Data-driven decision making is widespread at the level of data handling, but much less consistent at the level of analysis, governance and measurable business impact. In the UK Business Data Survey 2026, 86% of businesses reported handling digitised data of any type in 2025 to 2026, while 25% of businesses that handled digitised data analysed it to generate new insights or knowledge. The gap is important: having data available is not the same as using it to guide decisions.
Contents
- How widely businesses handle digital data
- How businesses analyse data for decisions
- What outcomes data use produces
- AI and automated decision-making adoption
- Data governance and decision-readiness
- Capability gaps and organizational performance
How widely businesses handle digital data
The starting point for data-driven decisions is basic data use: collecting, storing or processing information in digital form. The UK Business Data Survey 2026 found that 86% of UK businesses handled digitised data of any type in 2025 to 2026. In the second wave of the Business Data Use and Productivity Study, 83% handled some form of digital data in late 2024 to early 2025. The UK Business Data Survey 2024 reported a much higher figure, 99%, among UK businesses with at least 10 employees in 2024.
These percentages describe different survey populations and periods, so they should not be treated as a single trend line. The government sources use different survey populations and routing. Even so, each result shows that digital data handling is a normal part of business activity rather than a niche practice.
The types of data handled also matter. In 2025 to 2026, 67% of all UK businesses handled digitised personal data other than employee data, while 56% handled non-personal data other than employee data. In the second DUPS wave, 62% handled digitised personal data other than employee data and 52% handled any digitised non-personal data. The distinction is useful because customer, patient or client information creates different decision, privacy and governance requirements from operational or market data.
Storage arrangements varied in the UK Business Data Survey 2024. Among UK businesses that handled digitised data, 35% stored and processed it on their own premises, 19% used a public cloud provider, and 13% used a private cloud provider. These categories do not account for every possible storage arrangement, and the figures are not a complete measure of data maturity. They show instead that the infrastructure supporting decisions can be distributed across internal and cloud environments.
How businesses analyse data for decisions
The proportion of businesses handling data is substantially larger than the proportion reporting analysis for new insight. In the UK Business Data Survey 2026, 25% of UK businesses that handled digitised data analysed it to generate new insights or knowledge in 2025 to 2026.
Business size was associated with a large difference in reported analysis:
| UK business size | Analysed digitised data for new insights, 2025–2026 |
|---|---|
| Large | 69% |
| Medium | 49% |
| Small | 35% |
| Micro | 26% |
Source: UK Business Data Survey 2026 - GOV.UK. Percentages refer to businesses that handled digitised data.
The sector differences in the same survey were also visible. Data analysis to generate new insights was reported by 45% of UK finance and insurance businesses, 39% of information and communication businesses, 34% of professional, scientific and technical businesses, and 33% of human, health and social-work businesses. These are survey percentages for the stated sectors and period, not universal rates for every organization in those industries.
The second DUPS wave reported that 72% of UK businesses that handled digital data analysed it in late 2024 to early 2025. The same study reported that 4% engaged with big data. Because the two government studies use different survey populations and routing, the 72% figure should not be directly compared with the 25% UKBDS 2026 figure as if one were a simple increase or decrease. The measures may capture different types or levels of analysis.
A historical McKinsey survey provides a different perspective on decision-making. In a 2010 survey, 66% of surveyed executives said the business processes and decisions in their spheres of influence were becoming more data-driven. That result is explicitly dated historical evidence and reflects the surveyed executives, not a current universal business estimate. It is best used as context for the long-running shift toward data-informed management.
What outcomes data use produces
Data activity does not automatically produce better processes or new products. The UK Business Data Survey 2026 asked UK businesses that handled digitised data how often data use led to specific outcomes in 2025 to 2026. For more efficient internal processes, 8% said this happened always or most of the time, 18% said some of the time, 21% said hardly ever, and 49% said not at all.
Business size again produced a notable difference. Twenty-nine percent of large UK businesses said data use always or most of the time led to more efficient internal processes, compared with 11% of small businesses. For innovation or new business functions, 12% of large businesses gave the same always-or-most-of-the-time response, compared with 5% of micro businesses.
