Online card sorting: The comprehensive guide

You’ve just finished running your card sort. The study has closed and the data is waiting to be analyzed. It’s time to take a look at the analysis side of card sorting, specifically in our tool OptimalSort. Let’s get started.

‍

A note on analysis 📌

‍

When it comes to analysis, there are essentially two types. There’s exploratory analysis (when you look through data to get impressions, pull out useful ideas and be creative) and statistical analysis (which really just comes down to the numbers). These two types of analysis also go by qualitative and quantitative, respectively.

‍

You’re able to get fantastic insights from both forms.

‍

“Remember that you are the one who is doing the thinking, not the technique… you are the one who puts it all together into a great solution. Follow your instincts, take some risks, and try new approaches.” Donna Spencer, Maadmob.

‍

Getting started with analysis 🏁

‍

Whenever you wrap up a study using our card sorting tool, you’ll want to kick off your analysis by heading to the Results Overview section. It’s here that you’ll be able to see how many people actually took part in the study, the average time taken and general statistics about the study itself.

‍

This is useful data to include in presentations to interested stakeholders, just to give them a more holistic view of your research.

‍

Digging into your participant data ⛏

‍

With the Results Overview section out of the way, you can make your way over to the Participants Table. This is where you can find information about the individual people who took part in your card sort. You can also start to filter your data here.

‍

Here are just a few of the different actions that you can take:

‍

• Review your participants, and include or exclude certain individuals based on their card sorts. This is a useful tool if you want to use your data in different ways.
• Segment and reload your results. This function can allow you to view data from individuals or groups of your choosing.
• Add additional card sorts. If you also decided to run manual (in-person) card sorts using printed cards, you can add this data here.

‍

Analysing open and hybrid card sort data 🕵️♂

‍

The Categories tab is the best place to go for open and hybrid card sort results. Take some time to scan the categories people came up with and you’ll be able to quickly build up a good understanding of their ‘mental models’, or how they perceived the theme of your cards.

‍

Consider how different the categories might look for cards containing food items, for example. Some participants might create categories reflecting supermarket aisles, while others might create categories reflecting food groups.

‍

A good place to get started here is by refining your data. Standardize any categories that have similar labels (whether that’s wording, spelling or capitalizations etc). Hybrid card sorts have some set categories, and these will already be standardized.

‍

Note: Before you start throwing categories with similar labels together, take a closer look to see if people had the same conceptual approach. Here’s an example from our card sorting 101 guide:

‍

Of the 15 groups with the word ‘Animal’ in the label, 13 had a similar set of cards, but two participants had labeled their categories slightly differently (Animals and Environment’ and ‘Animals and Nature’) and had thus included extra cards the others didn’t have (‘Glaciers melting faster than previously thought’, for example).

‍

Reviewing the Similarity Matrix 🤔

‍

One really useful tool for understanding how your participants think is the Similarity Matrix. This view shows you the percentage of people who grouped 2 cards together.

The most closely related pairings are clustered along the right edge. Higher agreement between participants on which cards go together equates to darker and larger clusters.

‍

‍

There are a few different ways to use the insights from the Similarity Matrix:

‍

• Put together a draft website structure based on the clusters you see on the right.
• Identify which card pairings are most common (and as a result should probably go together on your website).
• Identify which card pairings are least common so you don’t need to waste time considering how they might work on your website.

‍

Spotting popular card groupings 🔍

‍

Dendrograms are a tool to enable you to spot popular groups of cards, as well to get a general feel of how similar or different your participants’ card sorts were to each other.

‍

There are two dendrograms to explore:

‍

• More than 30 card sort participants: The Actual Agreement Method (AAM) dendrogram gives you the data straight: “X% of participants agree with this exact grouping”.
• Fewer than 30 card sort participants: The Best Merge Method (BMM) tells you “X% of participants agree with parts of this grouping”, and so enables you to extract as much as you can from the data.

‍

Looking for alternative approaches 👀

‍

The Participant-Centric Analysis (PCA) view can be useful when you have a lot of results. It’s quite simple. Basically, it aims to find the most popular grouping strategy, and then find two more popular alternatives among participants who agreed with the first strategy.

