Book Review-The Ethnographic Interview

I’m about as far away from an ethnographer as you can get. I live in the heart of the United States and in the same home for over 20 years. And yet, I use ethnographic interviewing in one form or another every single week. How can it be that I’m not embedding myself into new and strange cultures, and yet I value skills that resemble those needed by an ethnographer so deeply? The answer lies in the techniques and thinking that The Ethnographic Interview teaches and in my work world.

I came to The Ethnographic Interview by way of Peter Morville’s work, Intertwingled. He recommended it as a way to understand information architectures – and corporate cultures – more completely. I agree. All too often, the issues we have in understanding one another are about how our cultures differ, and no one has bothered to understand the unwritten meanings behind the words we use.

Requirements Gathering

Before I share some of James Spradley’s insights into ethnography, it’s important for me to cement the connection between what people do today and what ethnography is, so that it’s criticality can be fully understood. In IT, business analysts – by role or by title – seek to understand the foreign world of the business. They learn about logistics, manufacturing, marketing, accounting, and more in an effort to translate the needs of these groups for the developers and systems designers that will create IT systems to support them.

Even the experienced business analyst who knows the company and the department well must do their best to remove all of their assumptions and start fresh in understanding what the group is doing and what they need. While it’s technically impossible to remove all assumptions, because they are so good at hiding, the ethnographer’s task is to eliminate as many as possible and to test those that remain.

I wrote a course for Pluralsight some years ago, titled “Gathering Good Requirements for Developers,” where I teach a set of techniques designed to expose assumptions, test them, and make things feel more real and understandable on both sides.

The requirements gathering process, whether a part of agile design or traditional waterfall methodologies, is absolutely essential to being able to deliver what the business needs. The process of requirements gathering is ultimately a process of eliciting and understanding what the foreign culture is saying – even if that foreign culture is inside of your organization.

What is Ethnography?

An anthropologist is expected to be off in a foreign land eating strange food and spending most of their time wondering what people are saying and what the heck they’re doing so far from those they love. Ethnography is their principle work, which is the systematic study of the culture they’ve embedded themselves in. Put differently, the goal of ethnography is (according to Bronislaw Malinowski) “to grasp the native’s point of view, his relation to life, to realize his vision of his world.”

Simply stated, it’s learning from people. However, there are several nuances. First, ethnographers invite natives to teach them. They don’t assume that they know or can learn the culture without help. Second, there are components of the culture that aren’t ever directly expressed. For instance, in the United States, the phrase “How are you?” is typically a greeting. The typical response is “I’m doing well, and you?” It doesn’t convey a real interest in the other person – until and unless it’s followed with, “I mean, really, how are you?”

Dig Deeper

If there’s one thing I’ve found that is a problem with requirements gathering, information architecture, or just working with other people, it is that we don’t truly understand. We believe we understand. We might be using the same words, but we just aren’t 100% in alignment. That’s where training in ethnography is really helpful.

Ethnographers observe behavior but inquire about the meaning. They understand objects but seek to discover the meanings that the culture assigns to these objects. They record emotions but go beyond to discover the meaning of fear, anxiety, anger, and other feelings.

In short, they dig deeper. They verify their understanding to ensure that what they believe they understand is actually right. Consider for a moment death. It’s the punctuation mark at the end of life – every life. Yet, different cultures view death differently. Some cultures keep death hidden – as is the Western point of view – while others embrace or celebrate it. Some cultures believe in reincarnation and others in an afterlife. It’s the same event, but it’s culturally very, very different.

Gary Klein explains in Sources of Power that we all make models in our head, and it’s these models that drive our thinking. He also shares how painful it can be to get these models to surface. The models are tacit knowledge that cannot be expressed in explicit language. In fact, Lost Knowledge differentiates between tacit knowledge and what’s called “deep tacit knowledge,” which are mental models and cultural artifacts of thinking that are so ingrained the person literally can’t see them.

The person the ethnographer is talking to, the informant, needs promped to access the information they don’t know they know. A good ethnographer can tease out tacit knowledge from even the worst informants – but finding the right informants certainly makes it easier.

Indispensable Informants

If you follow agile development practices, you may notice that agile depends on a product owner who is intimately familiar with the business process that the software is being developed for. Lean Six Sigma speaks of getting to the gemba (Japanese for “the real place”) to really know what’s happening instead of just guessing. Sometimes this is also used to speak of the people who really know what’s going on. They do the real work.

