The steps to follow to build a documentary fund

The creation of a documentary collection involves the development of a clear and formalized research and collection strategy, directly linked to the strategic question(s) that the monitoring is supposed to answer.

This is a process that unfolds in several key stages, each requiring attention and a methodical approach. Here is a guide to navigating these stages and developing a structured and relevant document collection.

1. Determine the search scope

The first step is to clearly define the subject of your monitoring. This involves specifying the area to be investigated, the specific questions you wish to answer, and the objectives of your research. This essential step allows you to target subsequent research more effectively and avoid collecting superfluous information.

2. Select relevant sources

Not all information sources are created equal. And a single source of information is rarely sufficient. This is why, since its creation in 2004, TKM has always prioritized the ability to collect and analyze a wide variety of information sources in the development of its software platform : patents, scientific publications, the web, theses, conferences, collaborative projects, startups, etc.

It is therefore essential toidentify the most relevant databases before embarking on a state-of-the-art review or setting up regular monitoring. Some databases are free, others require payment. Costs can quickly become a limiting factor. This is why TKM also offers privileged access to its own Data Lake, a global database of patents, scientific articles, projects, and startups.

3. Implement a collection strategy

Once the search scope is defined and the sources identified, it's time to list the keywords that will enable this exhaustive and relevant data collection. Each database has its own " grammar ," and it's useful to understand it.

However, Boolean operators (AND, OR, NOT), truncation operators , or proximity operators (NEAR, W, etc.) will most often help you refine keyword sets that combine efficiency and comprehensiveness. Ideally, it's best to keep a little noise (10 to 15%) rather than risk missing useful information.

But if this exercise becomes complicated, at TKM the team of analysts, composed exclusively of high-level scientists, has developed real expertise on the subject over the years.

4. Organize and index documents

Nowadays, regardless of the subject, collecting documents most often results in a substantial volume (several hundred to several thousand, or even tens of thousands). Before the analysis phase, a phase of meticulous curation, enrichment, and organization is essential.

This can involve adding useful metadata, extracting named entities useful for your research, or normalizing affiliations (names of organizations) and semantic indexing of documents.

Next, the implementation of a classification system, or even tags or selection will be useful to facilitate subsequent searches : yours but also those of the client teams in the end: management, R&D, IP, innovation, corporate venture, etc.

5. Extract useful information

Simply collecting information is not enough. But if your document collection includes several hundred documents, you probably won't want to read everything... and besides, that's not advisable.

Tools for semantic segmentation, research, weak signal detection and even “outliers” will be extremely useful for:

• to reinforce your vision of the field (which you already knew),
• but also to detect weak signals  (emerging technology, start-up, new patent, strategic pivot of a competitor…).

Thus, by segmenting your document collection, and without having to read everything, you can in a few hours build a strategic understanding of the subject, initiate an internal approach to collective intelligence and thus feed into the strategic management of your organization.

And finally, some particularly inspiring or strategic documents (opportunity or threat) will require in-depth reading.

The purpose of the document collection is to give you the opportunity to approach your strategic issues at a global level, while allowing for back-and-forth at the level of individual documents.

The strategic intelligence you build takes on a new dimension and never fails to capture the attention of your internal clients or management. The goal is to transform raw data into useful knowledge for decision-making.

6. Feed and update your fund

A document repository is a living tool that may require constant updating and enrichment to remain relevant. Even in the case of ad-hoc monitoring (mapping, state of the art, landscaping, etc.), it is quite common for the decision to be made to evolve it into a periodic monitoring system.

The frequency of these updates will depend on the strategic importance of the issue, but also on the technological developments in the field. For example, monitoring AI or cosmetics requires at least a monthly update, given the volume of knowledge produced by research and industry stakeholders.

For this reason, among others, it is essential to keep a precise formal record of the search strategies used.

Then apply them regularly to the different sources, unless you have access to a global document repository like TKM's "Data Lake", through which updates can be easily automated.

The new data will also need to go through the various steps outlined in this article. This task can be tedious and can quickly turn into a nightmare for the analyst. However, certain tools can save you valuable time.