‍

This approach is called Participant-Centric Analysis because every response (from every participant) is treated as a potential solution, and then ranked for similarity with other responses. What this is telling you is that if you see a card sort with a 11/43 agreement score, this means 10 other participants sorted their cards into groups similar to these ones. 

‍

Taking the next step: Run a card sort and try analysis for yourself 🃏

‍

Now that we’ve taken a bit of a deep dive into the analysis side of card sorting in OptimalSort, it’s time to take the tool for a spin and start generating your own data.

‍

Getting started is easy. If you haven’t already, simply sign up for a free account (you don’t need a credit card) and start a card sort. You can also practice by creating a card sort and sending it out to your coworkers, friends or family. Once you start to see results trickling in, you can start to make sense of the data.

‍

For more information, check out the card sorting 101 guide that we’ve put together, or our introduction to card sorting on the Optimal Workshop Blog.

‍

Happy testing! 

Usability guru Jared Spool has written extensively about the 'scent of information'. This term describes how users are always 'on the hunt' through a site, click by click, to find the content they’re looking for. Tree testing helps you deliver a strong scent by improving organisation (how you group your headings and subheadings) and labelling (what you call each of them).

‍

Anyone who’s seen a spy film knows there are always false scents and red herrings to lead the hero astray. And anyone who’s run a few tree tests has probably seen the same thing — headings and labels that lure participants to the wrong answer. We call these 'evil attractors'.In Part 1 of this article, we’ll look at what evil attractors are, how to spot them at the answer end of your tree, and how to fix them. In Part 2, we’ll look at how to spot them in the higher levels of your tree.

‍

The false scent — what it looks like in practice

‍

One of my favourite examples of an evil attractor comes from a tree test we ran for consumer.org.nz, a New Zealand consumer-review website (similar to Consumer Reports in the USA). Their site listed a wide range of consumer products in a tree several levels deep, and they wanted to try out a few ideas to make things easier to find as the site grew bigger.We ran the tests and got some useful answers, but we also noticed there was one particular subheading (Home > Appliances > Personal) that got clicks from participants looking for very different things — mobile phones, vacuum cleaners, home-theatre systems, and so on:

‍

‍

‍

The website intended the Personal appliance category to be for products like electric shavers and curling irons. But apparently, Personal meant many things to our participants: they also went there for 'personal' items like mobile phones and cordless drills that actually lived somewhere else.This is the false scent — the heading that attracts clicks when it shouldn’t, leading participants astray. Hence this definition: an evil attractor is a heading that draws unwanted traffic across several unrelated tasks.

‍

Evil attractors lead your users astray

‍

Attracting clicks isn’t a bad thing in itself. After all, that’s what a good heading does — it attracts clicks for the content it contains (and discourages clicks for everything else). Evil attractors, on the other hand, attract clicks for things they shouldn’t. These attractors lure users down the wrong path, and when users find themselves in the wrong place they'll either back up and try elsewhere (if they’re patient) or give up (if they’re not). Because these attractor topics are magnets for the user’s attention, they make it less likely that your user will get to the place you intended. The other evil part of these attractors is the way they hide in the shadows. Most of the time, they don’t get the lion’s share of traffic for a given task. Instead, they’ll poach 5–10% of the responses, luring away a fraction of users who might otherwise have found the right answer.

‍

Find evil attractors easily in your data

‍

The easiest attractors to spot are those at the answer end of your tree (where participants ended up for each task). If we can look across tasks for similar wrong answers, then we can see which of these might be evil attractors.In your Treejack results, the Destinations tab lets you do just that. Here’s more of the consumer.org.nz example:

‍

‍

Normally, when you look at this view, you’re looking down a column for big hits and misses for a specific task. To look for evil attractors, however, you’re looking for patterns across rows. In other words, you’re looking horizontally, not vertically. If we do that here, we immediately notice the row for Personal (highlighted yellow). See all those hits along the row? Those hits indicate an attractor — steady traffic across many tasks that seem to have little in common. But remember, traffic alone is not enough. We’re looking for unwanted traffic across unrelated tasks. Do we see that here? Well, it looks like the tasks (about cameras, drills, laptops, vacuums, and so on) are not that closely related. We wouldn’t expect users to go to the same topic for each of these. And the answer they chose, Personal, certainly doesn’t seem to be the destination we intended. While we could rationalise why they chose this answer, it is definitely unwanted from an IA perspective. So yes, in this case, we seem to have caught an evil attractor red-handed. Here’s a heading that’s getting steady traffic where it shouldn’t.