The same concept applies to ethnographic research. You need someone who is encultured, really a part of what you’re studying. While the manager who once did the job that you’re looking to understand might be helpful, you’ll ideally get to the person who actually is still doing the work. The manager will – at some level, at least – have decided that they’re no longer a part of that group, and, because of that, they’ll lose some of their tacit knowledge about how things are done – and it will be changing underneath their knowledge anyway.

Obviously, your informant needs to not just be involved with the process currently, but they also need to have enough time. If you can’t get their time to allow them to teach you, you won’t learn much. Another key is that the person not be too analytical. As we’ll discuss shortly, it’s important that the informant be able to remain in their role of an encultured participant using their natural language rather than be performing translation for the ethnographer – as they’ll tend to do if they’re too analytical.

You can’t use even the best interviewing techniques in the world to extract information that no longer exists.

Interviews

The heart of ethnography isn’t writing the report. The heart of ethnography is the interviewing and discovery process. It’s more than just asking questions. It’s about how to develop a relationship and rapport that is helpful. The Heart and Soul of Change speaks of therapeutic alliance and how that is one of the best predictors of therapeutic success.

Tools like those described in Motivational Interviewing can be leveraged to help build rapport. Obviously, motivational interviewing is designed to motivate the other person. However, the process starts with engaging, including good tips to avoid judgement and other harmful statements that may make a productive relationship impossible.

For his part, Spradley in The Ethnographic Interview identifies the need for respect or rapport and provides a set of questions and a set of interviewing approaches that can lead to success.

Types of Questions

At a high level, ethnographic questions fall into three broad categories – descriptive, structural, and contrast questions. These questions allow the ethnographer to dip their toes into the water of understanding, structure their understanding, and understand terms with precision.

Descriptive Questions

Descriptive questions are by far the most voluminous questions that will be asked. They form the foundation of understanding what is in the informant’s world and how they use the objects in their world. Descriptive questions fall into the following categories:

  • Grand Tour Questions – These questions ask for a tour around the topic
    • Typical Grand Tour Questions – Asking for a typical situation in their environment
    • Specific Grand Tour Questions – Asking for a specific time and what happened
    • Guided Grand Tour Questions – Asking to see the specific things happening in an area of the informant’s environment
    • Task-Related Grand Tour Questions – Asking the informant to explain a specific task that they do and how they do it
  • Mini-Tour Questions – Mini-tour questions are the same structure as grand questions but focused on a smaller area of the informant’s experience.
    • Typical Mini-Tour Questions
    • Specific Mini-Tour Questions
    • Guided Mini-Tour Questions
    • Task-Related Mini-Tour Questions
  • Example Questions – Asking for a specific example of something that the informant has answered in general
  • Experience Questions – Asking for experiences that the informant might have found interesting, relevant, or noteworthy
  • Native-Language Questions – Asking how the informant would interact with someone else from the culture – in the language that they use
    • Direct Language Questions – Asking what language they use to refer to something in their environment
    • Hypothetical-Interaction Questions – Asking questions about hypothetical situations that the ethnographer creates
    • Typical-Sentence Questions – Asking what kind of sentences that would be used with a phrase

Descriptive questions allow ethnographers to amass a large amount of information, but that information is unstructured and unconnected. While it’s necessary to spend some time in this space, after a while, it will become necessary to seek to understand how the informant organizes this information.

Structural Questions

As important as building a vocabulary is, understanding the relationships between various terms is more illuminating to the structural processes that the informant uses to organize their world. We use symbols to represent things, and these symbols can be categories that contain other symbols. This is a traditional hierarchical taxonomy like one might find when doing an information architecture (see Organising Knowledge, How to Make Sense of Any Mess, and The Accidental Taxonomist).

In truth, there are many different kinds of ways that symbols can be grouped into categories, and understanding this structure is what makes the understanding of a culture rich. Spradley proposes that there are a set of common semantic relationships that seem to occur over and over again:

1. Strict inclusion X is a kind of Y
2. Spatial X is a place in Y, X is a part of Y
3. Cause-effect X is a result of Y, X is a cause of Y
4. Rationale X is a reason for doing Y
5. Location for action X is a place for doing Y
6. Function X is used for Y
7. Means-end X is a way to do Y
8. Sequence X is a step (stage) in Y
9. Attribution X is an attribute (characteristic) of Y

Spradley proposes five kinds of structural questions designed to expose the semantic relationships of terms:

  1. Verification Questions – Asking for verification of a domain – or relationship between a set of terms
    1. Domain Verification Questions – Asking whether there are different kinds of a term that the informant has shared
    2. Included Term Verification Questions – Asking whether a term is in a relationship with another term
    3. Semantic Relationship Verification Questions – Asking whether there is a kind of term that relates other terms or if two terms would fit together in a sentence or relationship
    4. Native-Language Verification Questions – Asking whether the words spoken from the informant to the ethnographer are the words that would be used when speaking to a colleague
  2. Cover Term Questions – Asking if there are different types of a particular term
  3. Included Term Questions – Asking if a term or set of terms belong to another term
  4. Substitution Frame Questions – Asking if there are any alternative terms that could be used in the sentence that an informant has spoken
  5. Card Sorting Structural Questions – Asking informants to organize terms written on cards into categories and by relatedness. This is similar to an information architecture card sorting exercise. (See my post and video about Card Sorting for more.)

