Reference metadata describe statistical concepts and methodologies used for the collection and generation of data. They provide information on data quality and, since they are strongly content-oriented, assist users in interpreting the data. Reference metadata, unlike structural metadata, can be decoupled from the data.
Data on the Information and Communication Technologies (ICT) usage and e-commerce in enterprises are survey data. They are collected by the National Statistical Institutes or Ministries and are in principle based on Eurostat's annual model questionnaires on ICT usage and e-commerce in enterprises.
The legal basis for ICT enterprise statistics for survey year 2024 is Commission Implementing Regulation (EU) 2023/1507 of 20 July 2023 laying down the technical specifications of data requirements for the topic 'ICT usage and e-commerce' for the reference year 2024.Large part of the data collected is used to measure the progress in the implementation of one of the main political priorities of the European Commission for 2019 to 2024 – A Europe fit for the digital age.
Furthermore, ICT data facilitates the monitoring of the EU’s digital targets for 2030, set by the Digital Decade Policy Programme. Four of the key performance indicators (KPIs) of the current programme stem from the statistics for which the implementing and delegated acts are enclosed for adoption: Artificial Intelligence, cloud, big data (data analytics) and the digital intensity index for businesses (DII) - a composite indicator reflecting the digital transformation of business. The aim of the European survey on ICT usage and e-commerce in enterprises is to collect and disseminate harmonised and comparable information at European level.
All economic activities in the scope of Annex of the Commission Regulation are intended to be included in the general survey, covering enterprises with 10 or more employees and self-employed persons. These activities are:
Section C – “Manufacturing”
Section D, E – “Electricity, gas, steam and air conditioning supply”, “Water supply, sewerage, waste management and remediation activities”
Section F – “Construction”
Section G – “Wholesale and retail trade; repair of motor vehicles and motorcycles”
Section H – “Transportation and storage”
Section I – “Accommodation and food service activities”
Section J – “Information and communication”
Section L – “Real estate activities”
Section M – “Professional, scientific and technical activities”
Section N – "Administrative and support service activities"
Group 95.1 – “Repair of computers and communication equipment”.
For micro-enterprises see the sub-concepts below.
3.3.1. Coverage-sector economic activity for micro-enterprises - All NACE Rev. 2 categories are covered
Yes
3.3.2. Coverage sector economic activity for micro-enterprises - If not all activities were covered, which ones were covered?
All economic activities in the scope of Annex I of the Commission Regulation are intended to be included in the general survey, covering enterprises with 2 up to 9 employees. These activities are: NACE Rev. 2 sections C, D, E, F, G, H, I, J, L, M and N, division 95.1.
3.4. Statistical concepts and definitions
The model questionnaire on ICT usage and e-commerce in enterprises provides a large variety of variables covering among others the following topics:
- Access to and use of the Internet
- E-commerce and e-business
- ICT specialists and skills
- ICT security
- Artificial Intelligence
- Use of cloud computing services
The annual model questionnaires and the European businesses statistics compliers’ manual for ICT usage and e-commerce in enterprises comprise definitions and explanations regarding the topics of the survey.
3.5. Statistical unit
The statistical unit used is the business enterprise.
3.6. Statistical population
Target Population
As required by Annex of the Commission Implementing Regulation, enterprises with 10 or more employees and self-employed persons are covered by the survey.
For micro-enterprises see the sub-concepts below.
3.6.1. Coverage of micro-enterprises
Yes
3.6.2. Breakdown between size classes [0 to 1] and [2 to 9]
No
3.6.3. If for micro-enterprises different size delimitation was used, please indicate it.
[2-9]
3.7. Reference area
Detailed information on the provision of data on NUTS 2 regional level is available in “Annex I. Completeness“.
The geographic scope of the ICT-survey is the (European) Netherlands, i.e. excluding Caribbean Netherlands (which are also part of the Kingdom of the Netherlands).
3.8. Coverage - Time
Years 2023 and 2024.
3.9. Base period
Not applicable
Percentages of enterprises, percentages of turnover, percentages of employees and million euro (for selected indicators).
For the 2024 ENT survey, the reference period is the current survey period (2024) except for the modules on e-commerce, ICT specialists and skills and for (some questions of) ICT security where respondent should report 2023 values.
6.1. Institutional Mandate - legal acts and other agreements
Complementary national legislation constituting the legal basis for the survey on the use of ICT in enterprises:
Statistics Netherlands (CBS) carries out its statistical research using the encrypted version of the data. It goes without saying that CBS also adheres to all the relevant legislative requirements in order to safeguard data privacy. These include:
The General Data Protection Regulation (Algemene Verordening Gegevensbescherming, AVG 679/2016)
The Statistics Netherlands Act (CBS-wet effective from 2 March 2022 on Act of 20 November 2003 enacting a law governing Statistics Netherlands on acquisition, use and provision of data in the context of the supply of statistical information.)
