ONS Survey Improvement and Enhancement Plan for Economic Statistics

The Office for National Statistics (ONS) plays a vital role in delivering trusted, high-quality economic statistics that inform decisions across government, business, and society. At the heart of these statistics lie our social and business surveys – tools that provide the foundational data upon which our economic understanding is built.

In recent years, the ONS has faced significant challenges in maintaining the quality and reliability of its survey data. The COVID-19 pandemic accelerated long-term trends in declining response rates as societal behaviours and public engagement have evolved and confidentiality and security concerns have grown. These external challenges have been compounded by a long period of disinvestment in the core survey operation within ONS, as the organisation’s strategy has been to replace surveys with alternative data sources and expand its analytical portfolio during a time of tight financial controls. This has impacted the ability to maintain interviewer numbers, replace legacy systems and assess and update survey designs to meet evolving user requirements.

This Improvement and Enhancement Plan reflects the urgent need to address these challenges and a wider strategic pivot within ONS to sustain surveys into the future. It outlines a clear and ambitious roadmap to restore confidence in our survey operations, enhance data quality, and modernise our systems and methodologies. From the transformation of the Labour Force Survey to the development of a new Statistical Business Register, this plan reflects our commitment to innovation, transparency, and continuous improvement.

We recognise that high-quality statistics are essential for effective policy-making and public trust. This plan demonstrates our resolve to meet those expectations, ensuring that our surveys remain robust, representative, and fit for the future. We are confident that the measures outlined here will deliver meaningful improvements and reaffirm the ONS’s role as a world-leading producer of economic statistics.

Dr Alex Lambert OBE, Director of Social Surveys

Kate Davies, Director of Business Surveys

The Office for National Statistics (ONS) is the UK’s national statistical institute. ONS is responsible for delivering economic and population statistics that shape decisions made by national and local governments, businesses and households every day.

Integrating surveys with alternative data sources is widely acknowledged as the foundation of the future UK statistical system and some of ONS’s estimates already integrate multiple sources in this manner. As work continues to explore how we fully leverage alternative data across the statistical production process, surveys will remain the primary data source for statistical producers in the UK. ONS conducts over 80 social and business surveys, which underpin its key economic statistics. Although crime and health surveys are part of our portfolio, they are excluded from this document as they do not contribute to economic outputs.

For a variety of reasons, a number of issues have arisen on the quality of some key ONS economic statistics. These issues have been widely publicised, not least in the recent report on economic statistics by the Office for Statistics Regulation (OSR). In response, ONS has recently published Restoring confidence, improving quality: The plan for ONS economic statistics which sets out how ONS plans to address current quality issues, and the improvements to be made in each of the core economic statistics to address stakeholder feedback. The new plan for economic statistics highlights the importance of social and business surveys in providing the foundations for quality economic statistics. It provides a firm change in direction for ONS, focusing resources onto core economics statistics and prioritising investment to improve the quality of both business and social surveys.

Survey data quality challenges have contributed to the recent issues with ONS economic statistics. OSR’s report emphasises the widespread concerns from stakeholders about the quality of ONS’s survey data, which was focused on, but not limited to, the widely recognised problems with the Labour Force Survey. This survey Improvement and Enhancement Plan (IEP) directly addresses recent OSR recommended actions to demonstrate how ONS will restore confidence in social surveys and reduce risk in business surveys. With Restoring confidence, improving quality: The plan for ONS economic statistics capturing quality risks on statistical outputs and their relationship to surveys, this accompanying plan provides a further level of detail specific to surveys.

This plan builds on work over the last year to recover social survey data quality and represents a clear strategic commitment to deliver high quality business and social survey data now and into the future. It aligns with the commitments in social surveys originally made within the 2025 to 2026 ONS Strategic Business Plan and the additional attention to business surveys given most recently within Restoring confidence, improving quality: The plan for ONS economic statistics.

Our aim is that the publication of this plan will help users of our survey data understand the steps that ONS is taking to improve quality. Together with a commitment towards increased assurance and regular quality reviews of our survey portfolio, we will demonstrate transparency and delivery into the future.

We will continue to adapt the plan as our improvement work progresses, our research, trials and thinking matures, and opportunities are identified. We will publish regular updates and while some details and timings may evolve, we are confident the plan will deliver significant improvements to survey data quality underpinning core economic statistics.

Surveys are critical to the production of economic statistics, with a vast range of sources being used within these systems. Annex 1 demonstrates the relationship of ONS surveys to some of our critical economic outputs. High-quality survey data are essential as they ensure the integrity and reliability of economic statistics. Inconsistent, poorly designed, or infrequently updated surveys can introduce significant errors, misrepresent economic movements, or distort sectoral contributions.

Response rates, questionnaire design, and data validation processes are examples of the factors that play a crucial role in ensuring that survey data are both accurate and reflective of the true state of the economy. When businesses and households provide timely and accurate responses, the resulting data serve as a strong foundation for economic analysis, which can be used to inform and evaluate policy and decision-making.

Social Surveys

The production of the ONS’s core economic statistics relies on four social surveys. Three are large household surveys: the Labour Force Survey (LFS); the Wealth and Assets Survey (WAS); and the Living Costs and Food (LCF) Survey. The fourth is the International Passenger Survey (IPS), which interviews travellers at ports and airports across the country.

