You know how it feels like when you do not find something on a website? In the library? In your room? Well, it is all due to infirmation architecture!
We cover this topic in this episode. This episode is part of a 3-episode serie on information architecture, taken from my masterclass that I gave recently.
More specifically, this episode will cover:
You know how it feels like when you do not find something on a website? In the library? In your room? Well, it is all due to infirmation architecture!
We cover this topic in this episode. This episode is part of a 3-episode serie on information architecture, taken from my masterclass that I gave recently.
More specifically, this episode will cover:
Hello everyone, jeremy here and I'm very pleased to go on with this second episode of this series dedicated to information architecture. So, as a reminder in the first episode, if you haven't listened to it, I highly encourage you to do so, because it lays down the basics of information architecture and it will also help you understand why we are tackling the topics that I will tackle now in this second episode, which is dedicated to UX research methods for information architecture. But, just as a reminder, in the last episode we talked about, well, we introduced the topic of information architecture and we began with a definition of information architecture, saying it's the practice of organizing and labeling content to ensure it's logical and findable, with the idea that it's really about organizing the content, yes, in a way that makes sense. Yes, but for your users in the end. So it might make sense for you, but not necessarily for your users. So, and that makes the transition with this episode today about doing user research. Then we said that information architecture is made out of content, normal clientele and structure. We said that it's super important to consider information architecture at the very beginning of a UX design project, because it lays down the foundations of a logical, findable and logical app that makes sense for your users, and so that is. It is the key to provide an intuitive user experience, and also because it's present everywhere. So I gave some examples, for instance, from the e-commerce realm, but you have examples everywhere, so, for instance, in the grocery store, at the museum and so on. I also shared how we can help you at different levels and depending on where you are focusing right now. So, for instance, minimizing customer dropout can also help improving a website's reputation, seo ranking, reducing cost of support and so on.
Speaker 1:I also touched upon how to represent information architecture. So in general, it's done through sitemaps and or through spreadsheets, so you can use it more internally in a company to share about the internal structure of your information architecture and to make edits as a working document. But the sitemaps are more, let's say, visual representation and it's easier to communicate, to use sitemaps to communicate in the team. So main differences is as a sitemap, as I was saying, is more visual, but also it's readable vertically from top to bottom, with boxes, and each box is then divided into sub boxes and each one is then subdivided again and again until you arrive at the end of this architecture. So visually it looks more like a waterfall, although it's called a tree. But well, maybe it helps for you to imagine it. Or maybe you already saw some pictures on the Internet and I will also again embed the presentation linked to this topic so that you can read it. And then you have also the spreadsheet. So it's the same thing but laid out in a horizontal format, let's say readable from left to right Without the left the main, the first nodes, let's say in the hierarchy, the higher nodes of the hierarchy.
Speaker 1:So, for instance, you have the homepage, first column, then second column, you have the first categories, and then, as you progress to the right, you are going finer and finer in detail and in subdivisions and finally we ended up defining how you can define an information architecture, how you can design an information architecture, by first mentioning some examples of bad practices around us. So, for instance, no clear, no structure or unclear structure, then a lack of consistency regarding labels, and so the other extreme of having unclear structure is wanting to have too much structure. In the end, all of that in general happens because there is little or less little or non sorry UX research conducted. So in the end I was saying and this is the perfect transition for today, to avoid these kind of issues. Well, I will first give you the piece of advice of listing the product content that you have all of the features, all of the labels, kind of listed on a sheet or no working document on a mirror, wherever you want, but list it. Then run a competitive audit to see how your competitors are sorting out all this content. And finally, so these two first steps can be done without any research with end users or end customers, then, as I was saying in the last episode, this will ensure that the very minimum is done, but it doesn't ensure success in terms of having a usable and intuitive app or website or service, because you would need, in the end, to make sure that all this layout makes sense, all this architecture makes sense for your users, and the only way to do so is to run user research, and that is perfect, because that is what we are about to see today.
Speaker 1:So today, in this episode, we will focus on research methods for information architecture, and we'll begin by mentioning very briefly that there are mainly two research methods to study information architecture. You can define, redesign, improve information architecture. In my opinion, it's not the only two methods, they are the main methods, but you can use other techniques and gather data from other sources, with users, to improve your information architecture. So, for instance, you're running a usability test and one of your insights is that my user is not finding easily the my account section. Or I don't know, I'm making a financial app and my user is not. My customer is not finding easily the budget section and you did not run. Imagine one of the two methods I'm going to present, but it's one of the insights of your usability study, for instance. Well, even if you did not run the perfect method for this, it's still an insight that you can use to, of course, iterate on the information architecture. So I just want to make clear it's not clear cut in general the methods we can use in UX research, although the two methods I'm going to present are the main ones, because they will help you really gather the information. That is very specific for information architecture.