This latter is the cornerstone of the Statistical Business Register (SBR) of the CBS.
7.1. Confidentiality - policy
Regulation (EC) No 223/2009 on European statistics (recital 24 and Article 20(4)) of 11 March 2009 (OJ L 87, p. 164), stipulates the need to establish common principles and guidelines ensuring the confidentiality of data used for the production of European statistics and the access to those confidential data with due account for technical developments and the requirements of users in a democratic society.
At national level:
At National level, the Dutch Data Protection Authority (DPA) supervises compliance with regulations of the law concerning personal data protection. An important contact person for the Dutch DPA is the CBS Data Protection Officer. Statistics Netherlands has its own Data Protection Officer (DPO). This officer monitors the implementation of and compliance with the GDPR by CBS and keeps a register of all personal data processing.
The measures of Statistics Netherlands (CBS) to safeguard the privacy and confidentiality of the collected data include technical and logistical measures:
All CBS employees are bound by an obligation of confidentiality and have signed a confidentiality agreement.
CBS only uses data for statistical and scientific purposes. CBS is excluded by law from using the data for fiscal, administrative, verification and legal purposes. Furthermore, CBS is not allowed to use data for marketing.
All statistical processes at CBS are certified in terms of personal data protection. This privacy proof audit is carried out by an accredited external party.
The data protection policy concerning a minimum frequency rule has been updated. Recently, it is adviced to use 10 units for all NSI publications.
7.2. Confidentiality - data treatment
Data are transmitted via eDamis (encrypted) and delivered to a secure environment where they are treated. Flags are added for confidentiality in case results must not be disclosed.
At national level:
More specific measures of Statistics Netherlands (CBS) to safeguard the privacy and confidentiality of the collected data are:
When respondents complete a survey or submit any data, this information is delivered to CBS in encoded form. The data are received by CBS in a secured environment. Only authorised personnel shall have access to these data.
At the earliest possible stage in the process, all directly identifiable personal data are removed from the files. This means datasets for research will never contain any data such as names, addresses or citizen service numbers.
Additional information about confidentiality policies and measures are available via Statistics Netherlands webpage and brochures.
8.1. Release calendar
The target is to release the obtained statistical data within the survey year. Announcements on the release of the outputs from the survey on ICT usage in enterprises are published in a publicly accesible manner, namely the CBS-Publication calendar. The statistical outputs (StatLine tables) are free and available via the thema page ICT usage in enterprises (in Dutch).
Five tables about the results of the ICT usage in enterprises will be published in December 2024. By that time an announcement will be placed in the CBS-Publication calendar and short after a map will be made available via ICT usage in enterprises (in Dutch).
In general, annual news releases are available online in December of the survey year or in January of the following year.
10.2. Dissemination format - Publications
The National dissemination of results is carried out twice a year:
- At the beginning of the data collection. Enterprises are provided the links to the statistical outcomes of the previous year along with the invitation to participate in the survey "ICT usage in enterprises" (in Dutch).
- At the publication and release of the outcomes of the current year via Statline and MKB Statline. The latter is an open data site dedicated to statistics on Small and Medium-sized Enterprises (SMEs).
Furthermore, the yearly dissemination of results includes also:
- The 2023 publication "ICT, knowledge and economy" (in Dutch) is available as a web publication and it is free of charge. (Notice that this sort releases are based on data collected in the survey of a year earlier.) A link to all yearly publications is found Thema page. Since 2024, this publication has evolved into a dashboard containing a selection of the most relevant figures. These are presented in interactive manner and in time series. The release of the "Dashboard Digitalization and Knowledge Economy" (in Dutch) is publically available since last February 2025.
- The publication "Cybersecuritymonitor 2024" (in Dutch) is available as a PDF file and as web page. Both are also free of charge. (Notice that the release 2024 is based on data collected in the 2024 survey.) The publication series on Cybersecurity is also free of charge and publically available.
10.3. Dissemination format - online database
See detailed section 10.3.1.
10.3.1. Data tables - consultations
Results for selected variables collected in the framework of this survey are available for all participating countries on Digital economy and society of Eurostat website.
At national level:
Outcome tables are available via de websites Statline and MKB Statline. This later is a dedicated website for SME's. The number of public accesses is measured and reported per quarter and per table.
10.4. Dissemination format - microdata access
Microdata requests are granted under a number of conditions. See Microdata.
10.5. Dissemination format - other
Not requested
10.5.1. Metadata - consultations
Not requested
10.6. Documentation on methodology
The European businesses statistics compilers’ manual for ICT usage and e-commerce in enterprises provides guidelines and clarifications for the implementation of the surveys.