Like other data collection agencies nationally and internationally, the ONS has faced unprecedented recent challenges in household data collection as long-term trends in declining response rates were accelerated as a result of the pandemic. Whether it is increased confidentiality and security concerns, reduced levels of public engagement or an unwillingness to have an interviewer in their home, it is evident that in recent years a significant societal change has occurred with households noticeably less willing to take part in surveys. ONS is not alone in facing this challenge, with the whole survey sector adjusting to response rates between 10 and 20 percentage points lower than pre-pandemic. However, the quality of ONS’s social surveys has been disproportionately impacted due to their voluntary nature, significant length, and continued reliance on in-home face-to-face interviewing. These challenges highlight the importance of our ongoing Transformed Labour Force Survey (TLFS) strategic transformation programme. For the TLFS, we are implementing a shorter ‘online-first’ survey design to minimise respondent burden, and increase data quality, value-for-money and sustainability of household data collection.

The effect of post-pandemic data collection challenges on our core surveys, and the additional demand of testing the developing TLFS alongside the LFS, has significantly increased the workload on our interviewing community. This has coincided with a drop in our interviewer community capacity and capability due to long term financial disinvestment and cost savings in social surveys, recruitment restrictions and high levels of interviewer attrition. As reported in February 2025 to the Treasury Select Committee, the high-profile suspension of the LFS for labour market statistics in October 2023 occurred at a point at which 145 interviewers were available to work on the survey against a full requirement of 446 interviewers. With this mismatch in interviewer supply and demand being a key driving factor in recent data quality reductions, rebuilding our interviewer community capacity and capability has been a central focus across our improvement work.

Building on early recovery steps initiated in October 2023, a major initiative to improve the quality of social surveys was launched in April 2024 following a £19 million Treasury investment. This additional investment has been key to unlocking much needed quality improvements, given that the cost of a successfully achieved interview on ONS’s most important social survey (the LFS) has doubled in the last five years.

This plan outlines the progress improving social survey quality to-date and a very significant forward portfolio of further work aligned to further strategic investment within ONS’s latest Strategic Business Plan. However, recovering data quality is not a quick fix, particularly considering the multi-year nature of some of our survey designs and datasets. Annex 2 provides a more detailed assessment, but a high-level quality summary for each of the current social surveys underpinning economic statistics is given in the following subsections.

Labour Force Survey

Significant effort has been put into the recovery of LFS since Autumn 2023 when response reached a critical low. Our recent LFS quality update details how Wave 1 response levels are now very close to pre-pandemic levels with accompanying improvements in representativeness. Response rates and achieved interviews for Waves 2 to 5 are below target and will be the primary focus for significant additional interviewer resource and improvement in 2025 and 2026.

Living Cost and Food Survey

We are not currently meeting achieved interview targets on LCF and a key focus is currently to stabilise interviewer capacity and capability on LCF in order to bring performance back up to levels seen during 2024. We are also investing in improvements to LCF to be introduced from April 2026. These include an improved questionnaire, a further sample boost to 30,000 issued households and updated processing infrastructure and pipelines to introduce automated methods where possible.

International Passenger Survey

The IPS is meeting expected response levels. Improved methods and processing were implemented in July 2024 as part of the Travel and Tourism Reform Project (TTRP). These are expected to be refined throughout 2025 before the closure of Phase 1 of the TTRP. No significant concerns with IPS quality.

Wealth and Assets Survey

We are significantly below achieved interview targets on the Wealth and Assets Survey and have also experienced a significant delay with data processing and therefore timely publication of wealth statistics. These statistics have recently had accreditation removed and OSR have reviewed and provided recommendations for key improvement areas. Focus in 2025 and 2026 is on recovering field capacity and capability, as well as finishing the development of an automated processing pipeline to improve timeliness.

With a commitment to increased assurance and transparency, ONS will publish a rolling programme of formal reviews of its surveys. These end-to-end reviews will understand the latest user requirements, assess the performance of the survey in meeting these requirements and the current risks, improvements required and opportunities. Future versions of this plan will reflect the outcomes of these reviews and the actions that will be taken as a result.

In determining the quality of social surveys, response rates have traditionally been used as the key measure of survey quality. This remains a strong focus for us, but with sample sizes, boosts and field effort now regularly under review and scrutiny, monitoring of achieved sample sizes has become increasingly important. With lower response rates post-pandemic, there is also an increased need to ensure social surveys remain representative, as increasing sample sizes does not address the critical risk of non-response bias.

For the TLFS, representativity is actively monitored and the survey employs an adaptive survey design to reduce bias. The adaptive survey design allows field interviewer resource to be focused on underrepresented subgroups of the population (for example, urban, more deprived areas and addresses more likely to be occupied by those under 45 years old). This bias reduction approach is more challenging to implement on our other social surveys due to limitations of legacy systems; however, our recent LFS publication directly addresses this issue and we aim to further understand survey bias across our portfolio and improve our data and infrastructure to allow more regular monitoring in the future.

Business Surveys

The core production of economic statistics relies on 65 of the 74 business surveys run by the ONS. Those surveys not used in core economic statistics are collected on behalf of Other Government Departments or Devolved Governments.

Chief among the surveys used in economic statistics are:

• Annual Survey of Hours and Earnings, Business Register and Employment Survey, and Monthly Wages and Salaries Survey, for Labour Market;  

• Short terms surveys such as the Monthly Business Survey, Construction and Allied Trades, Retail Sales, and Business Impacts and Conditions survey;

• Quarterly structural surveys such as Quarterly Capital Assets, and Quarterly Stocks surveys;

• Annual structural surveys, for example, the Annual Business Survey, Annual Purchases Survey, Products of the European Community (PRODCOM) Survey, and Annual Survey of Goods and Services;

• Producer Price Indices, Services Producer Prices, and Import and Export Price surveys, for the Prices theme;  

• Quarterly and annual Foreign Direct Investment surveys and quarterly and annual International Trade in Services surveys for Trade; and

• Specialist financial sector surveys, Financial Services Survey.