Speaker 1:So, without further ado, let's begin with card sorting. So card sorting is the method of choice when you want to design an information architecture, let's say from scratch. So let's say it's a more exploratory method. I will develop then the different types of card sorting and how you can use them depending on the stage you're in in your user research project or your product development lifecycle. But let's to begin with, let's just see card sorting as a whole and come passing all types of card sorting. So what is card sorting?
Speaker 1:Card sorting is a method that is used to define or redefine an information architecture that makes sense to the user. So it's really about when you want to define something, and, in this case, information architecture. How does that work? Well, you have your participants categorize cards according to what makes sense to them. It's as simple as that. You ask them to cluster cards and we will see how, but you ask them to cluster the cards according to what makes sense. Sometimes you can say things like classify or sort cards according to what goes well together and what doesn't.
Speaker 1:You don't put it together Something like this, but this is already a bit more specific. You can just say sort the cards into piles according to what makes sense for you. Just that, why?
Speaker 1:Why would you use card sorting to design or redesign an information architecture? Well, it can help you answer several types of questions. The first one is where do users expect my content to be found? Basically, imagine you're again designing an app to help users control their finances. You can have several features. You can have the budget feature, you can have the tracking expenses feature, you can have some pieces of advice to manage your expenses, you can have some investment advice, and so on and so on. So you have all these features laid out in a spreadsheet, for instance as I was saying, referring to the first episode Spreadsheet, or a whiteboarding tool such as middle and so on, and you're like okay, now how should I sort my content? Well, you should sort your content in a way that your users would expect it to be found, and so this is exactly what card sorting does. It will help you organizing it by asking your users to sort it themselves. Then, okay, they sort the content.
Speaker 1:Okay, but, as I was saying, as I was referring to in the first episode, it's not because your content is nicely laid out that your users will find it, because there is another layer to consider, which is the label, the namings. Even if you sort the categories well, if you place them well in your website, you know what they refer to, you know that by clicking on that category, you will access that feature. Okay, but your users may not. Maybe the naming doesn't make sense. Maybe they don't know that by clicking on budget then they will have a tool which is called the tracking expense tool expense tracking tool sorry, behind budget, maybe it doesn't make sense. So that is why the naming is also important, and so card sorting also helps you for that.
Speaker 1:What names should be given to your categories? Because one aspect of the card sorting includes asking your participants to sort content and label the content they sorted, label the categories they sorted and, of course, what categories should go where in a site hierarchy. So the first step again you have your content. People sort it into categories. Where do users expect the content to be found? This is the first answer. Then you ask them to give labels to the categories. What names should be given to the categories? Second question answered, and then the categories themselves. You can ask some follow-up questions to your users to ask them how they would like to access it. How should it be? Between the categories themselves. They can also continue the sorting. So basically that. So this is card sorting 101, let's say Now, if we dive deeper into card sorting, how does that work in practice when you are running the activity?
Speaker 1:Well, it's really simple. So let's take the example of open card sorting. So I don't want to over complicate things here, but just know that there are several types of card sorting, and we'll begin with the open card sorting, which is the, let's say, the most common one when we think about card sorting. And so then we will dive into the other types of card sorting. Okay, but for the open card sorting, how does that work? Well, basically, you include this methodology in any user session. So and again we will take the example of a moderated one so you are, for instance, running an interview with a user directly through Zoom or in person. Okay, let's take that example.
Speaker 1:And you, during this interview, you decide to include some card sorting activity. How does that work? Well, how does that work? You grade the participants, you thank them for being here again, and then you begin the activity. So the activity is really really simple to teach to the participants. So, basically, you tell them something like here you have some cards laid out and I will ask you to sort them into piles according to what makes sense for you, and the best is, if you are in a moderated exercise, you just say that to them. You don't talk about the labels yet. So you say just sort the content into categories that make sense for you and then you add some explanations, of course. So, for instance, you can sort the cards into as many categories as you want and you can have as many cards as you want into each category.
Speaker 1:There is no right or wrong answer. There is no judgment here. It's really about your way of seeing things. That is interesting for us. So this is the first aspect, and then you can also tell them sorry if a card doesn't make sense or you don't understand it.
