How-to guide for intranet card sorting

Comments6

Card sorting is one of the most common and important tools for building user-friendly intranets. It’s a core user-centered design technique for both intranets and public-facing websites and plays an important role in the standard process for developing an intranet’s information architecture (the site navigation).

3 examples of intranet navigations
Three top-level navigation examples from ThoughtFarmer intranets

For teams who aim for rapid and sustained intranet adoption, building an intuitive site navigation is a critical pillar of success. By helping employees easily find the information they need every day, you can increase site visits and get employees to stay on the new intranet longer.

Card sorting can be a quick and inexpensive process and, in addition to its usability benefits, will help engage your users in the intranet project in a meaningful way.

What is card sorting?

Card sorting is a structured exercise that asks volunteer intranet users to sort a list of sample intranet content into logical groupings. Card sorting comes in two flavours: 1) open card sorting and 2) closed card sorting. Open card sorting is more popular and provides more useful insight in the early stages of building an intranet’s information architecture. For this reason we only cover open card sorting in this blog post.

Card sorting can be conducted online using special software or offline (in-person). During the exercise a user is provided with a list of examples of intranet content, hand selected by the tester (the person running the card sort). The user sorts the items on the list into groups of related cards and labels each grouping. When a number of users have completed the exercise the tester compares each of the users’ results in aggregate to look for patterns and insights as to where similar groupings did and did not occur.

Why is card sorting important?

At its most fundemental level card sorting is about building an intranet navigation structure using terms and groupings that match users’ perspectives. The result is navigation that is easy to use. An effective card sort can help you find patterns and surface users’ mental models, the interpretation of which guides the intranet’s design.

Many poorly designed intranets are built by one person or a small group who create the navigation around their own personal perspectives. This can lead to sites that are self-centered, overly prescriptive, or mechanical. These sites match how the designer thinks everyone else should see the content, rather than how users actually understand the material at hand.

Card sorting can help intranet teams look beyond their own blinders and build an information architecture that employees will be more likely to adopt. The practices listed below can help you produce good results, but card sorting analysis is not terribly formulaic. Keep in mind that card sorting can lend insight and provide direction, but does not necessarily give exact answers. This is why it is as much an art as a science.


Video: Build a site navigation in ThoughtFarmer

Once you have conducted card sorting and other exercises to create a user-friendly information architecture, you then need to build the navigation using your intranet software.

This brief video demonstrates the simple steps to building a complete hierarchical intranet navigation using ThoughtFarmer:

If you like the look of these simple site-building tools, sign up for a live demo to see more of ThoughtFarmer in action.


When to use card sorting

Card sorting is an early stage tool for crafting an information architecture and is considered a “design” tool rather than an “evaluation” tool. This means you use it to create a navigation, but not to validate the effectiveness of your proposed structures. Card sorting can result in a draft high-level site navigation, which task testing can then revise and refine further.

New site, redesign or section update: People typically employ card sorting when a new site needs to be built or a site needs a full redesign. It can also play a critical role in updating a specific section of content rather than a whole site.

Following a content audit: The content audit provides a snapshot of all your potential intranet material. From the material listed in this snapshot you can select representative content for card sorting.

Before task testing: Task tesing is an evaluation/validation step that helps you refine your information architecture. Before getting to that stage you must first complete a round of user-centered exercises such as card sorting to come up with a draft design.

Before visual design (sometimes): For many websites and intranets, card sorting (and task testing) come before the visual design stage (wireframing, sizing and selecting page elements, selecting colors, fonts, logos, etc.). With out-of-the-box social intranet software like ThoughtFarmer, where much of the site layout is already determined, this sequence is less relevant.  In this case, the site navigation becomes the largest effort of your intranet site customization and can take place before, during or after visual design.

Without a live site: Card sorting, task testing, and other information architecture activities can take place without having a live site up and running. As long as you document the results in a MS Excel spreadsheet or something similarly formatted, you can apply the information architecture whenever you eventually get your intranet site online.