A list of surveys and the outputs where they are used can be found in Annex 1.

All business surveys, except the Business Impacts and Conditions Survey, and UK Innovation Survey (run on behalf of the Department for Business and Trade) are mandated under the Statistics of Trade Act 1947. Most of the surveys are digital by default, using an electronic questionnaire (eQ) or Secure Electronic File Transfer. This has resulted in quicker response from businesses and reduced the time they take to complete a questionnaire, therefore reducing burden.

For business surveys, achieving the target response rate does not guarantee quality, and therefore, where possible, an estimate of coverage of the population is used as a second measure. A third is clearance rates (used internally), which determines the number of returned data items validated.

While business surveys are mandated, it does not guarantee response, and considerable effort is expended on both obtaining and validating returned data (confirming whether the data supplied is correct). Economic Statistics teams responsible for outputs provide target response rates, coverage of the population (measured by turnover) and clearance rates.

At present, the collection team is carrying a number of vacancies, affecting the team’s ability to meet these targets. Team members also typically work on one survey at a time, a siloed approach which results in businesses receiving multiple calls, and where the data they return to different survey requests is not always checked for coherence.   

The data requested can often also be complex, based on both business accountancy and National Accounts concepts. While some more experienced members of the collection team are confident engaging businesses on such data, there is a need to increase the capacity of this expertise.  

Even with teams at full capacity, increased skills and more, non-response remains an issue. It is vital that we understand why a business does not respond to all survey requests and we are addressing this through business engagement and a non-response strategy.

The statistical and survey design of each business survey has been created to ensure quality, with a representative sample design. Many of these designs are now out of date and need review. Like social surveys there is a need to embed a rolling review of business surveys. These end-to-end reviews will understand the latest user requirements, assess the performance of the survey in meeting these requirements and the current risks, the statistical methodology, improvements required and opportunities. Future versions of this plan will reflect the outcomes of these reviews and the actions that will be taken as a result.

The introduction of a new sampling frame for business surveys, the Statistical Business Register (SBR) with an increased coverage of small and micro businesses, also necessitates the need to review and redesign all business survey samples.

Imperative to survey quality and ensuring we are using the most appropriate methods and survey designs, is a move from our current technological estate, deemed as legacy technology, to new platforms such as the Data Preparation Platform (DPP) and Repeatable Analytical Pipelines. Any changes to survey design and methods are limited while a survey remains on legacy technology.

Whilst globalisation is not a new phenomenon, it does bring with it data collection challenges for National Statistical Institutes, including the ONS. Business surveys continue to adapt to the complex international guidance set for reporting global production, as there is currently no survey or administrative data source that captures the economic ownership of goods. To address this, the Large Case Unit and Globalisation team continue to work collaboratively with international businesses on improving data collection methods.

The substantial progress recovering data quality on the LFS wave 1 made over the last eighteen months demonstrates that much can be achieved with sample boosts, higher value respondent incentives and increased interviewer resources, all supported by appropriate additional investment. Building on this progress and with further strategic investment within ONS, the data quality improvement priorities over the year ahead are to recover LFS waves 2 to 5, followed by the LCF and then WAS.

With clear quality improvement objectives in place, it is recognised that additional interviewer numbers and sample boosts will increase achieved samples but only recover data quality and improve overall survey health so far. Addressing non-response bias issues or the inherent challenges of operating voluntary household surveys requires a more comprehensive approach to put social surveys on a sustainable footing into the future.

Our recent work in developing a new survey strategy defines survey sustainability as having flexible survey designs to meet evolving needs, a resilient and flexible workforce, high quality and reliable data, technological adaptability, inclusive and accessible methods leading to representative statistics, and effective quality monitoring – all offering the best use of public funds.

To achieve both urgent data quality improvements for economic statistics and progress towards long term sustainability, this improvement and enhancement plan outlines a comprehensive set of activities against four themes:

1. Sustainable Field Operations – focussed on improving capacity and capability across our interviewer community;

2. Refreshed Citizen Relationship – to improve trust, brand recognition and understanding of the importance of our surveys;

3. Improved Survey and Statistical Design – redesigning our surveys to be more streamlined, reduce respondent burden, and improve accessibility for the population; and

4. Technology Improvement – focussed on improving our systems and processes, taking opportunities to improve efficiency and agility of our technology.

The progress and plans against each theme are summarised in the following subsections, with full details of our projects and milestones in Annex 3.

Sustainable Field Operations

Improvements to Date

We have listened to our interviewer community to understand how their role has evolved since the pandemic and the underlying reasons for increased attrition. They vividly describe the challenges of accessing gated communities, the robust daily refusals on the doorstep, rising cost-of-living challenges on low pay, high workloads and a lack of career progression opportunities. This information has guided our actions to date and the forward plan presented here.

Extensive work over the last year has increased our face-to-face field community from its lowest size of 517 interviewers in March 2024, to over 800 in June 2025. This has been achieved by adopting a hybrid approach of recruiting permanent and agency colleagues and outsourcing our field recruitment to an external resourcing partner.

To support our new interviewers and improve retention of our experienced colleagues we have implemented a new progression and promotion pathway, through the creation of 42 new Coaching and Learning Partners roles at the higher Executive Officer grade. These roles retain a level of interviewing while building the capability of interviewers working within our tougher respondent environment. We have also sought to shift the culture of our field interviewer community, introducing continuous improvement and coaching approaches, and empowering line managers and interviewers to enable operational decisions at the lowest possible level.

Collectively, these measures have resulted in a marked increase in interviewer engagement, reduced long-term absences and reduced interviewer attrition, which has dropped from a peak of 32% in August 2023 to 25% in May 2025. Attrition levels will remain a focus for further improvement in our plans for 2025 to 2026.