Descriptive questions will be interspersed with structural questions to prevent monotony and to allow the ethnographer to fill in gaps in their knowledge. Though structural questions help provide a framework to how terms relate, the relationship strength between terms isn’t always transparent. That’s why contrast questions are used to refine the understanding of what the strength of the relationship is between terms.

Contrast Questions

Sometimes you can’t see differences in the abstract. For instance, our brains automatically adapt to changing light and convert something that may look blueish or pinkish to white, because we know something (like paper) should be white, even when the current lighting makes it look abnormally blue or pink. So, too, can the hidden differences between terms be obscured until you put them right next to each other. That’s what contrast questions do. They put different terms side-by-side, so they can be easily compared.

The kinds of contrast questions are:

  1. Contrast Verification Questions – Asking to confirm or disconfirm a difference in terms
  2. Directed Contrast Questions – Asking about a known characteristic of a term and how other terms might contrast on that characteristic
  3. Dyadic Contrast Questions – Asking the informant to identify the differences between two terms
  4. Triadic Contrast Questions – Asking the informant to identify which one of three terms is least like the other two
  5. Contrast Set Sorting Questions – Asking the informant to contrast an entire set of terms at the same time
  6. Twenty Questions Game – The ethnographer selects a term from a set and the informant asks a set of yes/no questions of the ethnographer until they discover the term. This highlights the hidden ways that informants distinguish terms. (This is similar to techniques like Innovation Games, where the games are designed to reveal hidden meanings.)
  7. Rating Questions – Asking questions about the relative values placed on different terms – along dimensions like easiest/most difficult and least/most interesting, least/most desirable, etc.

The sheer number of types of questions can seem overwhelming at first. However, many of these forms flow automatically if you develop a genuine interest in the informant and their culture. Still, sometimes it’s hard to try to learn a new language and think about what’s the next question that you need to ask to keep the conversation moving.

Multiple Languages

In the case of an anthropologist who is working with a brand new culture, it could be that they’re learning a whole new language – literally. However, in most cases, it’s not that the language is completely different and new to the ethnographer. In most cases, it’s the use of the terms that are different. Just experiencing the difference between UK English and American English can leave someone a bit confused. A rubber in England is an eraser in the US, and a cigarette in the US is a fag in the UK. While both are English, the meaning and expectations of the word are quite different.

We often forget how we speak differently in a profession. A lexicon – special language – develops around industries that aren’t a part of the general consciousness. It’s the ethnographer’s job to discover not only that lexicon but also what the words mean to the rest of us.

Who Should Translate, and When?

When there are multiple languages, there is always the need to translate from one language to another. However, who does that translation – and when is the translation done? Informants, in their desire to be helpful, are likely to try to translate the information of their culture into terms that the ethnographer will understand. While the intent is helpful, the result is that the ethnographer doesn’t get to understand that aspect of the culture.

So, while translation is necessary, it’s best to continue to discourage the informant from being the one who is doing the translation. The ethnographer can leave their notes in native language and then translate later. This also allows them to validate information with structural and contrast questions. Sometimes, it’s this review that reveals some underlying themes of the culture.

Themes

In most cultures, there’s a set of recurring themes that appear. It isn’t explicit or stated, but there are those sacred cows that everyone worships that shapes the way the organization thinks. An entrepreneurial company has agility or velocity at the heart of the way that they organize their thoughts. A brand-focused company may be inherently focused on status or image. While these values aren’t typically articulated, they’re assumed, and they shape the way that the organization thinks – about everything.

By having the opportunity to review and rework translations, these themes begin to emerge. The semantic relationships appear over and over again until it becomes apparent that they’re not specific ways of organizing a topic but are instead a way of organizing everything.

Depth

One of the challenges that I often see in requirements is that the business analyst doesn’t always spend the time drilling into the details and verifying understanding in a way that results in requirements that fully express the needs of the business and how they do work. The ethnographic process – including the variety of questions – is one way to combat this challenge. It’s possible to leverage the ethnographic process to more deeply understand what is happening and how the systems are expected to help.

