Dutch patent applicants and applications

What does the survey comprise?

Purpose

The statistics on patent applicants and applications from the Netherlands describe which companies from the Netherlands are applying for patents, broken down by company class, size class, country of control and technology field. It also shows whether patent applications were submitted in cooperation with a foreign party.

Target population

The target population refers to all direct patent applications from the Netherlands to the European Patent Office (EPO) and the Netherlands Patent Office (OCNL), a department of the Netherlands Enterprise Agency (RVO). The majority of the patent applicants are linked to a company in the general business register used by Statistics Netherlands.

Statistical unit

Enterprises. These are legal enterprises or natural persons. In official terminology used by Statistics Netherlands it is referred to as the enterprise group. This is the most comprehensive unit of persons established in the Netherlands over which control can be exercised. The enterprise is considered the actual actor in financial processes, especially the financing and income distribution process.

Date/year survey started

The research started in 2022 with the publication of results for the period 2013 to 2019.

Frequency

Annually

Publication strategy

In the summer of 2022, the data for 2013 through 2019 were published. The 2018 data are revised provisional, and the 2019 data are provisional. The publication strategy is as follows: in the year T, the dataset is supplemented with data for year T – 3. The statistics for year T – 5 become final, and the statistics for year T – 4 become revised provisional. Because the data for all years can change very slightly on an ongoing basis, a revision could be made after a few years; further research will determine whether this is necessary.

How is the research conducted?

Research type

The statistics are based on registry information.

Research method

Electronic retrieval of registers of patent applications.

Respondents

All data on patent families are obtained from PATSTAT online, a database made available for a fee by the European Patent Office (EPO). The data retrieved from the database concern only information on Dutch applicants. Both applications filed with the EPO and applications filed with the Netherlands Patent Office (OCNL) are retrieved from the database. PATSTAT is based on the central EPO documentation database (DOCDB). The address information of Dutch applicants is sometimes incomplete in PATSTAT, which is remedied by extra address information supplied by the Netherlands Enterprise Agency (RVO). This information is only used to link applicants to the business registry and is not published.

Sample size

Not applicable (integral observation from registers).

Checking and correction methods

Based on the name, address and residence data of the patent applicant, a link was established with the statistical unit enterprise from the general business registry. The choice was made to link a patent application to an enterprise because it is conceptually clearer and practically easier than to any of the other statistical units. The ownership of a patent is usually at the enterprise level and is difficult to translate to particular (local) parts or branches of an enterprise. It should be noted that universities and institutions are also included in the enterprise unit. Not all patent applications could be linked. Patent applications that could not be linked to an enterprise from the ABR include patent applications filed by a (natural) person.

Weighting

The results of the applicants (enterprises with patent applications) and applications have been weighted. If two or more enterprises are involved in one and the same patent application, then the application counts for a fraction (depending on the number of enterprises involved) in the classification by industry, size class or country of control. Also, in cases where an application is classifiable by two (or more) technology fields, it will be counted fractionally in each technology field. If an applicant has applied for patents in two or more technology fields then this also counts as 1 applicant in total, while the underlying technology fields are counted fractionally.

The presented classification by technology field is based on the International Patent Classification code. Technology codes are assigned to patent applications from a technological perspective by the patent granting organization. The code(s) assigned to a patent application depends on the function of the invention or the field of application. The IPC is divided into sections, classes, subclasses, groups and subgroups (in total, the IPC covers some 70 000 categories). Based on the IPC, Statistics Netherlands has compiled its own aggregates. In addition, the table also shows an aggregate of technology fields classified according to the technology concordance table of the WIPO (World Intellectual Property Organization). This classification links the symbols of the International Patent Classification (IPC) with thirty-five technology fields. Each patent application is assigned to one or more main fields according to the IPC codes to which it relates.

Quality of the results

Accuracy

The statistics are based on registry information that is enriched by Statistics Netherlands with information from the general business registry. Linking the registry information to the general business registry will not be completely error-free; it is possible that in a rare case an application is linked to the wrong enterprise or that a patent application is erroneously not linked.

PATSTAT is based on the central EPO documentation database (DOCDB). The EPO also produces patent statistics. The numbers for the Netherlands as reported by the EPO are higher than the numbers published in this study. There are several explanations for this. Firstly, PATSTAT only includes published patents, while the EPO also incorporates applications that have not (yet) been published or applications that are later withdrawn. In addition, worldwide applications are included by the EPO, while these applications are not included in this study. This refers to so-called EURO-PCT applications: these are patent applications to the World Intellectual Property Organisation (WIPO), for which patent rights are also sought in one or more countries that are members of the European Patent Convention. Applications outside Europe that have reached the European phase are also included. These applications were not included in this study because the worldwide applications from PATSTAT do not contain address information, so the link to the ABR is not possible with the current method. This explains most of the difference from the EPO publication.

Sequential comparability

The figures compare well sequentially. The data from the most recently published year T – 3 and the year T – 4 are considered provisional and revised provisional, respectively, when a new publication is made. In particular, the year T will be further supplemented at the next publication with data that were not yet available in PATSTAT online at the time of the previous publication.

In PATSTAT online, data may be added to earlier periods, which happens with the two updates of the database in spring and in autumn. This is because the EPO depends on the data provided by national patent offices, which is sometimes delayed. These changes in the earlier periods are sufficiently small that the figures in these years can be considered final for now, but after a few years it may be necessary to revise the statistics if the changes become more significant.

Quality strategy

The matching between the patent registries and the general business registry was initially done automatically through a number of algorithms that compare name and address information. A score was assigned to each automatic match that can be interpreted as a "report grade" of the match. It indicates the extent to which the match can be considered correct. After the automatic matching process, a number of operations were performed to increase the quality of the match. This involved not only checking the matches but also making extra efforts to manually link matches with low scores and unlinked applications.