Plans for 2025 to 2026

We aim to increase our face-to-face field interviewer resource to 1,023 by the end of the 2025 to 2026 financial year to fully resource our surveys. Only with fully resourced surveys is it possible to improve data quality by ensuring every household in the sample receives appropriate interviewer attention (measured by adherence to calling patterns, a set of daytime, evening and weekend calls designed to maximise response and reduce bias). Increasing face-to-face interviewer resource levels is a key ONS strategic priority, but actual final numbers will be dependent on the performance of our external resourcing partner and whether our attrition assumptions built into our recruitment models remain valid.

In response to the increasing sample sizes on the LFS, the size of the telephone interviewer community will be increased this year to 206 interviewers. An ongoing Telephone Services Review is identifying key areas for development within this operation.

For both our face-to-face and telephone interviewers, the ability to successfully persuade respondents to take part in our voluntary surveys increases during their first three months in role. Our new interviewer learning and coaching partners aim to support this process, but an understandable lag exists between a field capacity increase and benefits to survey performance. With the priority being LFS waves 2 to 5, then LCF and then WAS, the aim is to reach monthly target achieved interview levels for each survey within the following timescales:

• LFS (all waves) – November 2025

• LCF – January 2026

• TLFS (new short core design, in-home supported completion, data rotation and increased sample sizes) – January 2026

• WAS – March 2026

To further improve recruitment and retention across our interviewer community, we plan to trial a new recognition scheme that enhances the benefits offered to interviewers. The new scheme will better recognise and reward the valuable work interviewers do and aims to increase the number of achieved interviews for our critical surveys. It will also support the development of skills and capability across the entire interviewer community.

Refreshed Citizen Relationship

An increasingly challenging respondent environment has driven a response rate decrease of between 10 and 20 percentage points since the pandemic and above inflationary cost growth. With appropriate investment, ONS has many levers it can draw on to improve social survey quality, and the aim through the ‘Refreshed Citizen Relationship’ work is to directly address an increased reluctance to complete surveys. While citizen engagement, targeting marketing and communications campaigns are routinely employed by many organisations, this area of work within the plan has some of the highest levels of complexity and uncertainty. It remains a key area to improve and research, but realism is required as to the ability to cost-effectively reverse societal trends seen across the whole survey sector.

Improvements to Date

We increased the respondent incentive for taking part in our surveys across the survey portfolio in 2024 and 2025, and have been exploring more novel incentives approaches to improve response. For example, we have been considering the use of early bird incentives to encourage earlier responses on the TLFS and differential incentives to encourage response from underrepresented groups of society.

We have undertaken a pilot communications campaign designed to address barriers about survey legitimacy, data security, and brand recognition. This has been delivered in Birmingham, a local authority area with historically low response rates, particularly among young people, ethnic minority groups, and economically disadvantaged communities. The pilot draws on insights from the Census 2021 communications campaign and industry best practice. It includes use of broadcast video and local digital billboards to target messages to survey respondents using postcode data. We are now undertaking a comprehensive analysis of the data to look at the impact of the campaign on awareness, trust, and response rates.

Plans for 2025 to 2026

We will optimise our communication approach and deliver an informed decision as to whether to introduce targeted communications campaigns in areas of particularly low response. We will embed findings from recent inclusivity research to reach all sections of society including under-represented communities. For example, we will review our methodology on how best to engage respondents with English as a second language. We will also trial differential incentives, targeting younger age groups and ethnic minorities through incentives to improve response from underrepresented groups.

As highlighted by the Independent Review of the UKSA, to mitigate the fall in response rates there is a need to consider legislation to mandate citizens to participate in some key Government household surveys. Mandation helps facilitate high response rates on the census and our business surveys, and is routinely used to gain higher response rates internationally. However, continuous household surveys are not covered by the Census Act 1920 so if the Government did decide to take this approach new primary legislation would be required. In line with the review recommendation, UKSA has started work considering the options around mandation, including an initial Board discussion. Before deciding on a course of action, further consultation and evidence gathering would be required to understand the impact, including public acceptability, in a UK setting. We will explore this further through 2025 and 2026.

Improved Survey and Statistical Design

Improvements to Date

The TLFS is the ONS’s strategic solution to the quality challenges on the LFS and we have invested substantially in transforming the LFS to a multi-mode online-first design. Providing an ‘online-first’ approach to voluntary household surveys (backed up by interviewer contact if needed to reach all parts of society) is more flexible and attractive to citizens, boosting response rates to around 40%. It significantly boosts the number of achieved interviews and the data quality. The new TLFS design addresses long-standing survey design and questionnaire length issues and uses an adaptive design to target field interviewer resources in under-represented areas to reduce bias.

A complex and challenging transformation, the last year has seen substantial work on the TLFS as results from the parallel run of the LFS and TLFS became available. At a headline labour market statistic level, the LFS and TLFS results were similar with reasonable hypotheses around any differences observed. Larger differences occurred, though below headline level, with these differences relating to the biases resulting from partially completed responses to the online TLFS survey. To address this issue, an improved TLFS design has been developed and tested with respondents, as reported in our detailed technical design review. This design uses a substantially shorter and simpler core survey, as well the opportunity for non-responding households in under-represented areas to be supported to complete the survey in-home by an interviewer. The new design also rotates data between the longitudinal waves of the TLFS survey, further reducing the respondent burden.

The last year has also seen significant changes to the International Passenger Survey as part of the successful Travel and Tourism Reform Project (TTRP). While the IPS is a key source of travel and tourism data, it also feeds into economic statistics by capturing expenditure data from travel and tourism, which contributes to GDP calculations. With the aim of improving data quality and sustainability, the TTRP has implemented significant survey, operational and methods changes, as well as creating partnerships with other organisations to ensure efficient data collection. Since February 2025 the project has continued to develop and quality assure the new methods, and recently published Travel and Tourism Quarter 3 and 4 data.