‍

Evil attractors are usually the result of ambiguity

‍

It’s usually quite simple to figure out why an item in your tree is an evil attractor. In almost all cases, it’s because the item is vague or ambiguous — a word or phrase that could mean different things to different people. Look at our example above. In the context of a consumer-review site, Personal is too general to be a good heading. It could mean products you wear, or carry, or use in the bathroom, or a number of things. So, when those participants come along clutching a task, and they see Personal, a few of them think 'That looks like it might be what I’m looking for', and they go that way.Individually, those choices may be defensible, but as an information architect, are you really going to group mobile phones with vacuum cleaners? The 'personal' link between them is tenuous at best.

‍

Destroy evil attractors by being specific

‍

Just as it’s easy to see why most attractors attract, it’s usually easy to fix them. Evil attractors trade in vagueness and ambiguity, so the obvious remedy is to make those headings more concrete and specific. In the consumer-site example, we looked at the actual content under the Personal heading. It turned out to be items like shavers, curling irons, and hair dryers. A quick discussion yielded Personal care as a promising replacement — one that should deter people looking for mobile phones and jewellery and the like.In the second round of tree testing, among the other changes we made to the tree, we replaced Personal with Personal Care. A few days later, the results confirmed our thinking. Our former evil attractor was no longer luring participants away from the correct answers:

‍

Testing once is good, testing twice is magic

‍

This brings up a final point about tree testing (and about any kind of user testing, really): you need to iterate your testing —  once is not enough.The first round of testing shows you where your tree is doing well (yay!) and where it needs more work so you can make some thoughtful revisions. Be careful though. Even if the problems you found seem to have obvious solutions, you still need to make sure your revisions actually work for users, and don’t cause further problems. The good news is, it’s dead easy to run a second test, because it’s just a small revision of the first. You already have the tasks and all the other bits worked out, so it’s just a matter of making a copy in Treejack, pasting in your revised tree, and hooking up the correct answers. In an hour or two, you’re ready to pilot it again (to err is human, remember) and send it off to a fresh batch of participants.

‍

Two possible outcomes await.

• Your fixes are spot-on, the participants find the correct answers more frequently and easily, and your overall score climbs. You could have skipped this second test, but confirming that your changes worked is both good practice and a good feeling. It’s also something concrete to show your boss.
• Some of your fixes didn’t work, or (given the tangled nature of IA work) they worked for the problems you saw in Round 1, but now they’ve caused more problems of their own. Bad news, for sure. But better that you uncover them now in the design phase (when it takes a few days to revise and re-test) instead of further down the track when the IA has been signed off and changes become painful.

‍

Stay tuned for more on evil attractors

‍

In Part 1, we’ve covered what evil attractors are and how to spot them at the answer end of your tree: that is, evil attractors that participants chose as their destination when performing tasks. Hopefully, a future version of Treejack will be able to highlight these attractors to make your analysis that much easier.

‍

In Part 2, we’ll look at how to spot evil attractors in the intermediate levels of your tree, where they lure participants into a section of the site that you didn’t intend. These are harder to spot, but we’ll see if we can ferret them out.Let us know if you've caught any evil attractors red-handed in your projects.

In 2009, Bob Bailey and Cari Wolfson published published findings that changed how we approach first click testing and usability testing. They analyzed 12 scenario-based user tests and found that if someone gets their first click right, they're about twice as likely to complete their task successfully. This finding was so compelling that we built First Click Testing (formerly Chalkmark) specifically to help teams test this.  But we'd never actually validated their research using our own data, until now.

‍

Turns out, we're sitting on one of the world's largest databases of tree testing results. So we analyzed millions of task responses to see if the "first click predicts success" hypothesis holds up.

‍

It does. Convincingly.

