A quantitative analysis of what characterizes BDA-using companies

This report is part of a framework project on big data analytics (BDA) adoption and use within Swedish firms. The aim of the report is to provide a knowledge base for the Swedish decision-makers to better understand the drivers of BDA adoption and use at the firm-level in Sweden. This, in turn, we hope can lead to informed policymaking for the achievement of Sweden’s ambitious digitalisation objectives.

The report maps the use of BDA by Swedish firms, their characteristics and how they compare to other firms within the European Union (EU) context. The main data source for the analysis are the annual IT-surveys conducted by the respective statistical agencies of EU member states under the supervision of the statistical office of the EU (Eurostat). In this report, we make use of two sets of data: firm-level data from Statistics Sweden for the Swedish case and aggregated data from Eurostat for the purposes of cross-country comparisons.

Main findings

Almost one in five Swedish firms use BDA

The report shows that almost one in five Swedish companies used BDA in 2020, ranking Sweden ninth highest in Europe. Although this places Sweden above the EU average, it positions is slightly behind leading comparable countries such as Denmark, the Netherlands and the United Kingdom.

Sharp increase since 2016

The share of Swedish firms' using BDA doubled between 2016 and 2020, with strong growth across all company sizes, sectors and regions. This result is in line with similar trends observed for the leading countries.

BDA use is more prevalent in larger firms

The use of BDA is concentrated in large firms, with 45 percent of these firms reporting using BDA in 2020, compared with 29 percent of medium-sized firms and 17 percent of small firms. This structural pattern of BDA use is similar across-countries within the EU and persists regardless of the sector and region in which firms operate.

Technology-intensity and location key drivers of BDA use

Branch of activity and location within the country are found to be important factors that determines the likelihood for BDA use. Within the Swedish context, firms in technologyintensive industries, located in the Stockholm region are found to be most likely to use BDA than others. BDA use likely to increase rapidly with AI Our analysis indicates that companies using BDA are also widely using AI. The two can be seen as complementary and mutually reinforcing. Previous analyses, both internationally and within the Swedish context, has shown that AI has a huge potential to contribute to economic growth (Tillväxtanalys, 2023b). Thus, a similar conclusion should therefore partly apply to BDA use, given its complementarity of AI development and use. In general, the results in this report are similar to findings in a previous study on AI use and productivity among Swedish firms (Tillväxtanalys, 2023a).

However, the findings of the report should be interpreted with caution. Although the data is from 2020, it is already at risk of being dated to some extent due to rapid technological developments. The arrival of Chat GPT in 2022, the first commercial generative AI model, made AI available to the public and contributed to an explosion in its use. Given that these models require huge amounts of data, it is reasonable to assume that BDA use by firms has also increased since 2020.

Policy implications

Support schemes for increased use of BDA by small firms may be warranted

Only 17 percent of small firms used BDA in Sweden in 2020, compared with 45 percent for larger firms. This result is not surprising as smaller firms often struggle with adoption and use of new technology due to a number of factors such as limited investment capacity, higher financial vulnerability, and access to a skilled workforce. To the extent that policymakers want to see increased use of BDA, small businesses represent a group with significant potential, especially with respect to untapped economic growth.

Need for better analysis and evaluation of digitalisation policy

We see that it is possible to improve the conditions for impact evaluations of BDA use, in line with our previous recommendations with respect to AI use (Tillväxtanalys, 2023a; 2023b). This can be done by expanding the range and nature of questions in the statistical surveys being collected. Specifically, more in-depth questions about the challenges faced by companies when investing in and implementing BDA, as well as their access to and need for specific BDA skills, would also be valuable