Online card sorting: the new normal

Traditionally, card sorting was completed in a room with the participants using paper index cards, which they physically stacked. Today, web-based card sorting tools offer a flexible online alternative. Our favorite tool is OptimalSort, which offers easy setup and rather extensive analysis and data visualization features. While online card sorting is quite handy, offline card sorting has its own useful attributes.

Example of an online card sort using OptimalSort

Benefits of onilne card sorting

  • Easily involve geographically dispersed audience
  • Reach more people
  • Faster for participants
  • Timing more flexible for participants
  • Analysis tools built in
  • Exportable data

Benefits of offline card sorting

  • Option of group-based card sorting
  • Rich contextual discussions (for group-based card sorting)
  • Opportunity for interviews & surveys
  • Greater participant commitment, higher level of engagement

Alex Manchester of Step Two Designs recently wrote a more in-depth comparison of the two methods in his article Card sorting: online vs. offline.

If you have the time and resources to conduct both you may find that online and offline card sorting complement each other well. However, most of our clients house employees in distant offices and don’t have the resources to conduct in-person card sorting at multiple locations. Because of this we often default to online card sorting for intranets and, therefore, only cover that in detail in this post.

For in-depth instructions about offline card sorting, see Donna Spencer‘s cornerstone blog post Card sorting: a definitive guide as well as James Robertson‘s evergreen reference post Information design using card sorting.

While online card sorting software does provide useful analysis and visualization results, you’ll still need to spend time analyzing the results. The automatic analysis presents the data, which you still must interpret and extrapolate.

How intranet card sorting is special

Card sorting is frequently used in the design of public facing websites — the ones we can search for and find through Google, Yahoo!, etc. — as well as for intranets. But intranet card sorting carries a few specific characteristics and benefits that are different from card sorting for more marketing-focused sites.

For starters, the information on an intranet is uniquely focused on a company, and employees need the information to succeed in their daily work. In a company, if an employee can’t find needed material on the intranet, she will find other ways to get it.

Paradoxically, employees are very familiar with much of the content, but if they can’t find it easily on the intranet they’re prone to abandoning the intranet wholesale.

This issue of intranet adoption through trust of the content is related to the second major way intranet card sorting is different from normal website card sorting: user engagement.

Card sorting as intranet user engagement

While card sorting for public websites is a critical part of the design process, the users involved typically have no long-standing interaction with the web team or the website. The web team finds people who fit the website’s target demographic, brings them in to participate in card sorting and then sends them on their way, perhaps never to hear from them again.

Conversely, the participants in an intranet card sorting exercise are your users and will return to the site day after day (hopefully). In fact, the participants could end up being stakeholders, super users, or intranet champions.

For this reason, card sorting plays a dual role, both as a design tool and an engagement tool. A common truism about intranet projects is that by involving users in intranet development you achieve stronger sustained adoption. Additionally, any employees who participated in card sorting will be less likely to complain about the resulting intranet navigation. Why? Because they helped build it!

For more details on how participitory intranet design can lead to increased user engagement and adoption see our article 15 ways to engage users in building a social intranet.

Step-by-step guide to intranet card sorting

With all of that background provided, now it’s time to get down to brass tacks. These steps lay out our standard process for conducting intranet card sorting.

STEP 1: Identify participants

This step could theoretically come after Step 5, but we’ve listed it at the beginning for two reasons: 1) Many intranet teams put off user engagement until the very last minute and have to hustle to secure involvement and 2) it’s important to involve a broad array of users, not just those you can reach most easily in a short amount of time; this requires some time and planning earlier on.

Card sorting aims to capture employees’ perspectives on how different pieces of information fit together, but not all employees have the exact same perspective. The way people see and use a specific piece of content may vary based on their job levels, departments, locations, etc. It is therefore important to involve employees who span the spectrum on a number of different criteria.

Employee characteristics to consider:

  • Hierarchy
  • Location
  • Department
  • Role

Try to involve employees from throughout the hierarchy, from admins up to executives, as well as people from every department and location. Not only does this help you capture a full array of perspectives, but it also supports your user engagement efforts. Imagine having a couple employees in every department and location who participated in this fascinating exercise and who can talk up the new intranet project to colleagues.

Finally, how many participants do you need?