Speaker 1:So there are two things. The first thing is that they can ask you, but probably you won't want to answer because then all participants will not have the same level of explanation, but it's a possibility. If you see that it's really tricky, you may be able to give some definition or so, but usually that's not the case. Usually you will answer them. Second option, which would be what you don't understand, you can leave it off the side what is not clear for you, because it's also an insight. Anything that is not clear isn't an insight. Basically, that's what you want to test If your label is clear and where does it belong. So if you begin to answer something to your participants, depending on what they don't understand, you will bias your results. So I'm really not a fan of explaining.
Speaker 1:Just say to the participant put into a category that usually is called I'm not sure. So these are for the explanations. Once it's done, you can then ask the participant okay, now I will ask you the groups that you did, to label them, to give a name to the groups, and sometimes that's super interesting, because this happened to me Sometimes participants are super a surprise when you ask them that they're like oh, I don't know, they feel pressured. So it's really important to reassure them here and to say, okay, you're not being tested, it's not a test of your vocabulary, of your knowledge or whatever. It can be anything. It can be a word, it can be a sentence, it can be I don't know a scene that you have in mind, like describe a picture, describe something that comes to your mind. It can be anything, just to refer to this category and anything will be of use. That is the goal of open car sorting.
Speaker 1:And then what's super important is after this activity, and that is why I took the example of the moderated activity. Moderated means you will be here to moderate, you will be a facilitator and you will be present to interact with the user while they are doing the activity, and so the goal of being moderated moderated, sorry is that you can, of course, ask questions, interact with the participant, so you use that method when you want to go deeper into the activity and when you want to have more qualitative insights, and so, at this step, you just have to take advantage of this by simply asking the participant, following up questions such as okay, what did you think about this activity? How did you feel? This is very broad, but I like to ask this question this way because then I will not I will not bias or direct too much to the participant, but of course, some of you may prefer a more directive approach if you have fewer time, which would be something like how do you feel about the sorting that you made? How do you feel about the groups that you made, or can you please explain me the rationale behind the groups you made, and so you will have more explanation? Of course, I didn't mention, but also a great part of car sorting is that you can ask your participants to think aloud while they're doing the activity. So you may already have the answers to your questions, or you may already have some qualitative insights while they're doing the activity. So that's why it's so important to take notes while they're doing it and so you can ask them that how do you feel?
Speaker 1:And then there are some follow up questions, such as were there any cards that were easy to sort or difficult to sort? I like to ask it this way easy or difficult? Because then it's less biased. If you focus only on the difficult, well, it will make your participant focus on half of the story and maybe some of them are super easy, and it's also an insight. It's not because things are easy that it's not interesting. Maybe having some categories that are easy to sort. If your participant knows about it, you will know that this is something that maybe you should capitalize on on your website so that it's accessible, directly visible, because, well, the labels are easy, the categories are easy, you know. So capitalize on that, maybe. And what is difficult? It's also something you have to work on, of course. So if they say, oh, this card was a bit more difficult because the label I was not sure what it referred to or these two cards, they seem really close together, but I don't know. It seems that the topic is quite different. So I was hesitating. Whatever I'm making this up, of course, but you can feel how the type of answer you could get, and so this is the second type of question.
Speaker 1:And lastly, of course, when you're in an open card sorting, usually you have a category which is I'm not sure or I don't know, and you can ask your participants to put cards into this one when they don't know what they refer to, when they don't know how to sort it, and so you can ask them can you tell me a bit more about the I don't know category? Can you tell me a bit more about the card that you left and sort it? And this is super important, because if you have an information architecture with all your features let out and let's say, 30% of them are I'm not sure cards what can you do about it? If you don't ask your participants, you will never know. You will know that they put this content on the side, but you will never know. And so that is why it's important to run open card sorting and moderated card sorting first.
Speaker 1:In my opinion, at the beginning of a project and by project here I mean from scratch, developing an information architecture from scratch, not every project when you will enter. You will enter with zero information architecture done. Sometimes you enter a project and there is already a menu. There is already a hierarchy somewhere. This is not this that I'm talking about. I'm talking about starting a project from scratch. From a product design perspective, you don't have any information architecture. So that is why it's super important to do this kind of methodology. And then I will also speak about other types of card sorting. And so, just to sum up, about how to run a card sort.