Plans for 2025 to 2026

On the TLFS, work is now in the final build and test stages to implement the new short ‘core’ labour-market-focused survey from July 2025. The additional components of the new TLFS survey design will be subsequently implemented over the following six months (see our technical design review article and Annex 3, TLFS) ready for a full data quality assessment in July 2026.

We will also complete the current phase of the TTRP for the International Passenger Survey. During 2025 and 2026 we will refine and finalise improved methods and processing that were initially implemented in July 2024, before the closure of Phase 1 of the TTRP by December 2025.

Given the current quality concerns on the LCF survey, it is the focus for significant work over the coming year, specifically:

• An updated questionnaire with the aim of introducing COICOP2018 (see section A1.2: Classifications in The plan for ONS economic statistics), incorporating critical questions from the recently stopped Survey of Living Conditions (SLC) and reducing overall respondent burden;

• Updated processing infrastructure and pipelines, including automated receipt coding; and

• Pending financial approvals – a further sample boost to 30,000 households from April 2026.

On the WAS the interviewer resource will be expanded by March 2026 to effectively support the annual issued sample size of 33,000 households to ensure monthly minimum achieved sample requirements are met. The full benefits and impact on yearly achieved sample sizes will then flow through during 2026. However, the impacts will take longer to flow through to ONS wealth statistics, given the biennial nature of WAS. More broadly, following the recent OSR review and accreditation decision, ONS is developing a plan for the collection and processing work required to bring wealth statistics back to the standard needed for our key users, which we intend to publish in the Autumn.

A new Strategy, Research and Innovation Hub will be established in Summer 2025 to explore the use of alternative data sources and artificial intelligence integration within our surveys. The Hub will also drive forward and improve national and international collaboration, drawing on the latest internal and external thinking for survey research ensuring best practice.

Technology Improvement

Improvements to Date

Data collection and production systems need ongoing investment and development to exploit new and emerging data, tools, and techniques, adapting to meet new user needs.

The TLFS programme has enhanced the technology used to process survey returns by leveraging open-source technologies such as PySpark, R, and SQL to develop a centralised, modular, and user-friendly system known as the Configurable Integrated Processing and Assurance System (CIPAS). CIPAS is operational for processing TLFS data and is being built within a secure data platform in line with best practice and industry standards, ensuring the system is future-proof, capable of integrating third-party applications and able to integrate with the Integrated Data Service. We have also developed a Management Application, a user-friendly interface that lowers the barrier for colleagues to create and update Reproducible Analytical Pipelines, which will lead to increased efficiency and productivity, and enabling faster delivery of higher-quality outputs.

Plans for 2025 to 2026

To reduce our reliance on legacy systems and unsupported technology as well as increase efficiency, through 2025 we will develop a technology roadmap. This will include consideration of increased automation, use of questionnaire tools, upgrades to existing technology for questionnaire builds, and requirements for future tools that allow a more dynamic allocation approach for our interviewing community.

We will investigate opportunities for use of novel collection tools and artificial intelligence (AI) within our systems. For example, exploring new ways of helping respondents to provide accurate detail for complex variables in online mode – like Standard Industrial Classification and Standard Occupational Classification, using AI. Early indications from the prototype are promising with significant work required to refine and implement.

On the LCF survey, alongside exploration of a digital diary, we are also developing a tool that can read receipts automatically, reducing the need for manual receipt coding. We aim to implement this tool in April 2026, aligned to the wider LCF survey changes.

With the research and tests within this plan maturing, further details on medium and longer-term improvements will be included in future versions as the details on our forward funding profile are confirmed. At a high-level, across the timeframe of the Spending Review (SR), our main priorities will be to:

• Complete the development of the TLFS;

• Decommission the LFS; and

• Stabilise our expanded interviewer field community whilst enhancing capabilities and improving retention – enabling quality improvements across the breadth of social surveys.

More broadly, we recognise that a new Household Financial Survey (HFS) design is required to ensure we can sustain financial survey collection in the future. This should be complemented by exploring the opportunities of both administrative data to replace survey questions, and online-first data collection. The development of HFS is currently unfunded but we aim to exploit opportunities during this forthcoming SR period to conduct initial research, design and options analysis, with the aim of delivering a clear way forward ready for future SRs.

Building on the design for the TLFS and as part of our commitment to driving up quality, we plan to explore adaptive and responsive survey designs for all household surveys, that focus on sampling, field management and communication strategies for under-represented population groups. To do this, we will also need to investigate options for a centralised sample management system to deliver a dynamic system for allocation of work to interviewers.

In addition to stabilising our expanded field community, we will also look to establish specialist interviewer teams, supported by community engagement expertise, for under-represented communities, such as Gypsy or Irish Traveller communities.

To reduce our reliance on legacy technology, we will upgrade all social surveys onto the latest version of our Blaise data capture software, in-line with a new technology roadmap. We also aim to embed early projects that explore the use of alternative data sources and the integration of artificial intelligence, supported by our new Strategy, Research and Innovation Hub. This work will include using alternative data sources, where possible, to:

• Rationalise questionnaire content where it has been identified that user requirements are still being met;

• Tailor data collection procedures for different population subgroups in our sampling designs, field management and communication strategies (adaptive design);

• Adjust survey data collection protocols as a means of driving a more efficient use of our field capacity; and

• Transform our processing systems (particularly paradata and respondent feedback).