Documentation on quality management is recently updated as a part of the Year Plan of Statistics Netherlands (2023 and 2025, in Dutch). See section 4.2 (Kwaliteitsmanagement, in Dutch).
The aims is to develop a central management of the quality management system. Since 2022 a Chief Quality Officer advises on organization-wide policy frameworks in the field of quality assurance. This task is carried out in cooperation with appointed quality and process coordinators in the various departments and with the internal audit departments.
Main development is connecting the management systems for quality, information security and privacy and setting up an Improvement Register, whereby outstanding points from audits in the areas of quality, information security and privacy are monitored in a structured manner and demonstrably followed up.
11.1. Quality assurance
The European businesses statistics compliers’ manual for ICT usage and e-commerce in enterprises provides guidelines and standards for the implementation of the surveys. It is updated every year according to the changed contents of the model questionnaires.
At national level:
The general quality management principles in Statistics Netherlands are actively promoted by means of the Statistical Netherlands’ Quality Assurance Framework at Process Level (Quality Guidelines for Process Assurance 2014, 2017 and 2019). These guidelines are continuously revised and operationalized in a template to support National Information Protection. The template includes general information of the statistical (production) process as well as workflows and descriptions at different detail levels (top-down). Moreover, documentation on connected information systems and software are also provided along with self-assessments and GDPR-documentation. Monitoring of compliance and audits are carried out by the department of Methodology of Statistics Netherlands. The updates to these documentation about the process ICT-usage in enterprises are executed yearly.
Moreover, in July 2022, CBS was the subject of a peer review under supervision of the European Statistical System (ESS)
11.2. Quality management - assessment
At European level, the recommended use of the annual Eurostat model questionnaire aims at improving comparability of the results among the countries that conduct the survey on ICT usage and e-commerce in enterprises. Moreover, the European businesses statistics compilers’ manual for ICT usage and e-commerce in enterprises provides guidelines and clarifications for the implementation of the surveys.
At national level:
Production process is designed to be fully automatized from reading the survey data to the ready-to-upload data delivery. Evaluations and consistency tests are rule-driven. Process includes also rest-points to check data and metadata quality. Higher-level script is updated yearly based on Model Questionnaire such that handlings are minimal and reproducibility of results is increased. Lower-level scripts (and packages) are brought under a version control and management system. Updates of these scripts follow technical requirements of delivery.
12.1. Relevance - User Needs
European level :
At European level, European Commission users (e.g. DG CNECT, DG GROW, DG JUST, DG REGIO, DG JRC) are the principal users of the data on ICT usage and e-commerce in enterprises and contribute in identifying/defining the topics to be covered. Hence, main users are consulted regularly (at hearings, task forces, ad hoc meetings) for their needs and are involved in the process of the development of the model questionnaires at a very early stage.
User needs are considered throughout the whole discussion process of the model questionnaires aiming at providing relevant statistical data for monitoring and benchmarking of European policies.
At European level, contacts within the Commission, the OECD and other stakeholders give a clear picture about the key users' satisfaction as to the following data quality aspects: accuracy and reliability of results, timeliness, satisfactory accessibility, clarity and comparability over time and between countries, completeness and relevance. Overall users have evaluated positively (good, very good) the data quality on the ICT usage and e-commerce in enterprises.
National level :
The current level of disaggregation of enterprises' size classes reaches the requirements of the main users of the data of the ICT usage and e-commerce in enterprises.
As for the user satisfaction, this is not available.
12.3. Completeness
Detailed information is available in “ Annex I. Completeness “ - related to questionnaire, coverage, additional questions, regional data.
12.3.1. Data completeness - rate
Not requested
13.1. Accuracy - overall
Comments on reliability and representativeness of results and completeness of dataset
These comments reflect overall standard errors reported for the indicators and breakdowns in section 13.2.1 (Sampling error - indicators) and the rest of the breakdowns for national and European aggregates, as well as other accuracy measurements. The estimated standard error should not exceed 2pp for the overall proportions and should not exceed 5pp for the proportions related to the different subgroups of the population (for those NACE aggregates for the calculation and dissemination of national aggregates). If problems were found, these could have implications for future surveys (e.g. need to improve sampling design, to increase sample sizes, to increase the response rates).
More detailed information is available in “ Annex II. Accuracy “ - related to European aggregates, comments on reliability and use of flag.
13.2. Sampling error
For calculation of the standard error see 13.2.1.1.
13.2.1. Sampling error - indicators
Standard error (for selected indicators and breakdowns)
Precision measures related to variability due to sampling, unit non-response (the size of the subset of respondents is smaller than the size of the original sample) and other (imputation for item non-response, calibration etc.) are not (yet) required from the Member states for all indicators. Eurostat will make basic assumptions to compute these measures for all indicators produced (e.g. stratified random sampling assuming as strata the crossing of the variables “Number of employees and self-employed persons” and “Economic Activity” as it was defined in the 3 tables of section 18.1).
