
• the proportion of the caseload on legacy systems and the timeline for decommissioning legacy systems

2.39 We suggest that MoJ use our Quality Assurance of Administrative Data (QAAD) framework to present this quality information.

2.40 The quality statement provides clear advice for users on interpreting the statistics. However, not all users will read the quality documentation, and the statistical bulletin contains little commentary about data limitations and uncertainty.

Recommendation 2: To help users interpret the statistics, MoJ should include caveats and explanations of uncertainty alongside the statistics in the statistical bulletin.

2.41 Currently, quality information is split between the quality statement and user guide, which makes it hard for users to get a complete picture of the quality of the statistics. Some content is duplicated across the two documents. For example, both the user guide and the quality statement state that there are inconsistencies in the underlying methods for Crown Court estimates between Xhibit and Common Platform, but only the user guide explains the differences between measures from the two systems. To improve the accessibility of quality information for users, MoJ should consider combining existing quality documentation and publishing it in an HTML format.

Official statistics in development

2.42 The criminal court statistics are accredited official statistics. MoJ also produces official statistics in development on different aspects of the criminal courts, including the timeliness of cases and the age of the open caseload. It is good to see the development of new statistics that add insight on different aspects of the criminal court caseload. However, many of these statistics have been in development for years, without a clear plan to move them to official statistics.

2.43 MoJ told us that only the age of open caseload measure is still in development. This is a measure of the average age that a Crown Court case in the open caseload has been open. MoJ developed the method for the age of open caseload measure with HMCTS and policy colleagues and sought feedback from users. We welcome that MoJ intends to bring this statistic out of development once it has completed the second phase of the One Crown project. MoJ should communicate changes to the status of the age of open caseload measure and other official statistics in development measures, explaining why they are now of sufficient quality to be published as official statistics.

Applying learnings

2.44 We commend MoJ and HMCTS on their collaborative approach to investigating and addressing the Crown Court data quality issues. This work has led to a better shared understanding of the data and case management systems and stronger working relationships between MoJ and HMCTS. We encourage MoJ and HMCTS to extend the learnings from criminal court data to family and civil court data.

2.45 Family and civil court data have not received the same level of attention as criminal court data, but these data are equally valuable and provide important insights on these areas of the justice system in England and Wales. There are data challenges and data gaps that are unique to the family and civil courts. For example, the Nuffield Family Justice Observatory, in a recent briefing on data in family justice system, highlighted the fragmented nature of the data. Closer working between MoJ and HMCTS may help tackle these data challenges.

Related links

Ed Humpherson to Abbie Self: Review of MoJ’s criminal court statistics