While I may be far from the fields of a foreign land, speaking to people whose language I don’t speak, I often move from industry to industry and company to company, learning their languages and the way that they think about the world. The Ethnographic Interview is, therefore, a useful tool for helping me get a better understanding and better requirements.

Book Review-Intertwingled: Information Changes Everything

It might seem odd that one of the forefathers of information architecture would proclaim that everything is intertwingled and thus hard to force down into specific categories, but that’s what Peter Morville is saying. You can’t separate the parts from the whole. Intertwingled: Information Changes Everything is the start of a journey in realizing that, no matter how hard you try to cut apart, partition, chunk, and dissect, most things can’t be broken down without losing something. That won’t stop us from doing it, because we need to do it to simplify our world. However, it may change the way that we approach the process.

Thinking in Systems

The vehicle in our journey is the ability to see everything as a system. There’s an old Tootsie Roll commercial where a jingle is played that ends with “Tootsie Rolls are all I see.” It’s a catchy tune that was designed to get people thinking about Tootsie Rolls, so they couldn’t forget about them. However, there’s some truth in life to this commercial. Once you see something, you can’t un-see it. It’s impossible to not know what a hippopotamus looks like after you’ve seen one. If you’ve learned to see things in systems, it’s hard to un-see them.

For Morville – and for me – everything is a system. It’s a system that can be manipulated, changed, and adapted often with unintended and inexplicable outcomes. Whether you realize that steel axe heads degenerated some aboriginal societies (as in Diffusion of Innovations), you got caught up in Peter Singe’s wildly popular The Fifth Discipline, or you stumbled across the posthumously published book Thinking in Systems by Donella Meadows, once you’ve been infected with the awareness that everything is a system of stocks and flows (inbound and outbound), it’s hard to not see it any longer.

However, just because you can see things in systems doesn’t mean you can predict outcomes. Gary Klein recognized that fire commanders build mental models of the fire where they could predict what is happening. Their Sources of Power may not have been consciously known to them, but they did know when one of their expectations generated by the model was violated, and it was time to go back to the drawing board and try to learn how the system was really functioning.

Fire commanders were strikingly good at predicting how the fires they were used to would behave. Their mental models worked. However, knowing how fires worked made them no better at predictions in other areas of their life. They didn’t even realize that they were seeing the fire as a system and simulating its facets – they were blind to the fact that they were even seeing the fire as a system. (Seeing What Others Don’t is another work by Klein which helps to explain how we develop the insights that fuel our mental models.)

Wireframes and Storyboards

Wireframes are a stock tool of a user experience designer and of the information architect. These wireframes form the boxes. They’re the pictures of what’s being built. However, what Morville and others have discovered is that it’s what happens between the boxes that is truly interesting. In Understanding Comics, Scott McCloud explains that the story in a comic book happens in the gutter between the frames. In effect, it’s the arrows between the boxes that gives the story its power. It’s what the user fills in themselves that helps the most.

Story is about connecting one frame with the next. That’s why storyboards are an even more powerful tool for designers than the venerable wire frame. Storyboards, which are often used in designing feature films (see Creativity, Inc.), connect the dots and reveal what’s hidden in plain sight in the way of the wireframes.

Rewinding to Agility

One of the criticisms laid at the steps of information architecture is that it reeks of “big design up front.” This is just the sort of waterfall-based logic that agile development teams launch scud missiles at. The problem is that too few practitioners of agile have read the Agile Manifesto or understand what information architecture is. They argue against documentation when the original founders of the agile movement were more concerned with ceremony and eliminating the waste caused by it. They didn’t want to eliminate documentation. They were focused on building and adapting instead of unnecessary documentation and unnecessary rigidity. However, there’s nothing in agile that argues against understanding what you’re doing. In fact, agile is focused on learning and understanding. The argument that agile levels against traditional waterfall (one-iteration) approaches is in believing in the planning fallacy – that you can plan through things ahead of time. (See How Will You Measure Your Life? for more.)

In fact, agile aligns perfectly with systems thinking, where you do something, observe the results, and then do something else. Sometimes you reverse (or try to reverse) the latest change. Sometimes you head off in a different direction. Ultimately, you’re always making small changes to see how things react, then adapting.

Information architecture – like software development – can take on characteristics of agile or traditional waterfall development. It all depends upon whose hands the tools are wielding the tools. There are some activities that require the development of non-intervention understanding. However, this is not always the case. Good information architecture recognizes that you won’t get it right the first time – or at one time. The needs of users to access content shift as attitudes and options change.