More detailed information is available in“ Annex III. Sample and standard error tables 2024 “ – worksheets starting with “Standard error".
13.2.1.1. Sampling error indicator calculation
Calculation of the standard error
Various methods can be used for the calculation of the standard error for an estimated proportion. The aim is to incorporate into the standard error the sampling variability but also variability due to unit non-response, item non-response (imputation), calibration etc. In case of census / take-all strata, the aim is to calculate the standard errors comprising the variability due to unit non-response and item non-response.
a) Name and brief description of the applied estimation approach
CBS has been applying the theory of Horvitz & Thompson for many years to draw samples of enterprises from the Statistical Business Register (SBR). Horvitz & Thompson proved that it is always possible to calculate a good estimate if you draw the sample properly (random) and all the draw probabilities are known. The selection of the sample must be done using a chance mechanism to ensure that estimates are not systematically too high or too low (i.e. unbiased), and only then uncertainty margins can be calculated. Such a margin indicates how large the deviation between estimate and actual value can be.
The sampling error in the Dutch survey ICT usage in enterprises is hence calculated using the Horvitz and Thomson formula for the weighted sample mean. In an ideal situation, very high response rates (95% or higher), the estimates are very reliable. However, these estimators are less realible and actually not robust against non-response. Non-response may arise when part of the questionnaires cannot be completed, e.g. because respondents are not reachable, enterprises refuse to participate, or they are time-constrained and unable to participate.
Non-response affects the representativeness of the sample. This is because non-response often occurs among specific groups. For example, non-response can be very high in strata (sample cells) with small domain populations like Electricity, Gas and Water Supply (E35TE39) or Repair of computers (S951). Non-response can also affect specific questions or modules.
Whether or not observations become available is determined by two processes: sampling and non-response mechanisms. The probability of an observation is therefore equal to the probability of drawing the sample multiplied by the probability of response. However, the response rates are unknown. Thus, the theory of Horvitz & Thompson can no longer be applied and unbiased estimators cannot be calculated.
CBS tries to reduce estimates distortions by applying a weighing technique. Weights are assigned to the respondents to restore representativeness. An attempt is also be made to estimate the responses by using random (item) imputation based on response strata distributions, after which the theory of Horvitz & Thompson can be applied again.
By avoiding lower response rates, the simpler it becomes to solve bias problems. Nonetheless, there is no guarantee that these correction techniques can completely remove estimates distortions.
b) Basic formula
The weighted mean of stratum (sample cell) s of a variable p is calculated as
where wi,s are the weighting factors for cell i based on either enterprises, number of employees, or turnovers in stratum s with sample size ns.
The weighted variance of stratum j is calculated as
The standard deviation is estimated using the square root hereof
The standard error (S.E.) of a finite population is estimated by dividing the result by the square root of the number of records in the sample corrected by multiplying with a finite population correction (FPC):
where:
σw (ps):is the weighted standard deviation for stratum s,
ns: is the sample size of stratum s,
Ns: is the population size of stratum s,
1 - ns / Ns: is the simplified finite population correction (FPC) representing the relation between sample- and population sizes.
For large ns / Ns is the sampling fraction per stratum (fs):
CBS gains precision by sampling close to a larger percentage of the population in the case of enterprises with more than 100 employees. The effect on the FPC is that the error becomes zero when the sample size ns is equal to the population size Ns.
d)How has the stratification been taken into account?
Sample cells are defined as the product of sizes classes (14 size classes with 1+ employed persons) and NACE classes (using a more detailed stratification than the publication classes demanded by Eurostat, i.e. 63 NACE subgroups).
Enterprises with 100 or more employees (7 size classes) are included in its entirety in the gross sample. For enterprises with 2-99 employees (6 size classes) a random sample is used. Sample size is based on prior experiences and cost considerations. For optimal distribution of sample units across the strata, the Neymann allocation method is used.
Enterprises with 10-99 employees (3 size classes) which are very often drawn in business survey samples have a reduced likelihood of being drawn into the sample of the ICT-survey. This is done in particular to reduce the statistical burden on these enterprises. When possible, these enterprises are replaced by similar units in the same stratum.
e)Which strata have been considered?
All breakdowns for which the data were delivered to Eurostat, but only for the variables that were asked in this section of the QR.
13.3. Non-sampling error
See detailed sections below.
13.3.1. Coverage error
See concept 18.1.1. A) Description of frame population.
13.3.1.1. Over-coverage - rate
Over-coverage concerns two concepts 18.1. Description of frame population of the ICT- usage enquiry and 13.3.3.1.1. Enterprises out of scope (e.g. deaths, misclassified originally in the target population, etc.). See values provided in table 13.3.3.1.1.
