This is the eighth article in the series ‘The Art and Science of Taxonomy Development for Market and Competitive Intelligence’. Click here to view other articles.
Software plays a pivotal role in the success of a taxonomy implementation. With an explosion of digital information and the rise in popularity of new media types such as video, images, and audio, we need increasingly sophisticated software to manage taxonomies. The following are some key considerations to make while selecting the right software for your taxonomy implementation.
1. Browse, Search and Explore Taxonomy
It should be easy for a user to view the overall structure of the taxonomy and find out what is included and what is excluded. That is, view a term in the context of its parent-child relationships, synonyms, and its scope definition. In addition to browsing and navigation, the user should be able to search the taxonomy. The search should not be limited to the terms’ names but also search the scope definition, other synonyms, and related terms.
2. Automate tagging for information
The value of a taxonomy comes from the accurate tagging of information. To tag accurately, users need to remember the complete taxonomy and be willing to make additional manual efforts for tagging. And even if they are willing, the results won’t be consistent because of the subjective context of terms. At Contify, after several years of iterations, we have devised a hybrid approach of augmenting machine learning algorithms with a layer of human curation to drive the best results. Contify provides options to write tagging rules as well as use machine learning to automate the tagging process.
3. Suggest New Terms
The software should have features to identify suggest new terms. For example, it can save the user search keywords for analysis. Based on this search analysis, the software can help to identify new terms or alternate ones (synonyms) to be included in the taxonomy. It should also be able to identify new terms based on cluster analysis (unsupervised machine learning) of the aggregated information.
4. Workflows to Manage Taxonomy
A taxonomy evolves by adding new, deleting obsolete, merging redundant or splitting confusing terms. These actions are discussed in detail in the previous article. The software should be designed for the efficient execution of these actions.
After making changes to taxonomy, the existing content needs to be re-indexed (or re-tagged) so that the updated taxonomy is retrospectively applied to the archived content and not just to new content. To enable this, the software should have features to automate the process of re-indexing the archived content.
The software should make it easy to generate reports that are essential to maintain a healthy taxonomy. For example,
- Number of documents tagged with each term.
- Number of times a term was auto-tagged incorrectly and was manually corrected.
- Number of times a term was missed by the auto-tagging logic and consequently added manually.
- Changes made to the taxonomy over a period of time — this report is essential for the auditing changes and for preparing a regular taxonomy change log for all users.
7. Downloading a Copy of the Taxonomy
The number of terms in a taxonomy increases over time. Also, the relationships between terms become complex. It is, therefore, difficult to have a consistent and complete view of the complete taxonomy. The software should have an option to download the taxonomy. The download should include:
- Term name and synonyms
- Parent – child relationships
- Rules written for the terms
- Scope definition
- Date added/modified/discontinued
- History notes
8. User Support and Assistance
The software should allow users to report errors, request changes, and ask queries. The user interface of the software should be such that these options are easily accessible to the users as they use the information system. In Contify, users appreciate the in-app chat support the platform offers.
- Excel is still the recommended tool of choice for the initial phases of taxonomy development. The process of taxonomy development is already a pretty complex process, and it is prudent that we don’t add another variable to the initial phase. Excel makes it easy and straightforward to gather terms, arrange them in hierarchies, list scope notes, and add sample content. You can write notes for users and colour-code cells to highlight the terms that are new, being merged or deleted. More importantly, most users already know how to use Excel — they don’t need to learn a new tool.
- The software should be able to handle any special characters required by the language of the controlled vocabulary. Compliance with the latest Unicode standard is recommended for all non-Roman characters.