Memories

Our memories are fickle things. Our memories aren’t like video recorders accurately storing what happened. Instead, the memories become encoded and reduced into fragments and concepts that we can rearrange when we try to retrieve them. (See Science in Pseudo Science in Clinical Psychology for more on the fallacy of our memories.) Our memories don’t encode words. Instead, we encode concepts and ideas. When we retrieve the memory, we reconstruct the concepts using the words that we have in our current vocabulary.

If you don’t believe me, try to recall a conversation that you had when you were 8-10 years old. Recite the dialogue as you remember it into a recorder. Then go back and review the recording. You’re likely to find that your words weren’t words that you had in your vocabulary back then. Your memories were reconstructed with the help of your current vocabulary.

This simple trick reveals one of the ways that our brain tries to fool us. (For more on the way that our brain fools us see Incognito.)

Words

Even though we don’t remember words directly, words are deeply embedded into the way that we think. If you feed people negative experiences, you’ll get back negative words. One of the favorite parlor tricks of pop psychology is to listen for the Freudian slip. That is, when the truthful thing slips out from underneath the weight of polite society. Chris Argyris has an exercise of left and right columns, where the right column is what was actually said, and the left column is what was thought or felt. This exercise exposes how the words we use aren’t the words that we mean, and how these words can lead us down unproductive paths. (See Organizational Traps for more on the two-column method.)

Words are the way that we create the mental framing that we use for a problem. That’s why the words we use in our taxonomies are so important. They can conjure up the right ideas or ones that don’t connect with our audiences.

Maps and Territories

They’re called the badlands. Technically, this is a geological term referring to the erosion of clay-rich soil and softer sedentary rocks. However, the connotations of the term “bad” subtly influence millions of people to avoid these beautiful geological features. In this way, labeling an area as “badlands” on a map influences the visitors at Badlands National Park. The hidden connotations of the words we use shape how people see what we’re talking about, for better or for worse.

If I define a category for furniture, I’ll shape the way that people think about the category. They’ll picture in their mind what furniture means. They’ll see a chair, a couch, or a table. (At least 90% of the groups I do this exercise with say these three items.) If I’m looking for a rug, I won’t automatically look in furniture. Similarly, I’ll have a problem looking for a lamp. Is it in electrical or furniture? It has both properties.

So in one sense, a map isn’t a territory – but in another sense, it is, because it shapes that way that we think about the territory.

Multiplicity

There is an information architecture problem caused when the categorization labels make it difficult for consumers to decide which path they should follow. One approach to solving this is to use polyhierarchy – having one category included in multiple places in the hierarchy – said differently, having two parent categories for the same category.

While this is a necessary strategy for providing solutions to platypus-type problems (which don’t fit neatly in any one category), it can be an unnecessary crutch used because of poor category selection. Like other tools in the information architecture toolbox, it’s important to know when to use it – and when not to.

The best strategy for managing the problem of things that fall within multiple categories of an existing taxonomy is to use multiple taxonomies. That is, instead of creating one and only one hierarchy of terms, multiple taxonomies are used, each with its own set of terms that can be selected when appropriate – and omitted when not necessary.

Intertwingled

With all we know about how the mind works and how we categorize, we still don’t know enough, and we still can’t see enough. Breaking things down into their components so that we can learn about them is a good strategy. (See Efficiency in Learning and The Art of Explanation for more on learning approaches.) However, it’s not a good strategy if we fail to learn how those pieces fit together, how they work together, and how they break the rules when they’re connected to one another. That’s what makes things Intertwingled. They become that way because, when you put the pieces back together, you don’t always get exactly what you would expect. If you’re ready to start your own journey, perhaps it’s time to learn to see the world Intertwingled.

Book Review-How to Make Sense of Any Mess

When I explain my passion for information architecture to folks, they often wonder what I’m talking about. They understand intuitively that I’m not talking about designing buildings, but how can you design information? A better way of explaining information architecture is to say it is How to Make Sense of Any Mess. Abby Covert nailed the colloquial definition. In her book, she takes a practical tone to an often academic topic and explains how to make sense of our messes.

Fuzzy Lines of Information and Knowledge

The first thing that we’ve got to get out of the way is that all information is relative. What we know is true for our beliefs and our circumstances but not necessarily true to everyone or in all circumstances. The depths of the problem aren’t new to me. I’ve spent time exploring choice theory, where explaining your choices makes you less likely to like them. (See The Paradox of Choice.) I’ve peered under the covers of knowledge management to realize that not everything we know can be explained explicitly. Somethings are tacit – we just know them, and we can’t take them out of context. (See Lost Knowledge.)

