In this blog I will be introducing the KISS Taxonomy Implementation (KTI). Be sure not to confuse it with the rock band KISS, because KTI rocks too!
This new KISS Taxonomy concept allows you to implement taxonomies at the speed of light compared to the traditional plodding methods. The benefits of this taxonomy are:
- Surfaces the right results in search
- Gets users the information they need faster
- Reduces the number of refiners needed to get to a result
- Expands your term to get all results, making it broader
(In future blogs I will discuss how KISS Taxonomy provides a more personalized search experience to your users).
So, what exactly is a KISS Taxonomy Implementation (KTI)? KTI is an agile and iterative process that gets it done fast and is:
- Driven by usability versus analytical perfection (users are more important than anything)
- Only three levels deep – Broader is better
- Achievable quickly, providing the ability to measure impact and improve it
Driven by Usability vs Analytical Perfection:
A Kiss Taxonomy is all about usability. It’s more important to start moving and therefore take quick steps rather than experience paralysis by analysis. A lot of taxonomy projects start out with over analyzing all of the exact categories that a company is using. There is nothing wrong with this approach, but what I have seen is that the “over analysis” provides the scientific approach, but it is not necessarily how users work. In addition, you can create categories that are not used or needed. The key is to focus on what the users do. A great way to accelerate this process is to grab a screen shot of the folder structure of your CXOs and the Head of Product Engineering. I like to use their structures as a starter because at the end of the day, they are leading the company, and if you are going to all be on the same page, then the goal should be to get on your management team’s page!
Three Levels Deep:
Kiss Taxonomy is broader. You can go one, or two, but no more than three levels deep- no matter what. The KISS Taxonomy is about keeping it simple and providing faster access to content. There will be times in which a deeper taxonomy will either be required or make better sense, but what I have learned is that while added depth provides added granularity, it can also make it harder to manage. And, more importantly, harder for the non-subject matter expert (SME) to use.
Let’s take a quick look at an example of a traditional taxonomy. Since AI is such a hot topic, let’s use that as an example:
- Artificial Intelligence
- Cognitive Computing
- Machine Learning
- Natural Language Processing
- Neural Networks
Artificial Intelligence is the top category because it is the umbrella/top category of all the other topics. Cognitive Computing, Machine Learning, and Neural Networks are all subset categories of Artificial Intelligence. Natural Language Processing is a subset of Machine Learning. On the positive, it looks “pretty and neat”, and you can most likely guess that Artificial Intelligence is at the top and then it gets more granular as you go to the next level.
However, if we are building a KISS Taxonomy, then this would all collapse into just one level, with five categories. While there would be some value in the added depth, if there are only five categories, then it is just quicker to see the list, select a value, and filter the results to your topic. We will talk about when to add depth in our next KTI blog, but for this initial concept, we are going to collapse it into a list/one level taxonomy:
- Artificial Intelligence
- Cognitive Computing
- Machine Learning
- Natural Language Processing
- Neural Networks
Iterative:
Kiss Taxonomy is iterative. Traditional taxonomy tends to be set and forget. KISS Taxonomy, however, is always improving. There are small updates every two or four weeks. You will take your initial, very specific terms and start to expand them, regardless of how they look. However, you should expand the terms in alignment with how they are used within your company or department, or how they will be interpreted. If we take our original example, then after two weeks we might add these simple expansions to our initial query:
- Artificial Intelligence – AI
- Cognitive Computing
- Machine Learning – ML – Subset of AI
- Natural Language Processing – Subset of AI and ML – Understand human speech
- Neural Networks – Works like human brain
What does a bigger suggestion provide you? Well, if you look at the fourth and fifth terms, doesn’t this approach start looking like type ahead? Hopefully it should, because essentially you are expanding the terminology and aligning it with how your company would use that term to make it easier for the end user who might not know the difference between Natural Language Processing and Neural Networks.
Measurable:
This will be tough for a lot of people in the beginning because you might not have an easy way to measure. However, you should be able to at least use standard, out of the box reports for whichever tool you use as a search engine. This will allow you to at least see the search queries that people are using. You should observe a correlation of typical search terms in alignment with your KISS Taxonomy.
Essentially, we are taking a process that you implement with your website for SEO and applying those concepts to your intranet. A key difference between what you do internally versus what you do externally (WWW) is that externally you often pay for ads, and because you pay for ads, you demand a report that shows the success. There is also a tendency to scrutinize external search traffic and continue to refine keywords, etc., but this is not often the case with internal search. I could never quite understand why a company pays and cares so much about search externally, but as soon as we take that same process and run a search query on our intranet, we tend to not care or look the other way- or we just have pure indifference.
That’s it, that is the KTI process. Feel free to send me questions or make comments about your KISS Taxonomy Methodology experience. Our next blog will cover when and where to expand to a second and/or a maximum third level. I will also compare a KISS taxonomy to other types of taxonomies and when and where they make the best sense to use. Our third blog on KTI will review how to take the limitations of taxonomies and turn a known laborious manual tagging process into a lean, mean tagging machine. Our fourth and final blog on this topic will be how measurement and machine learning can be used in alignment with the tagging process to further automate KTI and provide a personalized search experience.