I’m currently researching different ways to provide content recommendations to the users of our technical documentation.
Content recommendations can dramatically improve the user experience of our readers, especially when it comes to guides containing hundreds and even thousands of topics. Because let’s be honest, nobody wants to read a technical guide cover to cover, just for the sake of it. Our readers want to quickly find a solution to a problem they’re facing with the software and keep on working. So the question is: “how can we help them find the information they need?” Content recommendations might be the answer.
Content recommendations allow tech writers to provide relevant information based on different criteria; e.g. user role, goal, previous clicks (user journey), searches, etc…and hide all other information that is not only irrelevant, but also contributes to overwhelm the reader.
Sometimes irrelevant information can be even misleading, causing customer frustration and probably more calls to support, which also translates into money loss for the software provider (“us”).
Providing recommendations is mainly about shifting the responsibility to the tech writers, who should identify relevant content, while saving the reader from hunting for buried information.
Have you ever wondered how can you spend hours watching one YouTube video after the other or how easily can you get hooked on Netflix? The answer is: these companies learn what you like and serve it to you on a silver platter. As simple as that. The question is, can we replicate this in a technical documentation context? (replace the verb “like” by “need”)
What is the real effect of providing content recommendations for technical documentation? Recommendations have a very positive impact on customer satisfaction and builds brand loyalty. The user needs are better met, as they are a click away from finding the solution they demand, and the user documentation plays its main role: problem solving.
So, how can we as technical writers provide content recommendations? By taxonomizing our content. Creating a taxonomized documentation sounds like a lot of work, but surely pays off in the short term.
But wait, what is exactly a taxonomy? A taxonomy is a system that you can use to organize and categorize your content.
Taxonomies can be structured in two different ways:
- As hierarchical relationships: which can be organized in a parent-child structure, with main categories and sub-categories with different specification levels.
- As facets: which makes possible to assign multiple classifications to the same item. Facets can be used for searches with multiple terms and support complex search queries.
Some of the most commonly-used taxonomies are: product name, version and user role. Additionally, you can consider the following categories:
- Technology (e.g. database type -> Oracle, MSSQL, Postgresql)
- Accessibility (How should the content be displayed?)
- Guide (administration, installation, integration)
- Goal (What can be achieved by reading this content?)
- Multimedia (in case you want to give more visibility to topics with videos)
There are some metadata models that can be used as a basis for creating you own taxonomy (like the “Dublin Core”) but the number and granularity of the metatags might vary greatly depending on the type of documentation you are creating and also on the goals you want to reach with the content recommendation. A taxonomy for technical content must be flexible, and adaptable. It must continuously evolve, and must be aligned with the business goals. This alignment drives more value from the documentation.
As a general rule, your taxonomy should:
- Be created on a global level (not on a project level)
- Consider specific needs and requirements of all types of readers
- Whenever possible, integrate with other systems containing metadata
- Be tested with real business data
- Don’t be too granular, as it is difficult to maintain
- Be constantly updated (it is not a one-time thing)
- Contain only relevant attributes (the obsolete ones should be removed)
Here you can find an example of a taxonomy of research in IoT technologies: https://www.researchgate.net/figure/Taxonomy-of-research-in-IoT-technologies_fig9_312957467
This idea expressed by Tom Johnson in his blog, really resonated with me: “If information is truly a corporate asset, leveraging it in different ways through metadata should be a key strategy. It leads to a major competitive advantage”
Once the taxonomy and metadata have been designed, it’s time to create a content model, which basically defines how the metadata should be used.
It is important to note that metadata plays also a very important role for developing chatbots for technical documentation.
I’ll keep looking into best practices to leverage the use of metadata/taxonomies and researching content recommendation engines for technical documentation, and write down my findings in a future post.