Predictive analytics is measured most consistently through business data-analytics activity: the use of internal or external data to identify patterns, trends, and insights for predictions and better decisions. In 2025, 39.85% of EU enterprises performed data analytics either internally or through an external provider, while 33.02% did so with their own employees (Eurostat, Digital economy and society statistics - enterprises). These figures describe the closest official measurement of predictive-analytics activity, but they are not identical to the narrower use of predictive models.
Table of contents
- How widely enterprises use analytics
- Transaction and customer data
- Digital traces and online behavior
- Location, sensors, and operations
- Open and satellite data
- Business size and geography
- AI context and measured business outcomes
How widely enterprises use analytics
Eurostat’s 2025 enterprise measure includes organizations with more than 10 employees or self-employed persons in the survey scope. It defines data analytics as the use of technologies, techniques, or software to analyze internal or external data and extract patterns, trends, and insights for conclusions, predictions, and better decision-making. That definition makes the figures useful for understanding predictive analytics adoption, while the scope limitation matters: analytics activity can include descriptive and diagnostic work as well as prediction.
In 2025, 33.02% of EU enterprises used their own employees for data analytics. A further 13.85% had data analytics performed by an external enterprise or organization. Counting either internal or external activity, 39.85% of EU enterprises performed data analytics. The measures overlap because an enterprise may use both its own employees and outside providers (Eurostat, Digital economy and society statistics - enterprises).
Eurostat also reported a broader technology measure. In 2025, 63% of EU enterprises used at least one of artificial intelligence, sophisticated or intermediate cloud services, or data analytics. The corresponding share reached 82% in Denmark, 81% in Finland, 80% in the Netherlands, 78% in Italy, and 77% in Sweden. Bulgaria recorded 27% for the same broad measure, the lowest proportion reported in that comparison (Eurostat, Digital economy and society statistics - enterprises).
These percentages should be read as indicators of organizational capability and activity, not as a direct estimate of the quality, accuracy, or frequency of predictive decisions. A business can analyze data without deploying a production forecasting system, and a broad technology measure can include tools unrelated to prediction.
Transaction and customer data
Transaction records are among the most established inputs for predictive analysis because they connect business activity with time, products, payments, and outcomes. In 2025, 26% of EU enterprises performed internal data analytics on transaction records such as sales or payment records. Customer data—purchases, locations, preferences, reviews, or searches—was analyzed internally by 18% of EU enterprises (Eurostat, Enterprises performing data analytics internally, by data source, 2025).
The country results show how unevenly these data sources are used. Internal analytics of transaction records reached 41% in Denmark, 40% in Czechia, 38% in Estonia, and 37% in Belgium. For customer data, Estonia and Denmark each recorded 31%, Belgium recorded 28%, and Sweden recorded 27%.
| Internal data source or activity | EU enterprises, 2025 | Selected higher country results |
|---|---|---|
| Transaction records | 26% | Denmark 41%; Czechia 40%; Estonia 38%; Belgium 37% |
| Customer data | 18% | Estonia 31%; Denmark 31%; Belgium 28%; Sweden 27% |
| Social-media data | 11% | Belgium 20%; Lithuania 20%; Sweden 18% |
| Web data | 11% | Belgium 20%; Lithuania 17%; Sweden 16% |
| Location data | 7% | Slovakia 16%; Lithuania 16% |
| Smart-device or sensor data | 6% | Lithuania 12%; Belgium 10%; Slovakia 10% |
The transaction and customer figures are measures of internal analytics, not the percentage of enterprises that use those records in a predictive model. They nevertheless identify the most common data foundations in this comparison and provide a practical baseline for assessing how far an organization has moved toward data-informed decision support.
Digital traces and online behavior
Digital trace data can add signals about attention, demand, sentiment, and changing behavior. In 2025, 11% of EU enterprises performed internal analytics on social-media data, and the same proportion analyzed web data such as search trends or web-scraping data (Eurostat, Enterprises performing data analytics internally, by data source, 2025).
Some countries reported substantially higher use. Social-media analytics reached 20% of enterprises in Belgium and 20% in Lithuania, while Sweden recorded 18%. For web data, Belgium recorded 20%, Lithuania 17%, and Sweden 16%.
These sources can support predictions about demand or customer response, but the statistics do not identify the model type, forecasting horizon, or business result. They also do not establish that social-media or web analytics caused better decisions. The figures show the prevalence of internal analysis using these data categories during 2025.