Although currently unfunded, based on research findings and decisions made this financial year, we may look to implement targeted communication campaigns to build trust and address security and confidentiality concerns amongst citizens.

Business Surveys Directorate are responsible for the collection and transformation of a suite of 74 surveys, 65 of which are used as input sources for economic statistics such as monthly Gross Domestic Product, Business Prices, and Average Weekly Earnings. Others are collected on behalf of Other Government Departments and Devolved Governments. A list of surveys and the outputs where they are used can be found in Annex 1.

To address the improvements and enhancements needed to increase the quality of business surveys, the Business Survey Strategy was created in 2023. The strategy sets out a vision for the future of business surveys to transform data collection to better measure the modern economy and adapt to the future. Specifically, it set out three core elements:

i. New integrated and modularised survey designs, moving away from silo based operations to a business centred and quality focused model, with sustainable technology, methods, and processes, with the ability to flex and react to changing user requirements.

ii. A new business engagement model to improve ONS relations with businesses, critical to response and quality, through improved education, communication, and support, shaping our future operating model to increase capability across our core collection operations.

iii. New innovative methods for data collection to include new technologies to reduce our reliance on legacy and automate our collection, removing blockers to adding value and improving efficiency.

As stated in Restoring confidence, improving quality: The plan for ONS economic statistics, fundamental to the improvements and enhancements needed for Business Surveys are to improve the foundations from which are the surveys are sampled, collected, and processed, this includes replacing the Inter-Departmental Business Register and the legacy technology used for collection and validation of business survey responses.

The Statistical Business Register

The existing sampling frame for business surveys, the Inter-Departmental Business Register (IDBR) brings together data from administrative and survey sources to provide a list of businesses within the UK. The IDBR is used to sample the business surveys that underpin core economic statistics and is widely used by Other Government Departments to sample businesses. Built on a legacy system, it presents challenges to delivering emerging and future user requirements.

The IDBR meets the current System of National Accounts and Balance of Payments requirements and businesses are classified to the current Standard Industrial Classification (SIC2007). But it will not support new international standards and future user requirements:  

• It has incomplete coverage of small and micro businesses: only businesses that have a VAT, PAYE and Companies House registration are included; 

• Samples can only be stratified by employment, turnover or a combination of the two – yet for financial surveys, to take one example, these variables are poorly correlated with the objective of the surveys; and

• It is not possible to incorporate additional administrative data to ensure more relevant sample designs. 

We are addressing these challenges through the creation of the Statistical Business Register (SBR), the addition of additional data sources (for example Corporation Tax data), extending the business population by ensuring all businesses with a Companies House registration are included and offering more flexible sample designs, and improving the quality of economic statistics and those produced by Other Government Departments. The addition of Self-Assessment data when available for use will further enhance the detail the register holds on the self-employed.   

To adopt the new Standard Industrial Classification (SIC2026), the SBR needs to be dual coded with the existing SIC and the new. Changes needed to meet new System of National Accounts (SNA) standards can also be incorporated.

The sampling and methods changes within the SBR along with the changes needed to business surveys to make use of these, will be implemented in an integrated way across our core economic statistics.

Improvements to Date

Through a phased implementation approach, we have launched a read only version of the SBR. An advanced search and view facility provides users with greater efficiency when searching for a business as well as more structural detail. We are now able to create reference list samples for the Building Materials surveys we run on behalf of the Department for Business and Trade. The addition of Corporation Tax data in the SBR also means we can select a more relevant sample for the Financial Services Survey.

Plans for 2025 to 2026

Over the next year, we will increase the business population within the SBR by incorporating the full Companies House data, providing better coverage of small and micro businesses. We are working towards the selection of large stratified random samples and are testing this with the Business Impacts and Conditions Survey.

In collaboration with Economic Statistics and Environment Group, we plan to develop an implementation plan for adoption of the SBR across core economic statistics, starting with Core Annual Surveys that provide the foundation of the National Accounts and sequencing future surveys to the Quarterly Gross Domestic Product timetable.

To address the requirements of economic statistics, our plans for the SBR extend past March 2026 into future years.    

By March 2027

The SBR will be dual coded with the current Standard Industrial Classification SIC2007 and the new SIC2026. In the first instance this will provide core economic statistics outputs to assess the changes in Business demography, and secondly it will provide the opportunity for coherent outputs with other countries who have adopted SIC2026.

When dual coded, surveys will transition to the SBR. There is a need to consider the technology used for each survey and it is expected that the collection and validation of these will be through the Data Preparation Platform, which is set to replace the legacy platforms currently in place. More information on this work is provided in the Legacy Platforms section.

The Business Register and Employment Survey, will provide the SBR granular data on businesses, providing information on the local units, for example location, employment, type of activity of site. We are planning to include the processing and therefore integration of BRES by March 2027, which will also facilitate moving this survey from paper to online collection, further reducing our legacy platforms estate.

By March 2028

A review of sample designs including parameters of all business surveys is planned to be completed by March 2028. This will incorporate the extended business population and the addition of new data sources, for example Self-Assessment data, to the SBR.

By the end of the 2027 to 2028 financial year, we plan to have transitioned all business surveys to the SBR.

Legacy Platforms

Improvements to the statistical methodology and survey design can be facilitated through the removal of legacy platforms. The technological estate for business surveys is largely based on legacy platforms. These platforms are maintained to ensure they remain operational but are out of date, costly to support and prohibit changes to the underlying methodology and therefore survey and statistical design.

While in the main these platforms are stable, if they fail, we are reliant on third party support to fix and restore systems, to allow the collection, validation and process of survey data to recommence. This can often lead to resource being diverted to meet the quality targets set by Economic Statistics, putting other surveys at risk.