I’ve addressed the fuzziness with which our words convey our meaning. Though we have powerful mind-reading skills that allow us to work together, these skills are not perfect (see Mindreading). A single word can have opposite meanings. Words with multiple, different meanings are called homographs. They’re particularly insidious to communication when they have opposite meanings. Consider dust – which can either refer to the act of removing dust or the dust itself. Consider “weather,” which can either mean to withstand a storm or to be worn away.

I know, too, that the way we arrange options can make us what Nudge calls “choice architects.” We can shape the choices that people take by their arrangement. However, the rules for this rely upon the murky depths of subtle cues that we use to make or decisions without knowing we’re making decisions.

To What End

The ultimate question for any information architecture effort is “To what end?” That is, who is going to use the information you’re organizing? What are their goals? What is it that you would like them to learn – without knowing it? By answering these questions, you can use the tools you have to create organizations that make it easier for users to find the information they’re looking for.

However, the current users aren’t the only people to be considered when designing the information architecture. Future potential users count, too – particularly when the existing user base is small and you’re trying to make it larger. So, too, do those people who are involved in the outcome of the structure that you’re using – like investors or managers of the departments.

Make It Visual

Organizing information isn’t an easy proposition. It’s a process designed to reduce the cognitive load necessary to learn. (See Efficiency in Learning for more on cognitive load.) One might argue that, if we’re speaking about organizing and making sense of messes, what does learning have to do with this seemingly unrelated topic? The answer is that we’re learning to navigate the information jungle to find what we’re looking for. We’re learning how to organize our thinking into more effective patterns. Thus, we’re always learning. Marcia Bates estimates that we absorb 80% of what we know not through formal learning. We just “get” it by experiencing life.

Our brains are inherently visual. We’re not wired to process information in the collection of ordered symbols called letters into words, sentences, and paragraphs. While we adapt in the direction of being able to process this information, we still make sense of pictures and diagrams easier. That’s why “a picture is worth a thousand words.”

Abby Covert walks us through several diagram types including block diagrams, flow diagrams, Gantt charts, quadrant diagrams, Venn diagrams, swim lanes, hierarchies, mind maps, schematics, and journey maps. Each of these visualizations has different benefits. In software requirements gathering, I use these types of diagrams as well as entity relationship diagrams, activity diagrams, state diagrams, data flow diagrams, and ecosystem maps. The idea is that each type of visualization can quickly relate an aspect of the relationship between entities in ways that the others – and certainly words – cannot.

Technically Right, Practically Useless

In designing a taxonomy, there are sometimes some hard choices to be made. In some cases, the users routinely categorize something “incorrectly.” While there’s no right or wrong way to organize things, once you’ve accepted a taxonomy, every item should have its place. A classic example is the tomato. Technically, it’s a fruit, but it’s important to know not to put it in a fruit salad. Most folks recognize that it has vegetable-type qualities when used in cooking. When classifying products in a grocery store, where would you put the tomato? In the technically correct location – where few people would find it initially – or in the incorrect location where it’s findable?

The answer depends on your goals. In the case of a grocery store, the answer may be that you file it with vegetables, because cashiers and shoppers who use self-serve lines expect to find it there. If you’re a college library on botany ,where would you place it, then? Because there’s a different objective, teaching the “correct” locations, you might be inclined to place it in the category of fruit. Of course, people still won’t be able to find it, but then they can be admonished by a professor for not realizing that the tomato is a fruit and not a vegetable.

Ambiguity and Exactitude

On the surface, exactitude would be our goal in organizing any mess. After all, if everything has a place, and everything is in its place, then all is right with the universe. However, exactitude costs us flexibility. We can’t put cooking scissors in a drawer labeled spices and expect anyone will find them. Spices is a specific and –relatively speaking – exact label. It means that we don’t have the flexibility to add scissors into the same container. Of course, the ambiguity of “kitchen stuff” is probably broad.

“Kitchen stuff” could be almost anything even – quite literally – the kitchen sink. Cooking stuff may be an alternate category label that restricts the contents some, but not by much. The flexibility offered by ambiguity increases the level of abstraction and reduces navigability.

There’s no answer to how to create the right balance of being ambiguous enough to be accommodating to unexpected events and at the same time exact enough to be clear and valuable. When you err in one direction, you move away from the other.