Speaker 1:Of course, you will begin by choosing a set of topics which refer to the different features of your website, of the content of your website or app or service. Usually we give the advice of having between 40 and 80 items, not more, because then it begins to be a lot for participants to sort out. So basically that. Then of course, it might be a trivial, but avoid topics containing the same words. So here it can be important to mention that.
Speaker 1:Imagine you have a website I don't know I'm making this up, maybe it's a bad example, but that's what comes to my mind right now. You are having any commerce website and you have some products that are really similar. You have imagine one phone that is in blue and one phone that is in red. I don't know. I don't know, and you don't have in your information architecture a way to put it into colors. I don't know. I'm really making this up. So you have the same name. What changes? Is the color? Naturally, if that is interest to you and you want to know if your participants sort them into a group and then they give the label color, okay, that's okay. But if not, maybe it doesn't make sense to have twice the cards with the same name, because you know that your participants will put them together. The example is not the best, but you can see what I'm getting at. So avoid topics containing the same words. And then you might ask okay, but then if I want to cross reference my topics in the website, so maybe I want to access I don't know I want to access swimsuits into the water sports category but also into the women category, because these are women swimsuits, imagine. Yeah, okay, but then you can ask the participant that you can say to the participant that one card can be duplicated and put into several categories. This is something I didn't mention, but this is super important.
Speaker 1:Cards can be duplicated if it makes sense to your participants. Usually they will do it for some cars because you say it at the beginning and then they will forget. You can remind them several times. You can say don't forget. You can duplicate cards. So you choose a set of topics. You avoid topics containing the same words, of course. Then you ask the participant to sort them into piles, mentioning of course and the piles can be of various sorts, of various sizes that they can change their mind as they work the participants that they can leave items unsorted. Then you ask the participant to name groups and you debrief with the participants with questions such as were any items especially easy or difficult to place? Did any items seem to belong to in two or more groups? This is super useful as a question. And what thoughts do you have on the items left unsorted? And then you can repeat with 15 to 20 users. So, as a rule of thumb, I let run my car sort with at least 15 users, and this is for moderated carsort. If you're doing unmoderated, we will cover this topic later. It's good to increase, crank up those numbers as much as possible.
Speaker 1:Now about the analysis. What kinds of analysis can we do, can we have for carsort? Well, it depends on the type of car sort that you did, of course. So again, we will touch upon I'll touch on the types of carsort later. But imagine you have several types of carsort open, closed, hybrid and so on and all of them are done differently, are conducted depending on if you are unsure, depending on your level of uncertainty regarding the information architecture, depending on your maturity of your information architecture, depending on the insights that you already have. So the less you are sure of your information architecture, the more you will go towards expiratory analysis. Of course this is basics of data analysis and basics of user research. The more you are unsure, let's say, the less you're sure, the more you will go to exploration. Of course you want to know what you should know, and then, the more you are sure, the more you will go with quantitative numbers, metrics, to validate statistically.
Speaker 1:Okay, so for the analysis of carsort, you have two types of analysis. The first part is the expiratory analysis. So this is more for a moderated carsorting. It doesn't mean it's not for a non-moderated, but it's more the case when you have qualitative insights. And this entails looking at the big picture, basically from a very subjective and, let's say, qualitative mindset, and so you will look at all your piles, all your cars, and you will make conclusions based on that. So it's really subjective. But at the same time, if you did your interviews well and if you can identify a pattern, it's not that subjective. So it's something that several researchers could agree on. So I will now mention some types of information that you can have and what is the type of insight you can have. So it's kind of a reading grid, okay.
Speaker 1:So imagine you have cars that are grouped into multiple categories, consistently what you can say, for instance, this card is too broad and should be broken down into several cars Because one user is putting it category one, one user into category two, one user into category three, and this is consistent. So this card will be too broad and so it should be broken down into several cars. Sorry, correcting a mistake here. When I say broken down into multiple categories, it's more consistently users divided into several categories. Sorry for the mistake. So this is the first type of insight. Then you have one category which is consistently unused. So this is more for closed car sorting. We will see that after. So, closed car sorting the difference basically is that you already put the categories to your participant, you already lay out all the categories and you ask them to put the cars into these categories that you already pre-formatted. If you see that a category is unused, it means it's useless, so you should take it out.