In addition, these legacy platforms inhibit our ability to create a sustainable operation, as they rely on manual intervention and operations to support the work needed and therefore are inefficient.

As stated in Restoring confidence, improving quality: The plan for ONS economic statistics, the ONS plans to adopt the new Standard Industrial Classification (SIC2026), SNA and Balance of Payments Manual (BPM). Validation and processing platforms, as with the SBR, will need to be dual coded with the existing and new SIC. This is not possible through our current legacy platform, and neither are other changes from SNA and BPM, further strengthening the case for removal of legacy platforms.

A programme of work is in train to replace these platforms with the strategic cloud-based Data Preparation Platform.

Improvements to Date

The ONS has re-platformed the Retail Sales and Business Enterprise Research and Development surveys, is in the process of adding three more surveys and has plans to tackle employment surveys as a next step. In addition, the BPM system has replaced a large majority of our Lotus Notes legacy estate, with Power Apps being investigated to replace the small percentage remaining.

Plans for 2025 to 2026

Alongside completing work to move the Monthly and Quarterly Business Surveys and the Construction and Allied Trades survey to the DPP and Data Access Platforms, we are prioritising the replacement of end-to-end legacy technology for the Annual Survey of Hours and Earnings (ASHE) and other Employment surveys.

A plan for other surveys to move to DPP will be produced and aligned with plans to transition to the SBR and programmes of work such as the Future Business Surveys and Statistics Programme.

As well as providing the foundations for quality core economic statistics, the SBR and Legacy removal work supports the business survey strategy. To achieve our full aims of the strategy, further work is needed, and the following provides details of how we will address the business survey strategy. The full business survey delivery plan for 2025 to 2026 is outlined in Annex 4. Work beyond April 2026 is subject to Spending Review discussions with HM Treasury.

Sustainable Collection Operation

A programme of work to reduce waste in the collection process was conducted, resulting in a focus on educating the businesses we collect data from on individual surveys, switching response chasing to the use of secure messaging and email, and prioritising the validation of responses by impact to results, allowing business surveys to meet an efficiency target of 15% of their budget.

To achieve a sustainable operation, we are investing in a new organisational design that reduces the burden on the businesses who provide data and increase the skills of people. While a change in ways of working was successful, the current organisational design means our people work on individual surveys. Businesses with an employment of 10 or more can be selected for more than one survey, leading to them receiving multiple questionnaires a year and multiple contacts. The data requested can often be complex and our people are required to understand business accountancy as well as National Accounts concepts.

Improvements to Date

Recognising the largest businesses have complex and global structures and can receive circa 500 questionnaires from us a year, the Large Case Unit (LCU) was created in 2018. Its goal is to improve the quality of data from businesses, which contribute significantly to their respective industries in ONS outputs.

The LCU offers an account management approach to data collection and a single point of contact. The unit collects and validates all surveys ensuring a coherent data picture for each business in scope. Any data discrepancies are picked up at the start of the statistical value chain and addressed before entering production cycle. A deeper understanding of the business models and motivations reduces the need to query and response chase survey data, making efficiencies for businesses and ONS.

Based on the success of this model, in April 2024 we successfully delivered an account management proof of concept covering large businesses with simpler business structures, further improving quality and relevance of communications with businesses.

Plans for 2025 to 2026

Based on the complexity of the surveys, we are developing career pathways based on themes such as financial surveys, where we are looking to increase our business accountancy skills; and trade where we will develop a greater understanding of international and interregional trade and will also look to extend to employment surveys. These career pathways will also increase the analytical skills of data collection operatives.

We will also increase the number of businesses dealt with by the Account Management Unit from 24 to approximately 500 and LCU from 125 to approximately 300.

Enhanced Business Engagement

To improve and enhance our business engagement, our vision is to create a partnership built on trust and mutual respect with the business, working with them to increase response and a better understanding of the UK economy, now and in the future.

Non-response negatively impacts the quality of data, it is therefore important we understand why businesses do not respond to surveys and what can be done to increase response, through:

• Engagement and education of businesses, to explain the surveys we collect and how they are used as well as the importance of their data;

• Targeted response chasing; and

• Better communications, questionnaire design and reminder letters

When businesses do not respond to three periods of a survey an enforcement process is used. The initial stages of this process generally result in the business returning the surveys required. Non-compliance is a criminal offence heard by a magistrate’s court, fines are levied at £2,500 and completion of the questionnaires is still required.

Improvements to Date

By understanding different respondent personas (Annex 5) we have focussed on increasing response from late and non-responders. A redesign of questionnaires, and letters has reduced the assumed reading age, front loaded key information and made use of colours, bold text and icons for emphasis.

A comparison of response demonstrated that the new style communications increased overall response. For some surveys, an increased first-time clearance, when the data returned passes validation gates when received without follow-up, was evident, improving the quality of data collected.

Plans for 2025 to 2026

Creation of a non-response strategy which will include tailored collection services to respondents and ensure a business-centred approach to the design of our questionnaires and all communications. This work will also include a refreshed approach to enforcement activities while maintaining the obligations we have under the Statistics of Trade Act.

This will result in a more tailored accessible service, flexibility, richer feedback on our questionnaires and increased quality, improving the ONS reputation, providing greater transparencies and efficiencies within the collection operation, while reducing the burden on businesses.

In addition, we are actively increasing our engagement with businesses, ensuring they understand the importance of our surveys and how they can use the data we publish, as a means of encouraging response.

Improved Survey and Statistical Design

Integrated and modularised business surveys will reduce the number and overlap, resulting in reduced respondent burden. The number of business surveys and the questions asked has increased over time, to meet the needs of economic statistic users. Despite most of these surveys now being collected digitally, their design and integration with our legacy systems makes it difficult to react quickly to a changing economic landscape.