Location, sensors, and operations
Location and sensor data can describe where activity occurs and how equipment or connected devices behave. In 2025, 7% of EU enterprises performed internal analytics on location data from portable devices or vehicles. Six percent analyzed data from smart devices or sensors (Eurostat, Enterprises performing data analytics internally, by data source, 2025).
The country-level results were higher in selected markets. Location-data analytics reached 16% in Slovakia and 16% in Lithuania. Smart-device or sensor analytics reached 12% in Lithuania, 10% in Belgium, and 10% in Slovakia.
The gap between the EU figures for transaction records and these operational sources is notable: 26% analyzed transaction records, compared with 7% for location data and 6% for smart-device or sensor data. This describes adoption of internal analytics, not necessarily the availability of data. A business may possess sensor or location data without having the systems, skills, permissions, or use case needed to analyze it.
Open and satellite data
External public data can extend a predictive system beyond the organization’s own records. In 2025, 7% of EU enterprises performed internal analytics on government authorities’ open data, while 4% analyzed satellite data (Eurostat, Enterprises performing data analytics internally, by data source, 2025).
Government open-data analytics reached 19% of enterprises in Latvia, 14% in Czechia, and 12% in Lithuania. Satellite-data analytics reached 12% in Lithuania and 6% in Slovakia.
The source mix matters when interpreting predictive-analytics statistics. Government open data and satellite data may be useful for forecasting conditions outside a company’s direct control, but Eurostat’s figures do not specify the applications, data frequency, spatial resolution, or prediction accuracy. They should therefore be treated as adoption measures rather than evidence that one external source is more effective than another.
Business size and geography
Enterprise size is strongly associated with internal analytics adoption. In 2025, 78.84% of large EU enterprises performed data analytics with their own employees, compared with 27.86% of small EU enterprises (Eurostat, Digital economy and society statistics - enterprises). The difference is 51 percentage points, showing a much wider reach of in-house analytics among large organizations.
When internal and external providers are counted together, the leading country figures in the supplied Eurostat comparison were 59.99% in Denmark, 55.99% in Estonia, and 55.96% in the Netherlands. Poland recorded 24.50%.
| Measure, 2025 | Percentage |
|---|---|
| Large EU enterprises using their own employees for analytics | 78.84% |
| Small EU enterprises using their own employees for analytics | 27.86% |
| Danish enterprises using internal or external analytics | 59.99% |
| Estonian enterprises using internal or external analytics | 55.99% |
| Dutch enterprises using internal or external analytics | 55.96% |
| Polish enterprises using internal or external analytics | 24.50% |
The size comparison concerns internal analytics, while the country comparison combines internal and external activity. Those denominators are different, so the figures should not be ranked as though they were a single measure. The results do support a clear decision-making point: access to external providers can be part of the analytics pathway for organizations that do not maintain a large internal team.
AI context and measured business outcomes
AI statistics provide context for predictive analytics but are broader than predictive analytics itself. In 2024, 13.5% of EU enterprises with at least 10 employees used AI technologies, up from 8.0% in 2023. AI use reached 27.6% in Denmark, 25.1% in Sweden, and 24.7% in Belgium. Machine learning for data analysis was one of the AI technologies measured by Eurostat within that 13.5% enterprise AI-adoption population (Eurostat, Usage of AI technologies increasing in EU enterprises).
U.S. evidence from the Business Trends and Outlook Survey showed business AI use rising from 3.7% in September 2023 to 5.4% in February 2024. The same source estimated a rate of about 6.6% by early fall 2024; that 6.6% figure is a forecast, not an observed result (U.S. Census Bureau, Tracking Firm Use of AI in Real Time).
For the United Kingdom, businesses with at least 10 employees reporting AI use increased from around 12% in late 2023 to around 35% in June 2026. The average number of AI technologies used per adopting UK business rose from around 1.4 to around 1.6 over the same period (Office for National Statistics, Artificial intelligence in UK businesses: 2023 to 2026).
A separate UK analysis reported around 19% higher turnover per worker associated with businesses adopting advanced digital technologies such as AI and cloud computing, after accounting for management practices and business characteristics. This is a descriptive association and does not establish causality. The underlying Business Data Use and Productivity Study wave 2 collected information from 3,796 businesses between 3 December 2024 and 28 February 2025 (UK Department for Business and Trade, Business data use and productivity study (wave 2): statistical report). The survey count is population context, not a predictive-analytics performance statistic.