Skip to main content
March 30, 2020

The benefits of a taxonomy (podcast, part 1)

In episode 73 of The Content Strategy Experts Podcast, Gretyl Kinsey and Simon Bate talk about the benefits of establishing a taxonomy.

“Filtering is possible through the use of taxonomies. They have a real world benefit for people looking to find something.”

—Simon Bate

Related links: 

Twitter handles:

Transcript: 

Gretyl Kinsey:     Welcome to The Content Strategy Experts Podcast, brought to you by Scriptorium. Since 1997, Scriptorium has helped companies manage structure, organize, and distribute content in an efficient way. In this episode we talk about the benefits of establishing a taxonomy. This is part one of a two-part podcast.

GK:    Hello, and welcome everyone. I’m Gretyl Kinsey.

Simon Bate:     And I’m Simon Bate.

GK:     And we are going to be looking at taxonomy today, and talk about some of the benefits that you might encounter if you establish a taxonomy within your organization. So I think the logical first place to start is just by defining, what is a taxonomy?

SB:     That’s a great place to start. So a taxonomy is an organizing scheme that helps us make sense of stuff. Let me give a couple of examples there. One is, if you go into the library and you try and find a book, usually you’ve used something like the Dewey decimal system or the Library of Congress system. That’s an organizing scheme. Another organizing scheme that we’ve learned about in school is the way that plants and animals get placed into kingdom, genus, species, and so on. These are a couple of the most well known taxonomies. And also if you’ve shopped on Amazon, you’ve encountered taxonomies there.

GK:     Right, if you have ever used any of the tools that they have to help narrow down some of the products to what you want to buy, that’s definitely a great example. So continuing along that path, can you use multiple taxonomies simultaneously?

SB:     Yeah you can. Let’s look at the Amazon example a bit more. Assume that you’re interested in buying a shirt. There are a number of characteristics of shirts that can be used to categorize or limit your search results. Do you want a red shirt? Do you want a green shirt? Color is one of the taxonomies. What size do you want? Small, medium, large and so on. That’s another taxonomy. Do you want a long sleeve shirt? Short sleeve, sleeveless. What material do you want, cotton, silk, rayon, casual or formal? All of these things which we call facets in taxonomies can be used to narrow down the options, so you can find just the shirt that you want. If you’ve ever used a used car finder, there’s exactly that same kind of filtering is done there. And that was made popular quite a number of years ago. All this filtering is possible through the use of taxonomies. They have a real world benefit for people looking to find something.

GK:     Yeah, absolutely. And I know that’s something I think all of us have used in our day to day lives at some points, not just in maybe our careers in terms of content, but going more in that direction, how else might you use taxonomies, what else are taxonomies good for?

SB:     Well, one of them is for standardizing data. So if you think about looking for a shirt and you’re looking for a medium, there’s a whole number of different ways, actually, that a vendor might describe something as being medium. They might just use a capital M, they might use Medium with a capital M, they might use medium, all lower case, medium, all upper case, they may use size range, so size 34 to 36 or something like that. If you’re getting this data such as shirt size from multiple vendors, and each vendor has a different standard for storing the data, and you blindly pass that along, your users are going to have a ridiculous set of choices to have to define to go through to find a medium shirt. So again, that’s not very user friendly.

SB:     So by correlating all of those into a single definition of medium, your taxonomy ensures, regardless of how you receive the data, it fits into a single definition everywhere.

GK:     Can taxonomies be reused?

SB:     Yes. The same idea of sizing can also work for other things, such as other than shirts. We could use it for coats, pants, gloves, anything that has a size, we can reuse that same taxonomy with those things.

GK:     That’s really awesome. We’ve been looking at this example, started out with Amazon and went more specifically into maybe clothing that you might use these taxonomies that are built in for that. But I want to shift gears and talk about some other ways that you can use taxonomy. So how might this be something that helps on a support site?

SB:     Well, the concept of taxonomy is the same there. In a support site we want our users to be able to retrieve information so they can perform their jobs. Taxonomies help with the users being able to locate the answers to their specific questions.

SB:     So in the retail site, we talked about clothing sizes. In a support site you can use the same ideas of taxonomy to help readers narrow down product type, product name, version, and so on.

GK:     When you’ve got this kind of a taxonomy built in, how do you know that your facets are correct?

SB:     That’s a good question. The taxonomies inherently reflect the person or group that created it. Diversity is key to ensuring that any biases are surfaced. That is, you can’t just create the taxonomy by yourself. You have to work and develop your taxonomy within a group of people. And the more diverse you can make that group, the more you can be assured that your taxonomy actually is as general as it can be, that it reflects all perspectives rather than just simply your perspective. So confirmation bias, selection bias, and these can limit your perspective on the facets.

SB:     Another problem that we have with taxonomies is they do enforce a sort of top down approach. Humans naturally want to group things and then break those groups down further. How do you know the thing at the top is really at the top, and how far down do you go in subgroups? One hint that Patrick Lamb offers in his book for us is to forget our scientific traditions. Rather than trying to find a single, perfect ideal spot for an object or a piece of information and put it where it’s most likely to be found, just don’t agonize over the perfect. You just find the information.