Speaker 1:You can also have cars that are grouped together but in different categories depending on the participant. So here it means that the participants agree more on the fact that the cards should be grouped together than on the fact that they should belong to a certain category. So here you should maybe work on the naming. You should maybe work on how should I display this information, because really it seems that it's not one category per se, but more the fact that they should belong together. You also can have cards that are left and grouped consistently, and so here it's important to have some follow-up. Again, asking participants can you tell me more about why did you leave this card unsorted? And so you also might need to reconsider these cards, because if they are consistently unsorted and participants tell you the same thing, like I'm not seeing the use of that or whatever, maybe you can take that off Again.
Speaker 1:Also take advantage and capitalize on what looks the most easy, the easiest, sorry, so you can have, for instance, the most popular categories, so something that comes really often and participants all say it's super easy to sort them because it makes a lot of sense. Well, take advantage of that and think about a place where you should put it so that it's most easily seen, capitalize on that, and so you can also have categories, meaning the same for your participants, but given different labels, so you can have, for instance, cities. So imagine you're sorting things from a travel website and you can have cities and destinations. So maybe in that case, if you see the cities Amsterdam, london, berlin sorted into these two different categories, maybe it makes sense to have one label because the groups are the same, the cards making up each group are the same, but the names are different, so maybe it makes sense to put it together. So this is really the first type of analysis, which is exploratory and so making sense of the big picture, before going deeper and by deeper I mean a more quantitative analysis, if you have the opportunity. So for that again, you need to have at least really it's a very bare minimum 20 participants, but I would say more is better, like 30.
Speaker 1:And so the type of analysis we have in this case is quantitative analysis, and so it's really about numbers. So one way to represent that is called the similarity matrix, matrix of similarity, and so basically it's a matrix and it's well, it's a half a matrix, let's say. So it's kind of a square. You divide it by the diagonal and then you lay out in the diagonal all the namings of the cards, and then you make the cards correspond one to another. So you read it at the intersection of rows and columns. So you have column A, row number three. You read at the intersection, and this what you read at the intersection is a percentage and it's the percentage of people who sorted the cards together. Basically, that the higher the number, the more participants put the cards together, so the more it makes sense to have the cards together. In your website it's basically that, and so this is only that, only a measure of distance between the groups. And then on top of that you have a second type of representation I'm not sure about the statistics here, I don't remember, to be honest, but it's a different way of visualizing it which is called the dendrogram. And so on top of that you have a dendrogram.
Speaker 1:It's something that is read from left to right or well horizontally, let's say not in a matrix, and so you have all your labels on the left and then you have kind of a tree. Basically, so on the left you have the most divided, the well let's say the groups are the most divided the cards. And then the more you go on the right, the more they are grouped together until arriving at one category only that, let's say, rules them all. How you should read it is that, basically, the more you go to the left, the more it's divided. The more you go to the right, the more it's grouped. But then you're like okay, where should be the cutoff? The cutoff is very simple, it's visual.
Speaker 1:So there are two elements that will help you. The first one is the level of agreement. The more you go to the left, the more in general. So you have 100% of agreement. So in general, all participants will agree that all the cards are separated. So it's like it's like. It's like because if you put all the data together, no participant in general will have the exact same categorization. So that is why the scenario in which all the cards are all taken apart, everyone agrees on that basically. So you have this first aspect and then the more you go on the right, like all cards grouped together.
Speaker 1:Well, it's difficult to have 100% participants agree that all the cards should be grouped together. So in general it's zero on the right and then in the middle, well, you have this differential in percentage and then so you can see that, the level of agreement, and then, based on that, you also have visually the ramifications that will begin to form. You will have this shape and you will be able to tell visually when you see that there is a bump. There is a bump in the ramification and visually you can say, okay, here, if I put my card off, here I have three categories and I have, let's say, 60% of agreement and that's okay for me. So it's really about dodging, and there is no perfect science. Here is what makes sense to you. Again, an information architecture should not only be made only on user research, it's also on the business goals of your company and also on the features of what makes sense to your users. So there are different levels and that will be all for the episode of today.
Speaker 1:We focus today on the open card sorts and on methods of analysis for card sorts. We will tackle in the next episode alternative methodologies for card sorts and tree testing. And, as I was saying, that is all. I'm super happy to have made this episode. Thanks again for listening to me. If you want to know more about this, I'll encourage you to go on the website. That will be indicated in the show notes. You will be able to access this presentation and you will be able to leave a comment. Again, if you like this show, please subscribe. Share it with a friend. I will be very happy, and if you want to get in touch, please do so on the website. Thanks again for listening and see you in the next episode. Bye-bye.