‍

Users who get their first click correct are nearly three times more likely to complete their task successfully (70% vs 24% success rate).

‍

Here's how we validated the original study, what our data shows, and why first clicks matter more than you might think.

‍

‍

‍

Original first click testing study: 87% task success rate

‍

‍

Bob and Cari analyzed data from twelve usability studies on websites and products with varying amounts and types of content, a range of subject matter complexity, and distinct user interfaces. They found that people were about twice as likely to complete a task successfully if they got their first click right, than if they got it wrong:

‍

If the first click was correct, the chances of getting the entire scenario correct was 87% if the first click was incorrect, the chances of eventually getting the scenario correct was only 46%.

‍

‍

‍

Our Tree Testing data: First clicks predict 70% task success rate

‍

‍

We analyzed millions of tree testing responses in our database. We've found that people who get the first click correct are almost three times as likely to complete a task successfully:

‍

If the first click was correct, the chances of getting the entire scenario correct was 70% if the first click was incorrect, the chances of eventually getting the scenario correct was 24%

‍

To give you another perspective on the same data, here's the inverse:

‍

If the first click was correct, the chances of getting the entire scenario incorrect was 30% if the first click was incorrect, the chances of getting the whole scenario incorrect was 76%

‍

‍

‍

How Tree Testing measures first click success and task completion

‍

‍

Bob and Cari proved the usefulness of the methodology by linking two key metrics in scenario-based usability studies: first clicks and task success. First Click Testing doesn't measure task success — it's up to the researcher to determine as they're setting up the study what constitutes 'success', and then to interpret the results accordingly. Tree Testing (formerly Treejack) does measure task success — and first clicks.

‍

In a tree test, participants are asked to complete a task by clicking though a text-only version of a website hierarchy, and then clicking 'I'd find it here' when they've chosen an answer. Each task in a tree test has a pre-determined correct answer — as was the case in Bob and Cari's usability studies — and every click is recorded, so we can see participant paths in detail.

‍

Thus, every single time a person completes an individual tree testing task, we record both their first click and whether they are successful or not. When we came to test the 'correct first click leads to task success' hypothesis, we could therefore mine data from millions of task.

‍

To illustrate this, have a look at the results for one task. The overall Task result, you see a score for success and directness, and a breakdown of whether each Success, Fail, or Skip was direct (they went straight to an answer), or indirect (they went back up the tree before they selected an answer):

‍

‍

‍

‍

In the pie tree for the same task, you can look in more detail at how many people went the wrong way from a label (each label representing one page of your website):

‍

‍

‍

‍

In the First Click tab, you get a percentage breakdown of which label people clicked first to complete the task:

‍

‍

‍

‍

And in the Paths tab, you can view individual participant paths in detail (including first clicks), and can filter the table by direct and indirect success, fails, and skips (this table is only displaying direct success and direct fail paths):

‍

‍

‍

‍

‍

How to run first click tests: Best practices for usability testing

‍

‍

First click analysis is one of the most predictive metrics in usability testing. Whether you're testing wireframes, landing pages, or information architecture, measuring first click success gives you early insight into whether your design will work.

‍

This analysis reinforces something we already knew: first clicks matter. It is worth your time to get that first impression right. You have plenty of options for measuring the link between first clicks and task success in your scenario-based usability tests. From simply noting where your participants go during observations, to gathering quantitative first click data via online tools, you'll win either way. And if you want quantitative first click data, Optimal has you covered. First Click Testing works for wireframes and landing pages, while Tree Testing validates your information architecture.

‍

To finish, here are a few invaluable insights from other researchers on getting the most from first click testing:

‍

• Jeff Sauro details a useful approach to running a first click test, and shares the findings from a test he ran on 13 people.
• An article on Neoinsight describes three common usability problems that first click testing can solve.
• Gianna LaPin describes a first click test she ran on Netflix, VUDU, and Hulu Plus.

‍

‍

About this study

‍

This analysis was conducted in 2015 using millions of task responses from Optimal’s First Click and Tree Testing tools. While the dataset predates recent UI trends, the underlying behavioral principle, that a correct first click strongly predicts task success, remains consistent with modern usability research.