Everything is a Mess

In the context of How to Make Sense of Any Mess, everything is a mess. At some level, something designed by someone else won’t make perfect sense for us. Things designed by others will at least seem messy to us. Even things that seem like they’re organized today won’t feel organized tomorrow. We don’t have to feel overwhelmed by the mess. Bit by bit, we can make sense of our messes – and try to move things forward. Even if we don’t expect to ever learn how to make sense of our messes, it’s worth learning How to Make Sense of Any Mess.

Book Review-The Accidental Taxonomist

There are dozens of things that I do each day that I didn’t set out to do. I do accounting and billing work without a desire or intent to do it. I do sales and marketing – and neither are at the top of my list of things to do. I accidentally picked these things up when I decided to be an entrepreneur and run my own company well over a decade ago. Working with taxonomies – and becoming a taxonomist – can happen by accident too. That’s why The Accidental Taxonomist is appropriate for someone looking to learn how to create taxonomies. I’ve never heard a child say, “I want to grow up to be a taxonomist.” Despite this, there are those who have taxonomy as a part of their job – whether they intended it to be or not.

Long, Long Road

Before I get to the heart of the matter, it’s appropriate to tell you that I didn’t read the book in one sitting. I didn’t read it in a week, a month, or even a year. I started the process of reading The Accidental Taxonomist about half a dozen years ago. It was as I was putting the final touches on my Pluralsight course The Art and Practice of Information Architecture. I got the course done and never finished the book.

In just getting back to it, I felt a bit like some of my clients that struggle to get their taxonomy projects off the ground. Or, rather, my clients that needed to get something accomplished and realized they needed a taxonomy to accomplish their goals. The taxonomy was started, the goals were achieved, and the taxonomy sat aside for a while – sometimes a long while. Before we get too far, we should explain what a taxonomy is.

What is a Taxonomy, Anyway?

Barry Swartz in The Paradox of Choice explains that filtering is one of the basic functions of consciousness. What he doesn’t cover is that so is organization. We’re hardwired to make sense out of our world, and, as Gary Klein explains in Sources of Power, that comes through simplification until we have a model that we can run in our heads. Taxonomies allow us to organize our thoughts and information.

We’re all familiar – willingly or not – with the hierarchical biological taxonomy of zoology. That is to say that we learned Carl Linnaeus’ organization of all animals. We learned Kingdom, Phylum, Class, Order, Family, Genus, and Species as a way for differentiating one animal from another and identifying their nearest cousins.

We also learned, but most of us promptly forgot, how Melvin Dewey organized his library. The system of organization held a brilliant discovery for extension. He figured out that he could make his system flexible and allow for increasing levels of detail through the use of a numbering system.

We probably never learned about S.R. Ranganathan’s different approach to classification. He was frustrated that things could only be placed in one spot. There was in effect one “right” way to find things. His insight was to introduce facets. Instead of trying to capture the uniqueness of any given item in a single hierarchical dimension, he proposed that items be classified in several different categories, or facets, and the combination of these facets would be how the item was classified. This approach was called colon classification, because he chose the colon to separate the various facets.

I include Ranganathan’s system to point out that taxonomies are about organization. They’re not about hierarchy, though that is often the way they’re executed. They’re not about books or animals. Taxonomies are, at their core, about how we make sense of this world that is far too complex for our minds to process.

What’s a Thesaurus?

I remember first “discovering” the thesaurus in grade school. You could make your writing sound more impressive by looking up words that no one knew. I could take a simple, common, everyday word and replace it with something more profound and meaningful. (Perhaps I even looked up the word profound.) To me at the time, thesaurus only meant synonyms. I could find words with similar definitions. Eventually, I found the antonyms. However, for the better part of 30 years, that’s all they were.

When I started diving into information architecture and how we organize information in ways designed to make them easier to access, I realized that my old friend the thesaurus was more powerful than I had given her credit for. More than just synonyms and antonyms, the thesaurus contained the relationships between words. Where a dictionary can tell you what meanings are associated with a word, it’s the thesaurus that can put the word on the map in relationship to other words.

Understanding which words had broader and narrower meanings allows you to respond with precision words that encapsulate the exact scope that you wish to cover. There can be alternative spellings to help you understand how there might be multiple different ways to spell a word – such as color and colour. The thesaurus had more to offer than I had anticipated.

What’s an Organization’s Thesaurus?

The role of a thesaurus in an organization is even more powerful. Inside the context of an organization, a thesaurus can identify preferred terms over terms that are less preferred. They can share common misspellings. They can define terms across languages. They can translate the scientific to the everyday – and vice versa.

An organization’s taxonomy provides a map of the terms that are used in the organization and notes about how those terms are used – or are intended to be used. They provide the basic relationships between the terms. When the relationships get more complex, then we’ve moved from the world of thesaurus into the world of ontology.