Improvements to Date

To reduce the number of business surveys and improve the quality of our economic statistics, the Future Business Surveys and Statistics (FBSS) programme has been created. This programme has three strategic aims:

• Improved measurement of the UK economy today, driven by future data user needs through more coherent and valid business statistics;

• Data collection with a business-centred approach to design and engagement and better use of alternative data sources; and

• An efficient, adaptable and sustainable system for producing business statistics. 

Progress to address these challenges is now well underway with plans to deliver a suite of around 40 surveys with an updated modular design.

To date we have created a consolidated business survey portfolio, reducing the number of surveys where possible and implemented an improved survey and questionnaire design for the Business Enterprise Research and Development Survey.

Plans for 2025 to 2026

Across the next year we aim to complete the survey and statistical design of the Integrated Annual Business Survey integrating all annual surveys over a phased implementation across the next three to four years, and we plan to adopt other survey themes or periodicities into the programme, in alignment with Economic Statistics plans. Starting with the Annual Surveys facilitates the adoption of the new System of National Accounts and new Standard Industrial Classification.

Technology and Automation

To improve the efficiency of the business process, underlying methodology and the flexibility, stability and sustainability of business surveys, it is paramount we invest in technology and innovation. In addition to our work on the SBR and Legacy platform removal we are focussing on:

• Continuous improvement of the online collection of business surveys and moving the remainder of paper-based surveys online; and

• the use of automation and use of AI, both across our technological estate and within the data collection process.

Digital Data Collection

Digital data collection improves the speed of response from businesses to our surveys and results in less data failing validation tests, reducing the number of calls made to businesses, therefore reducing burden, and provides the opportunity to provide a dataset with increased response and coverage to output areas. Approximately 2.5 million questionnaires are sent to a sample of the UK business population per year, historically dispatched on paper.

Improvements to Date

To date 1.1 million paper-based survey questionnaires have moved to online collection. The drive towards digital data collection also enabled our rapid response to establish the Business Impact of Coronavirus survey during the pandemic, adding an extra million questionnaires to the online platform, resulting in 2.1 million survey questionnaires online.

Plans for 2025 and 2026

There has already been progress in 2025. For ASHE, any business with more than five employees received an online collection instrument in April 2025, leaving only those with less than five on paper. This has resulted in quicker response from these businesses, providing more time for quality assurance of the data returned.

The Producer Price Survey was successfully moved online in June, and the Services Produce Price Survey will move to online collection in July. The remaining Producer Price Surveys (Imports and Exports) will be moved to online collection by September of this year, further reducing our legacy estate through decommissioning the Telephone Data Entry system currently used by these surveys.

The remaining paper-based surveys are largely complex in their nature will all be moved to digital data collection, either through the creation of the SBR, the FBSS programme or legacy reduction.

Automation and Artificial Intelligence

Automation and AI present the opportunity for a better organisational design and business process, leading to efficiencies and a sustainable operation, replacing the manual process undertaken as part of the collection activity.

Improvements to Date

The application and use of large language models to improve the quality of coding for the ASHE surveys have been adopted. The resulting application, ClassifAI, improves the quality of coding and creates an efficiency saving that can be reinvested into the collection and validation of ASHE data, again improving quality.

We have set up a team to work alongside the ONS Data Science Campus and Digital Services Automation team, to investigate where automation and AI can be used to replace manual processes and implement these across the operation.

We have developed a number of co-pilot agents to generate respondent communications, including emails and letters, resulting in standardised communications and efficiency savings.

Automation is used increasingly across Business Surveys both tactically and strategically. Currently automation is used to process PDF requests, consolidate complex respondent commentary into a dataset usable by front line staff, and support the production of operational MI.

Initial discussions with businesses on automated collection of data have been held. LCU businesses are engaging with us to discuss how this can be achieved, and we recognise that some accountancy packages for example, have built in applications that can provide the data needed for some of our more complex surveys.

Plans for 2025 to 2026

ClassifAI will replace the existing Standard Industrial Coding tool for the Business Register and Employment Survey, and we are investigating its use to code products on other surveys.

Business Surveys are looking at how AI can be used to support Dead Letters – each year a significant amount of post is returned unopened, which impacts business surveys’ response rates – replacing the manual process of searching for businesses’ registered addresses on the Companies House website, with an AI based solution that is ten times faster. A supplementary project is using AI to help teams get information in an efficient and structured way on business returns, which is otherwise difficult to find in a defined structure. In the short term the Innovation Hub will work with the relevant teams up and downstream to support, develop and implement a solution.

We will continue to investigate where automation can replace manual processes within the processing of business surveys and implement these when fully tested and agreed.

The work described in the previous sections sets the foundations for better quality economic statistics. In addition to the future work planned for the SBR and legacy platform removal, work will also be required to ensure we meet all aims of the Business Survey Strategy. Further details on how we will meet these aims will be described in future versions of this plan, alongside updates on what has been achieved.

At a high level, across the timeframe of the SR, the key priorities will be the removal of legacy platforms for all business surveys, ensuring all are digital by default, statistical methodology and survey design reviews and a reduction in manual processing. The latter will have a positive impact on our ability to increase the size of as well as the number of businesses within our Large Case and Account Management units, further increasing quality.

Working with Economic Statistics, we will continue to adjust and add to plans to ensure their needs are met, including additions or changes to their plans. Fundamental to this work will be the adoption of the new System of National Accounts and Balance of Payment Manual, as well as classification updates such as SIC2026.

We will continue to look at how automation and AI can support and enhance the work of business surveys, and continuously improve the technological estate, including the SBR.