13.3.1.2. Common units - proportion
Not requested
13.3.2. Measurement error
Restricted from publication
13.3.3. Non response error
See detailed sections below.
13.3.3.1. Unit non-response - rate
See detailed sub-concepts below.
13.3.3.1.1. Unit response
Restricted from publication
13.3.3.1.2. Methods used for minimizing unit non-response
After the invitation letters in February, we have 3 reminders in total until September in order to motivate enterprises to fill in the ICT-survey.
13.3.3.1.3. Methods used for unit non-response treatment
1. No treatment for unit non-response
2. Treatment by re-weighting
2.1 Re-weighting by the sampling design strata considering that non-response is ignorable inside each stratum (the naïve model)
2.2 Re-weighting by identified response homogeneity groups (created using sample-level information)
x
2.3 Re-weighting through calibration/post-stratification (performed using population information) by the groups used for calibration/post-stratification
3. Treatment by imputation (done distinctly for each variable/item)
x
4. Method(s) and the model(s) corresponding to the above or other method(s) used for the treatment of unit non-response. (e.g. Re-weighting using Horvitz-Thompson estimator, ratio estimator or regression estimator, auxiliary variables)
13.3.3.1.4. Assessment of unit non-response bias
The response rate was higher than 60% for EGE10, but it was 42% for E2T9 (micro-enterprises).
13.3.3.2. Item non-response - rate
Not available.
13.3.3.2.1. Methods used for item non-response treatment
1. No treatment for item non-response
2. Deductive imputation An exact value can be derived as a known function of other characteristics.
x
3. Deterministic imputation (e.g. mean/median, mean/median by class, ratio-based, regression-based, single donor nearest-neighbour) Deterministic imputation leads to estimators with no random component, that is, if the imputation were to be re-conducted, the outcome would be the same.
x
4. Random imputation (e.g. hot-deck, cold-deck) Random imputation leads to estimators with a random component, that is, if the imputation were re-conducted, it would have led to a different result.
5. Re-weighting
6. Multiple imputation In multiple imputation each missing value is replaced (instead of a single value) with a set of plausible values that represent the uncertainty of the right value to impute. Multiple imputation methods offer the possibility of deriving variance estimators by taking imputation into account. The incorporation of imputation into the variance can be easily derived based on variability of estimates among the multiply imputed data sets.
7. Method(s) and the model(s) corresponding to the above or other method(s) used for the treatment of item non-response.
13.3.3.2.2. Questions or items with item response rates below 90% and other comments
Other comments relating to the item non-response
Additional issues concerning "item non-response" calculation (e.g. method used in national publications).
Questions and items with low response rates (cut-off value is 90%) and item non-response rate.
Not applicable.
13.3.4. Processing error
We did not have processing errors.
13.3.5. Model assumption error
Not requested
14.1. Timeliness
See detailed section in the Full metadata report.
14.1.1. Time lag - first result
Not applicable
14.1.2. Time lag - final result
Data are to be delivered to Eurostat in the fourth quarter of the reference year (due date for the finalised dataset is 5th October). European results are released before the end of the survey year or in the beginning of the year following the survey year (T=reference year, T+0 for indicators referring to the current year, T+12 months for other indicators referring to the previous year e.g. e-commerce).
At national level:
Data are published nationally at StatLine in December 2024, the time lag will be T+1 for indicators referring to 2024 (most of variables), T+12 months for indicators referring to 2023 (e-commerce and ICT-security).
14.2. Punctuality
See detailed section below.
14.2.1. Punctuality - delivery and publication
Restricted from publication
15.1. Comparability - geographical
The model questionnaire is generally used by the countries that conduct the survey on ICT usage and e-commerce in enterprises. Due to (small) differences in translation, in the used survey vehicle, in non-response treatment or different routing through the questionnaire, some results for some countries may be of reduced comparability. In these cases, notes are added in the data.
Detailed information on differences in the wording of the questions in the national questionnaires is available in “ Annex I. Completeness “ - worksheets related to questionnaire, coverage, additional questions.
Comparability between regions:
In the year 2024, data on regions has been delivered for the second time to Eurostat. One of the main problems of comparability between regions of the country is that some economical sectors are concentrated in a few provinces. Therefore data can become strongly geographically unbalanced, e.g. the services sector is concentrated in NL3: West Netherlands when compared with NL1: North Netherlands.
The figures of the variables number of employees (EMPL) and number of enterprises (ENT) for the combination C10TS951XK (All enterprises) and EGE10 (10 or more employees) show strong business demographics difference of 10:1 and 7:1 between NL3 and NL1 regions, respectively.
