What does the survey comprise?
The survey underlying online turnover statistics gives an insight into changes in turnover generated by online sales in the Dutch retail sector.
Enterprises that are economically active and that are principally active in retail trade in the period under review. Retail trade refers to the sale of new and/or used goods not produced by the seller to consumers (not professional users). The survey covers enterprises with ten or more employees, petrol stations and dispensing chemists are not included. The online turnover of these enterprises accounts for about 70 percent of total online retail sales.
Date/year survey started
The figures are available from January 2014.
Statistics Netherlands usually publishes the first results six to eight weeks after the reporting period. Up to sixty days after the quarter, figures may be adjusted as increased response becomes available. After that the figures will not be changed. Five months after the reporting year, all late response is used to compile the definite figures, for all months in the reporting year. Until that point, figures are provisional. After that, figures are definite.
Definite figures are only revised if significant adjustments or corrections are necessary.
How is the survey carried out?
Sample survey among:
• enterprises in retail sale via mail order and internet (SIC 47.91) with ten or more employees
• enterprises in retail branches (SIC 47) other than retail sale via mail order and internet with 50 or more employees
A model is applied in order to estimate the online turnover for enterprises with 10 to 49 employees within the retail branches other than retail sale via mail order and internet.
Data on online turnover are collected as part of the Monthly Retail Survey. Respondents may submit the required data electronically or on paper.
The sample for online retail statistics comprises about 750 enterprises. These enterprises are also included in the Monthly Retail Survey.
Checking and correction methods
Statistics Netherlands conducts plausibility checks on data received from enterprises. Among other things the figures are compared with those for previous months/quarters and those for the same month or quarter of the previous year. If the latter figures are lacking or are too ambiguous, the figures are compared with the averages of similar enterprises ( i.e. companies of the same size active in the same sector). If inconsistencies are established, they are corrected automatically or manually. For reasons of efficiency, manual correction is almost only used for the largest enterprises, in view of the impact of the figures of these enterprises on the final total figures of the sector to which they belong.
The survey data of the responding enterprises are raised to the total number of enterprises in the retail sector in the period.
The indices are weighed according to their relative share of turnover within the retail sector.
A so-called base shift is performed every five years, in order to align the index series to the present day business structure. Currently, the reference year is 2013 (2013=100).
Quality of the results
As this survey is currently in an experimental phase, the quality will be continually assessed and methodology, population coverage and output may be adjusted. This should be taken into account when using these data. Changes in methodology and any resulting corrections in the figures will be specified in the table notes.
As the results are mainly based on a sample survey, there is a certain margin of error.
As the turnover statistics aim to show current and recent developments, Statistics Netherlands corrects for important breaks in series that occur as a consequence of changes in methods and definitions.
The survey results are only published after the plausibility checks. These checks include:
• Time-series analyses (consistency in time);
• indicator analyses (relationship between variables);
• confrontation with results from other sources.