Ontology’s Relationships

In a thesaurus, the focus is on words. They make up the tent poles on which the relationships are hung. However, ontologies focus much more on the relationships between words and the nuances of these relationships than the words themselves. Instead of being focused on the tent, ontologies are focused on the net that keeps circus workers safe. It’s not the individual ropes – or words – that keep performers safe. It’s the relationship and connection between the words that keep performers from falling.

Ontologies are a way of understanding a field of study or knowledge. Ontologies provide a rich map of how things in the field are connected to one another. The relationships are richer than simply one term being broader or narrower than another.

What is a Taxonomist?

If an organization organizes their content through a taxonomy in the form of a thesaurus and a set of ontologies, why do we call the role a taxonomist? At the root, it’s the development of an organizational structure – irrespective of which tools are used – that defines the core behaviors of a taxonomist. Their role is to organize and make easier to access information. The tools they use are just the tools of the trade.

The funny thing is that many taxonomists – but not all – aren’t in full-time roles. Few taxonomists have it in their title, though some have it in their job description. It’s more common to have taxonomy development as a prerequisite for something the role requires, so often the taxonomist isn’t a person who spends all day organizing structures. Most of the time, the taxonomist is someone who has a job to do that is made better by taxonomic development.

Special Skills

If categorization and organization is a part of the basic functioning of consciousness, then shouldn’t everyone be considered a taxonomist? At some level, yes. However, what differentiates every man from the taxonomist is in the tools that they’ve developed for clarifying, codifying, and communicating what the structure of organization is. By learning what humanity knows about psychology, neurology, and the organization of large information, taxonomists can distinguish their capabilities.

While these aren’t likely enough for a taxonomist to feel truly confident in every situation, this knowledge and these skills are useful.

Taxonomy Purpose

A taxonomy’s purpose is to help organize content, that’s easy. However, taxonomies provide structure and framing that shapes the way that people think. As a result, taxonomies are more than just a way to browse to the information you want. Taxonomies can be helpful in shifting the way the organization works.

Sometimes this is through the inclusion of detailed terms in a hierarchy to encourage users to be more specific about what they mean. Other times, it might be through the use of preferred terms. Preferred terms in the taxonomy can shift the thinking from package delivery to package assurance. It’s a subtle shift that focuses the corporate consciousness on assuring shippers and recipients that their package will make it to their destination.

Polyhierarchy

Atoms have a challenge that they can only exist in one place at a time. However, in our electronic taxonomies, we can put things into more than one place in the taxonomic tree. Consider a color taxonomy that starts with a level of red, blue, and green. Where does the second-level color blue-green belong? Blue sure, but green as well. This is a polyhierarchy, where an item has multiple parents. While logically this seems like the exception, polyhierarchies are more common than most would like to admit.

The truth is that taxonomy projects are messy. It’s only a matter of time before you’re going to run across the digital equivalent of a platypus. The platypus has a mixture of reptilian, avian, and mammalian genes. It’s a classic challenge for the zoological taxonomy that splits reptiles, birds, and mammals all the way at the top. With a polyhierarchy, the platypus can find its place in all three taxonomies.

Tagging

Being a taxonomist solves only one part of the puzzle. Taxonomists create the structures, but it’s often up to others to tag the content to fit into the taxonomy. This split means that, in many cases, the taxonomist must make a point to sit with those who are actually doing the categorization to understand what is and isn’t working. Similarly, they should sit with users who are actually trying to find the information.

The key challenge in taxonomic development isn’t in designing the taxonomy. The key challenge is getting the users – who are often not dedicated indexers – to enter the metadata necessary to make the taxonomy work. Too many taxonomy projects are abandoned before the work really gets started, because the people indexing the content refuse to do it.

Pre and Post Coordination

There are tricks that can be used to improve results. Search can aggregate terms by leveraging synonyms even if the users aren’t always using the preferred term. Facets can go a long way to simplifying the search process, and full-text indexing makes some level of taxonomic identification unnecessary. Automatic classifiers – whether rules or machine learning-based – can help the content get the correct metadata with minimal help from the indexers.

With all this mess, it’s hard to keep track of when the metadata is known and to judge its reliability. Whether it’s entered at or near the time of creation in the form of pre-coordination or it’s managed through the searching process, getting it right is hard. Maybe you find that you’re not getting the findability that you want, so to fix the problem, you’re going to become The Accidental Taxonomist. Perhaps a quick read can give you tips that will make the process easier and less painful.