Detailed information on the provision of data on NUTS 2 regional level is available in “Annex I. Completeness“ – worksheets related to regional data.
15.1.1. Asymmetry for mirror flow statistics - coefficient
Not applicable
15.2. Comparability - over time
See detailed section in the Full metadata report.
15.2.1. Length of comparable time series
The length of comparable time series depends on the module and the variable considered within each survey module. Additional information is available in annexes attached to the European metadata.
There are no changes with regard to last year, hence we do not expect differences inherent to this issue.
15.3. Coherence - cross domain
Not applicable
15.3.1. Coherence - sub annual and annual statistics
Not applicable
15.3.2. Coherence - National Accounts
Not applicable
15.4. Coherence - internal
Not applicable
Restricted from publication
17.1. Data revision - policy
Data have a provisional status from the first time they are published. Target is achieving a definitive status within two periods after.
In the period 2023 an unplanned revision took place. This revision originated from a technical fault and lead to methodological research and evaluation. Methodological improvements have been provided on the short- and long-term. The short-term improvements (e.g. improved imputation procedures) have been implemented. The long-term improvements (e.g. evaluation of consistency across variables due to the variable-specific imputation of item non-response) requires further planning, unit tests and training sets.
17.2. Data revision - practice
In case we need to correct a (set) values, we republish the tables on Statline along with an explanation on what has been changed. Eventually, it can be decided to temporally suppress a table until the corrections and corresponding check-ups are executed.
The main reasons causing data revision are linked to technical errors on the data collection or on the processing of the data. In those cases, methodological advice may also be requested.
17.2.1. Data revision - average size
Not requested
18.1. Source data
A) Frame population description and distribution
For more information see concept 18.1.1.
B) Sampling design - Sampling method
Description of the sampling method used (e.g. stratified random sample, quota sampling, cluster sampling; one-stage or two-stage sampling) and information which variables were used to stratify, the categories of those variables, in particular for the NACE Rev. 2 categories related to the "possible calculation of European aggregates", and the final number of strata:
Enterprises with 100 or more employees are all included in the gross sample. For enterprises with 2-100 employees a random sample is used, stratified by NACE (32) and size class (6). Moreover, sample size is also based on prior experiences and cost considerations. For the optimal distribution of sample units across the strata, the Neymann allocation method is used. Enterprises that are drawn in other business survey samples have a reduced likelihood of being drawn into the sample of this survey.
C) Gross sample distribution
More detailed information is available in “ Annex III. Sample and standard error tables 2024 “ (Worksheet: GROSS SAMPLE)
D) Net sample distribution
More detailed information is available in “ Annex III. Sample and standard error tables 2024 “ (Worksheet: NET SAMPLE)
18.1.1. Sampling design & Procedure frame
A) Description of frame population
a) When was the sample for the ICT usage and e-commerce in enterprise survey drawn?
07-02-2024
b) Last update of the Business register that was used for drawing the sample of enterprises for the survey:
02-02-2024
c) Indication if the frame population is the same as, or is in some way coordinated with, the one used for the Structural Business Statistics (different snapshots):
Not actively coordinated but both have as basis the SBR-stand 01-12-2023
d) Description if different frames are used during different stages of the statistical process (e.g. frame used for sampling vs. frame used for grossing up):
n.a.
e) Indication the shortcomings in terms of timeliness (e.g. time lag between last update of the sampling frame and the moment of the actual sampling), geographical coverage, coverage of different subpopulations, data available etc., and any measures taken to correct it, for this survey.
The Statistical Business Register (SBR) is daily updated. The sample was drawn at 07-02-2024 based on the population snap-shot of 01-12-2023. Because of the time lag between snap-shot date of sampling frame (01-12-2023) and the start of data collection (23-02-2024) some enterprises were found not active anymore during the data collection. Enterprises of the sample which were not active anymore at 23 February were excluded from the data collection. No measures were taken to correct for this except that the weights of the sample data change.
B) Frame population distribution
More detailed information is available in “ Annex III. Sample and standard error tables 2024 “ (Worksheet: FRAME POPULATION)
18.2. Frequency of data collection
Annual
18.3. Data collection
See detailed sections below.
18.3.1. Survey period
Survey / Collection
Date of sending out questionnaires
Date of reception of the last questionnaire treated
General survey
23-02-2024
01-10-2024
Micro-enterprises
23-02-2024
01-10-2024
18.3.2. Survey vehicle – general survey
General survey - Stand-alone survey
18.3.3. Survey vehicle – micro-enterprises
The collection of micro-enterprises was integrated with the general survey
18.3.4. Survey type
This is a fully web survey, meaning that responses are collected online only. The sample of respondents is randomly divided in two portions. The first portion is addressed around half February and the second portion is reached at the end of February.
