Welcome to the new version of the Patents Manual. Please note there are changes to the numbering and sequence of the chapters and pages in the manual. You are encouraged to take the time to explore and familiarise yourself with this new structure.

Annex VA - Internet searching 1

Date Published

Internet Patent Searching Guidelines


The Internet searching group was asked to investigate the usefulness of the Internet for patent searching purposes and to recommend guidelines with the following constraints:

  • Searching is of PCT minimum documentation

  • Searching patent information NOT NPL (NPL to be considered later)

  • Limited to sites currently available to examiners ("free sites")

The group collected data on current searching practice and consulted with a technical panel of examiners and senior examiners on the features offered by the sites available.

The following comments and guidelines apply to all original searching.

Usefulness of the Internet

The working group agrees that patent searching on the Internet presents opportunities to reduce costs and improve the quality of our searches but only where that tool is used intelligently to enhance the effectiveness of the standard WPAT or CAS search. In this regard the time resource expended on Internet searching must be regarded in terms of the time saving it will produce in the WPAT or CAS search. Generally this will be by a narrowing of the search field. Internet use without a clear strategic objective is not justified.

Generally the data collected showed that examiners are making appropriate judgements and are receiving benefit from their Internet use. However some feedback suggests that long Internet sessions are being conducted (up to 3 hours) without any benefit identified. Hence it is appropriate to provide some guidelines to assist examiners in their value judgements concerning the Internet.


The Commissioner, Deputy Commissioners and employees of the Patent Office are under legal obligation not to disclose information about matters proceeding under the Act except where required or authorised by the Act, eg at OPI (Section 183).  See also Article 30 of the PCT.  Our customers rely on these provisions in their commercial and IP strategies and a perceived failure in our confidentiality obligations would have serious implications for our ability to attract searching work and for the prestige of the Office; an actual failure could result in a legal claim for damages.

Before commencing a search on the Internet it is paramount that examiners consider whether their proposed search terms would amount to disclosing the searched invention over the Internet.  A selective keyword search, generally associated with mechanical and electrical technologies, is unlikely to disclose the invention.  To avoid such disclosure it is best to use only a few search terms, at least initially.  The number of search terms can then be increased in stages until an appropriately focused result is achieved.  However it is essential to check at each stage that the invention would not be disclosed by the increased number of search terms.  

In some areas of chemistry and bio-chemistry, the entry of a specific alloy, compound or gene sequence may disclose the invention over the Internet.  Consequently the entry of such terms over the Internet is unacceptable.  It is more appropriate to use a dedicated and secure communication medium for this purpose.  

Generally the Internet is an insecure communication channel.  Hence it does not matter whether database providers assure confidentiality or not, or whether providers log accesses to their databases or not.  The primary issue is the level of security of the communication medium to the database.  It may be that the Internet could be made secure if appropriate encryption techniques over the Internet were employed.  While such developments may overcome security concerns in the future, the principle to be followed is that we are not to enter anything over the Internet that discloses the invention.


The attached flow chart (Annex V.5) outlines the recommended procedure with guidelines for key decision points. Ultimately the optimum use of search tools for any particular case is dependant on those individual circumstances and examiners are expected to use their professional judgement in conjunction with the views of the three person search team and if necessary in consultation with senior or supervising examiners. The guidelines are repeated below with their explanation.

Will a quick Internet patent search assist in refining or validating the search strategy?

An Internet patent search may be appropriate where:

  • IPC doesn't cover invention well

  • Not confident about keywords

  • US uses different terminology

  • Art is largely US based

  • Technology is not well understood.

  • Search is not dependant on viewing drawings or chemical structures

  • Citation disclosing all relevant features is very likely


The points listed above are some of a number of situations where the benefits of using the Internet are likely to justify the time spent. It should be remembered that the databases on the Internet are a subset of those available through Derwent and if an Internet search is not likely to produce a change in the eventual WPAT/CAS search strategy, or confirm the validity of a narrowly focused strategy, it shouldn't be done.