GK:     Absolutely. And I think that that’s one area where, with some of the clients we’ve worked with, that’s where getting into things like user testing and analytics and just really acquiring the information that they need about the real world cases of how users are going to make use of those facets and how they’re going to search for information, what kinds of information they’re trying to find, and how they’re going to go about it, can really help them if they’re coming up with a taxonomy and trying to figure out what those facets need to be. If you just come up with it from the perspective of how you think it makes sense from the way you’ve designed your products or the way your marketing team wants to emphasize things in the way that they are putting out that messaging, that may not actually serve what your users need. So it’s important to try to get that information from them as much as you can, and continue to use that to make your facets better.

SB:     Absolutely.

GK:     So how are taxonomies presented to the users, what are some different examples of ways that they might come across to the user?

SB:     Well, there are several different ways that your taxonomy can be presented to the users. One way might be lists, so you might just have a simple list of items. For instance, a list of sizes or a list of product names, variety of things like that. Now the simple list, that’s the basis for where we go off into taxonomies because as soon as you get more than about 12 items or so, this gets really hard to use. If you’ve ever been on a site where you’re presented with a dropdown list that goes off the screen, it starts to get really, really hard to find the thing that you want.

SB:     A couple of examples of these symbolists might be a shopping list, or a list of animals, say. So you could just have a list of any animals, a lion, a cow, a dog, a cat, a rat, something like that.

SB:     The next way of looking at your information is with trees, or hierarchies. You divide your list into a set of related subgroups. For example, if you create a shopping list, you might want to divide your shopping list by food type. So you could have a section on your list, or a sub-list for produce, a sub-list for things you want to find in dairy, sub-list you want to find in groceries, and so on. If you’re looking at a list of animals, you might want to list them in a tree, say according to their habitat.

SB:     But of course, one problem here is that one person’s idea of how to organize these things might be different than another person’s perspective. The trees then lead us to what we call hierarchies. And hierarchy is a tree with a very strict rule about the subdivisions. The tree is exhaustive; that is, it covers everything that there is, and it is unambiguous. So for everything that you have on the list, there’s no way that it can actually exist under two different categories.

SB:     One great example of this is the standard Linnaean classification of animals, where we actually break things down into kingdom, phylum, class, order, family, genus, and species. Each one of those different divisions, and actually above kingdom, now they’ve added domain. All of these divisions actually have very specific rules about what it is that separates one from another in each of the different sub classifications.

SB:     Now, talking about these ways, we have the idea of facets, and we brought the idea of facets up before. And a facet is essentially, it’s an attribute that may represent a piece of information. The facet itself can be a list, a tree, or a hierarchy.

SB:     Now, another way that this could be presented is in a matrix, or matrices. In a matrix, actually you can have two or three facets presented in a table. Let’s take the simplest facets, which of course are lists. In a matrix, you could have a table which is two dimensional. So in the rows you could have one list represented in the columns. You actually could have another facet presented. An example of a matrix is a table in a catalog that associates a specific product number with two or more characteristics, such as capacity and operating environment. If I’m looking in my catalog, I know I need to find a piece of equipment. I know what the capacity of that piece of equipment needs to be, whether that be voltage, maximum voltage that it can handle, maximum pressure, all sorts of things. And then I also need to find that piece of equipment that works in a specific operating environment. Does it have to operate in subzero temperatures, does it have to operate in normal temperatures, does it have to work in tropical climate, anything like that. So I can use that matrix of those two different characteristics and find the specific product that I need.

SB:     And finally, a little looser than a matrix, there’s a relationship map. And a relationship map is a way you show the proximity and relationships among the different entities in your taxonomy. A relationship map could be a physical map, so it could be actually a public transport system, or the human body showing something like the lymphatic system, the nerve system, those are relationship maps. Or it can be conceptual. A conceptual relationship map is something like a mind map.

SB:     So that’s a long answer to a very, very short question.

GK:     Yeah. And I think it’s really common to see these combinations of these different ways that taxonomies are presented to users, maybe even in the same interface or the same site or what have you, kind of like we talked about earlier with the Amazon example. When you go to search, a lot of times when you have over on the left, all these different ways that you can sort the thousands of results that you get from a search, there are multiple taxonomies and ways of presenting those taxonomies at work, because you can choose things from a list, you can choose things from ranges of lists, there are all sorts of different things you can do. And so I think that, again, it gets back to what your customer base needs and how they tend to look for information about your products that you would then say, okay, which of these different ways are going to be most effective to present this taxonomy to our users? And it might be more than one. It might be a combination that you find works best.

SB:     Yep. And there’s an interesting thought here actually, that we’re talking here about all the various specific ways of taxonomies, but I think a lot of the people actually listening to this podcast are probably looking for specific solutions that usually will relate to a computer interface, such as a help system, or trying to find a specific manual or something. And I have to say that while there are all these divisions, the lists, trees, hierarchies and so on, really a lot of the time we’re going to find ourselves, for the most part, really focusing on lists primarily, and perhaps then trees and hierarchies. But really, lists are the thing, mostly because of the interface of the computer. There’s not a really effective way on a computer, just with a standard HTML interface, to be able to show somebody a tree structure. And actually within the types of things we’re talking about, there’s no real need for it either. The tree structure really works very well with animals and things. A hierarchy works very well with animals. But we really are interested in lists.

GK:     Right, and I think that’s a good place to wrap up part one of this podcast. We will be back next time with part-two.

GK:     And thank you for listening to the Content Strategy Experts Podcast, brought to you by Scriptorium. For more information, visit Scriptorium.com or check the show notes for relevant links.