18.3.5. Survey participation
Voluntary
18.4. Data validation
We have used the server-based validation tool provided by Eurostat (both Acceptance and Production environments) to validate the data deliveries plus an in-house developed tool for a year-on-year comparison.
Internal validation is executed by adding the last results to the existing time series. Positive and negative trends are detected structurally. The monitoring of the results is based on step-wise filtering of the results using a top-down approach, i.e. the first results to be reviewed are the top NACE code aggregations and thereafter the higher enterprise size classes. In the next step, cross tables between size class versus medium and low NACE code aggregations are performed.
The log files of the processing tool are also reviewed. Those contain data about coverage and imputation rates.
Relating the output datasets of the statistic ICT usage in enterprises with other datasets should be explored yet, given that such a comparisons are cross-domain statistics. Those require often other identification codes and geolocation considerations to match those consistently.
18.5. Data compilation
Grossing-up procedures
Depending on the variable that has to be reported, background variables such as “number of persons employed”, "number of enterprises" and "turnover" are needed for grossing up the responses from industry level towards economic sector level. Most variables collected are qualitative ones. A few are quantitative variables, i.e. percentage or real values.
Three variable cases are considered:
Case 1. Variables such as (overall) percentage of persons employed having access remotely to IT-services of their enterprises
Case 2. Variables such as (overall) percentage of enterprises using social media or having a website or
Case 3. Variables such as (overall) percentage of turnover via online platforms.
We use weights to aggregate from industry-level towards economic sector-level. Subgroup weights are based on the number of enterprises that are within the strata (i.e. survey size classes by sector levels) of the target population and the corresponding number of enterprises in the response. These weighting factors (partially) correct biases caused by the unit non-response for target variables of Case 1 and Case 2. The e-commerce turnover-related variables are estimated using weighting factors based on enterprises' turnover (variables of Case 3) and are less robust to unit non-response.
18.5.1. Imputation - rate
This value varies from breakdown to breakdown and from item to item. These values are reported in the logging files and depend strongly on the filters provided in the specifications. We do not report these in a separate table.
18.6. Adjustment
Not applicable
18.6.1. Seasonal adjustment
Not applicable
Problems encountered and lessons to be learned:
In general, the tertiary sectors (G45TS951XK) seem to use AI-tools more extensively and intensively than primary or secondary sectors (C10TF43). The ICT-sector, retail trade(G47) and whole trade (G46) seem to be processing data of individuals using AI-tools and correcting those results for bias.
Moreover, questions on AI may be improved by giving an example on the use of tools (chatbox such as ChatGPT or any other) and the processing of the outcomes like writing letters, writing reports, or translating those. AI-tools are used to make scripts to interpret and compare the content of images, pictures or videos too.
Due to the speed but also complexity of the penetration and adoption of AI-tools, it may become relevant to ask enterprises if they are involved in pilots on the use of AI or if there is an AI-coordinator/responsible next to the ICT-responsible. Reason is that the latter may be more aware on the IT-infrastructure, but less on the specific use or adoption of for example Large Language models (LLM’s) within an enterprise.
Last but not least, wide spread automatization of processes increase efficiency. However, the use of an assisted control across the process leads to reduce the probability of technical failures.
Data on the Information and Communication Technologies (ICT) usage and e-commerce in enterprises are survey data. They are collected by the National Statistical Institutes or Ministries and are in principle based on Eurostat's annual model questionnaires on ICT usage and e-commerce in enterprises.
The legal basis for ICT enterprise statistics for survey year 2024 is Commission Implementing Regulation (EU) 2023/1507 of 20 July 2023 laying down the technical specifications of data requirements for the topic 'ICT usage and e-commerce' for the reference year 2024.Large part of the data collected is used to measure the progress in the implementation of one of the main political priorities of the European Commission for 2019 to 2024 – A Europe fit for the digital age.
Furthermore, ICT data facilitates the monitoring of the EU’s digital targets for 2030, set by the Digital Decade Policy Programme. Four of the key performance indicators (KPIs) of the current programme stem from the statistics for which the implementing and delegated acts are enclosed for adoption: Artificial Intelligence, cloud, big data (data analytics) and the digital intensity index for businesses (DII) - a composite indicator reflecting the digital transformation of business. The aim of the European survey on ICT usage and e-commerce in enterprises is to collect and disseminate harmonised and comparable information at European level.
Name of data collection
ICT-gebruik bij bedrijven 2024
25 June 2025
The model questionnaire on ICT usage and e-commerce in enterprises provides a large variety of variables covering among others the following topics:
- Access to and use of the Internet
- E-commerce and e-business
- ICT specialists and skills
- ICT security
- Artificial Intelligence
- Use of cloud computing services
The annual model questionnaires and the European businesses statistics compliers’ manual for ICT usage and e-commerce in enterprises comprise definitions and explanations regarding the topics of the survey.