Choose Internet site/database.

Irrespective of the site/database chosen, we are not to enter anything over the Internet that discloses the invention.

Keyword searching:  USPTO site (abstract search) is most recommended followed by Espace.

Piggy back searching:  IBM site is most recommended followed by USPTO.

Complex Search > 5 TermsFull TextAbstractClaimPiggyAnnexes V.6-V.8 provide a comprehensive comparative list of features







*Not recommended


The choice of the internet site used in any particular case will ultimately depend on the nature of that case.  However, there are a number of limitations that should be taken into account when deciding which site to use.  In summary these are:

USPTO  At the USPTO site it is possible to conduct complex searches of US patents on either the abstract database or the full text database. In most cases the abstract search would be adequate.  Full text searching is much slower and should only be done in cases where there are good reasons for doing so (perhaps the three person team should be consulted in these cases if a significant amount of time is going to be spent). Claim searching is possible and may be a useful option in some cases. It is possible to do piggy back searches however titles of the prior art are not provided.

Espace  In Espace it is not possible to perform a complex search.  The maximum number of search terms is about 5. Another major disadvantage is that truncation of the search term is not possible.  The major advantage of Espace is that one search can cover patents from the main countries. It can also be useful for searching excluded subject matter that would not be patentable in the US. Piggy back searching is not possible in Espace since the prior art is not listed anywhere as a hyperlink.

IBM  The IBM site is a commercial site.  Consequently we need to be sensitive to our customers' views about using this site.  For instance a patent applicant, who may be a competitor of IBM, may view our use of the IBM site as leading to possible disclosure of their invention to a competitor.  The use of the IBM site must be determined with such considerations in mind.  The same considerations apply if use of any other commercial site in any other technological field is contemplated.

At the IBM site it is possible to conduct complex searches of the abstract and claims of US patents and complex searches of the abstracts of EP, WO and JP patents.  It is noted that on this site each country must be searched separately.   

Piggy back searching is very convenient on this site because the titles of the prior art documents are provided with each patent.  This allows scanning of the titles before looking at the abstract of the prior art.  This makes the IBM site the preferred choice for piggy back searching.  If appropriate, patent numbers can be extracted from a keyword search on USPTO or Espace and then transferred to the IBM site for piggy backing purposes.

Continue Internet search?

Stop if highly relevant art is not found after several set of keywords are entered. Normally the benefit of an Internet search is obtained in the first 10 to 15 minutes.

Any search taking more that 15 to 30 minutes which has not produced significant value should be terminated and the matter discussed with the three person team.


The data collected indicates that best practice lies in short Internet searches that produce either citations or better keywords that help to confine the WPAT/CAS search. Elaborate keyword strategies developed on the Internet (often with the full text option) are usually too narrow to transfer to WPAT and this suggests that the time taken to develop them is not well spent.

There are however rare circumstances when narrow full text searching is appropriate, eg if a key distinguishing feature of a claim is likely to be mentioned only in the detailed description of the prior art.

Piggy backing?

Piggy backing (on the Internet or otherwise) is justified where there is some certainty that better X citations will be found and (if done before the WPAT search) there is some risk that they will not be identified in the WPAT/CAS search.


Piggy backing (or forwarding) is an old technique that is facilitated in many ways by the linking of documents on the Internet, particularly from US reference lists but also from EP and GB search reports. It is however a time consuming activity and should not be done merely because the facility exists. Usually it is done after the WPAT/CAS search and potentially produces highly relevant art that would not otherwise be located. If it is done before the WPAT/CAS search there must be a good indication that better citations will be obtained that allow a narrowing of the subsequent search or that could not be located by other means. Using piggy backing to locate documents that are not more relevant or that will be turned up anyway by the WPAT/CAS search is a waste of time.

Amended Reasons

Amended Reason Date Amended

Publishing for testing

Back to top