The survey recorded lower frequent results for innovation overall. Five percent of UK businesses that handled digitised data said data use always or most of the time led to innovation or new business functions, while 16% said it did so some of the time. These findings suggest that data collection and analysis need an operational pathway into a process, service or product before they become visible business outcomes.
The second DUPS wave measured outcomes using agreement statements rather than the same frequency categories. In late 2024 to early 2025, 55% of UK businesses strongly or somewhat agreed that acquiring, collecting or analysing data led to creating or improving services and products. Fifty-four percent agreed that data activities led to more efficient internal processes or cost savings, and 60% agreed that they led to a better customer experience. Eighteen percent agreed that data activities led to data products the business sold or licensed directly.
These results are not interchangeable with the UKBDS 2026 frequency measures. They use a different question format, study period and survey design. The practical lesson is still clear: perceived value can be common even when frequent, direct operational impact is reported by a smaller share of businesses.
AI and automated decision-making adoption
Artificial intelligence is becoming part of the data-to-decision pipeline, but adoption varies sharply by business size, sector and geography. In the UK Business Data Survey 2026, 41% of UK businesses that handled digitised data reported using AI-based technologies in 2025 to 2026. The rate was 82% for large businesses and 58% for medium businesses.
Sector and regional results included 62% of UK information and communication businesses using AI-based technologies and 48% of finance and insurance businesses. London reported 60%, compared with 36% in the East of England. These figures describe reported use during the survey period; they do not establish that AI use improved decision quality.
Businesses using AI also reported more data activity than businesses not using AI. Forty-one percent of businesses using AI collected data, compared with 26% of businesses not using AI. Thirty-seven percent of businesses using AI analysed data, compared with 17% of businesses not using AI. The figures show an association between AI use and data practices, not a causal effect of AI on data capability.
Integration and automation were less widespread. Twenty-one percent of UK businesses using AI said their AI tools were integrated into existing business systems, while 5% of UK businesses using AI reported using automated decision-making tools or systems. UKBDS 2026 asked automated-decision-making questions only of AI-using businesses, so this 5% is not a percentage of all UK businesses.
Data governance and decision-readiness
Governance is part of decision readiness because people need to know which data can be accessed, how AI may use it and which decisions require oversight. Among UK businesses using AI in 2025 to 2026, 17% said they had a policy or guidelines on AI use or development, while 5% said they had a formal written AI policy.
The business-size difference was substantial: 56% of large UK businesses using AI reported a formal written AI policy, compared with 3% of sole traders. Among businesses with an AI policy or guidance, 62% said it included guidance on AI access to business data and files.
Awareness of regulation was higher than confidence in its clarity. Fifty-three percent of UK businesses using AI said they had at least heard of AI regulatory guidance. Among businesses aware of that guidance, 19% agreed that it was clear and easy to understand. The result identifies a decision environment in which awareness may exist without broad confidence about how to apply the rules.
Capability gaps and organizational performance
The second DUPS wave found that 83% of UK businesses strongly or somewhat agreed that they had the data needed to make decisions in late 2024 to early 2025. Yet only 51% agreed that findings from data processing or analysis were available to their business. The difference points to a common practical gap: data may exist, while usable findings are less consistently accessible.
Recent IBM research reported that 92% of surveyed chief data officers agreed that AI success depends on connecting data to business outcomes in 2025. At the same time, 33% or fewer of those surveyed chief data officers believed their organizations connected data to business outcomes well. These are survey results for IBM’s stated executive population, not universal business estimates.
Older McKinsey evidence helps separate capability barriers. In its 2010 survey, 25% of surveyed executives said their companies lacked the data needed to increase data-driven decision making. Twenty percent identified a culture that prioritised experience over data as a barrier, 17% said their companies lacked skills to translate and synthesise analysis for decision makers, and 15% said they lacked the analytical skills required to carry out analysis.
The dates matter when interpreting these findings. The McKinsey observations are retained as explicitly dated historical evidence, while the UK government and IBM figures describe later periods and different respondent groups. Taken together, they map the decision chain: businesses commonly handle digital data; fewer report analysis for insight; fewer still report frequent process or innovation outcomes; and governance, integration and translation skills remain important conditions for turning information into decisions.