The statistical unit used is the business enterprise.
Target Population
As required by Annex of the Commission Implementing Regulation, enterprises with 10 or more employees and self-employed persons are covered by the survey.
For micro-enterprises see the sub-concepts below.
Detailed information on the provision of data on NUTS 2 regional level is available in “Annex I. Completeness“.
The geographic scope of the ICT-survey is the (European) Netherlands, i.e. excluding Caribbean Netherlands (which are also part of the Kingdom of the Netherlands).
For the 2024 ENT survey, the reference period is the current survey period (2024) except for the modules on e-commerce, ICT specialists and skills and for (some questions of) ICT security where respondent should report 2023 values.
Comments on reliability and representativeness of results and completeness of dataset
These comments reflect overall standard errors reported for the indicators and breakdowns in section 13.2.1 (Sampling error - indicators) and the rest of the breakdowns for national and European aggregates, as well as other accuracy measurements. The estimated standard error should not exceed 2pp for the overall proportions and should not exceed 5pp for the proportions related to the different subgroups of the population (for those NACE aggregates for the calculation and dissemination of national aggregates). If problems were found, these could have implications for future surveys (e.g. need to improve sampling design, to increase sample sizes, to increase the response rates).
More detailed information is available in “ Annex II. Accuracy “ - related to European aggregates, comments on reliability and use of flag.
Percentages of enterprises, percentages of turnover, percentages of employees and million euro (for selected indicators).
Grossing-up procedures
Depending on the variable that has to be reported, background variables such as “number of persons employed”, "number of enterprises" and "turnover" are needed for grossing up the responses from industry level towards economic sector level. Most variables collected are qualitative ones. A few are quantitative variables, i.e. percentage or real values.
Three variable cases are considered:
Case 1. Variables such as (overall) percentage of persons employed having access remotely to IT-services of their enterprises
Case 2. Variables such as (overall) percentage of enterprises using social media or having a website or
Case 3. Variables such as (overall) percentage of turnover via online platforms.
We use weights to aggregate from industry-level towards economic sector-level. Subgroup weights are based on the number of enterprises that are within the strata (i.e. survey size classes by sector levels) of the target population and the corresponding number of enterprises in the response. These weighting factors (partially) correct biases caused by the unit non-response for target variables of Case 1 and Case 2. The e-commerce turnover-related variables are estimated using weighting factors based on enterprises' turnover (variables of Case 3) and are less robust to unit non-response.
A) Frame population description and distribution
For more information see concept 18.1.1.
B) Sampling design - Sampling method
Description of the sampling method used (e.g. stratified random sample, quota sampling, cluster sampling; one-stage or two-stage sampling) and information which variables were used to stratify, the categories of those variables, in particular for the NACE Rev. 2 categories related to the "possible calculation of European aggregates", and the final number of strata:
Enterprises with 100 or more employees are all included in the gross sample. For enterprises with 2-100 employees a random sample is used, stratified by NACE (32) and size class (6). Moreover, sample size is also based on prior experiences and cost considerations. For the optimal distribution of sample units across the strata, the Neymann allocation method is used. Enterprises that are drawn in other business survey samples have a reduced likelihood of being drawn into the sample of this survey.
C) Gross sample distribution
More detailed information is available in “ Annex III. Sample and standard error tables 2024 “ (Worksheet: GROSS SAMPLE)
D) Net sample distribution
More detailed information is available in “ Annex III. Sample and standard error tables 2024 “ (Worksheet: NET SAMPLE)
Annual
See detailed section in the Full metadata report.
The model questionnaire is generally used by the countries that conduct the survey on ICT usage and e-commerce in enterprises. Due to (small) differences in translation, in the used survey vehicle, in non-response treatment or different routing through the questionnaire, some results for some countries may be of reduced comparability. In these cases, notes are added in the data.
Detailed information on differences in the wording of the questions in the national questionnaires is available in “ Annex I. Completeness “ - worksheets related to questionnaire, coverage, additional questions.
Comparability between regions:
In the year 2024, data on regions has been delivered for the second time to Eurostat. One of the main problems of comparability between regions of the country is that some economical sectors are concentrated in a few provinces. Therefore data can become strongly geographically unbalanced, e.g. the services sector is concentrated in NL3: West Netherlands when compared with NL1: North Netherlands.
The figures of the variables number of employees (EMPL) and number of enterprises (ENT) for the combination C10TS951XK (All enterprises) and EGE10 (10 or more employees) show strong business demographics difference of 10:1 and 7:1 between NL3 and NL1 regions, respectively.
Detailed information on the provision of data on NUTS 2 regional level is available in “Annex I. Completeness“ – worksheets related to regional data.