Renowned usability expert Jakob Nielsen wrote a blog post which provides some scientific insights into how many users to test in card sorting. Spoiler alert: Nielsen recommends including 15 people in the card sorting.

We see that as your base number and recommend that you ensure a minimum of 15 users, but aim for double that. More user testers won’t hurt the results and will broaden your reach for user engagement.

So, if you are going to send email invitiations to colleagues inviting them to complete a 15-minute card sorting exercise online within a two-day period, you may need a list of up to 50 people just to ensure an adequate number of responses.

STEP 2: Review content audit & select topics for cards

We presume that prior to the card sorting you have completed a full content audit that will provide you with a rich list of content from which to select your examples. See the ThoughtFarmer how-to guide on intranet content audits to learn more about that process.

To select topics for cards, begin by first scanning your full content audit to get a lay of the land.

You’ll want a total of 50 to 60 cards, each with an individual piece of content/topic.

Try to pick content that is universally relevant. Building a new intranet is a tremendous opportunity to learn about your organization. One thing you may discover is that people in different locations and roles use very different information than you use. But you don’t want to learn this lesson by picking cards that are foreign to some of your testers. If people see cards they don’t understand, they will have trouble completing the exercise and may feel the intranet is not relevant to their work. To address this issue you can complete a comprehensive content audit and then select content you think is used across the entire company.

Note that if your content audit revealed large gaps in your existing material, or you have big dreams about new content for the new intranet, it is okay to include future content in the card sorting. If you do this, just be sure the content is labeled in simple lay terms that people will understand.

You may find it easy to use a spreadsheet to list and track the topics for the card sort.

STEP 3: Analyze topics for consistent level of granularity

Next, start to review your cards for a consistent level of granularity. What does this mean exactly?

Will you include only individual pages and files, such as “Travel Expense Form,” or will you move up to a broader section/folder level of content, such as “Travel forms”? These two examples offer different levels of granularity. One is a specific piece of content while the other is a section that contains specific pieces of content.

This is important because without a consistent level of content granularity, you may confuse users and skew results.

One rule of thumb to help you achieve consistency: ask the question, “Can this card fit as a topic underneath any other card?” If the answer is yes, you may need to replace broader content with more specific content (or vice versa).

Keep in mind that the goal of a card sort is to surface users’ mental models and assumptions about content. Considering that, try to avoid listing “manuals” or “guides” or “handbook” or similar terms that pre-suppose content groupings that may not make sense to users. When you use such terms you are embedding your own assumptions in the cards, which can skew results.

STEP 4: Write brief card descriptions (optional)

Once you’ve got a list of 50-60 cards, type up a brief description for each. If you’re doing a paper card sort, then you can write the descriptions on the back sides of the cards. If you’re doing an online card sort using software such as Optimal Sort, you can enter in a description for each term.

A description should be a very brief at-a-glance statement that clarifies a card’s meaning. You don’t want to over-explain a term and provide words that will heavily influence how users sort the cards.

Examples:

“Travel expense form” | Submit to request funds prior to a business trip

“Company dictionary” | Lists common terms & acronyms with explanations

“Instructions for remote email access” | How to access company email from outside of the office

STEP 5: Create & test online card sort

Once you’ve created your list of cards along with brief descriptions (optional), log in to your selected online card sorting tool and enter your cards into a new card sort. This should be a quick and easy process. If it isn’t, consider switching to a different tool (we prefer OptimalSort).

After saving your new card sort online, run a few tests to make sure it is working properly and that the cards and descriptions display the way you want.

STEP 6: Invite participants

Once you are happy with the online card sort and have fully tested it, select a closing date and set the card sort to go live.

Craft an email that you can send to all participants with a link to the card sort. This instructional email should include the following information:

  • Introduction to the intranet project
  • Explanation of the important role employees will play in crafting the the structure of the new intranet
  • A link to the card sort itself
  • Expected amount of time required to complete the card sort
  • The deadline by which the card sort must be completed
  • Who to contact with questions

While we often provide clients with a template for this email, it is important to use your own language and style that matches your company culture and norms.

Additionally, it may be useful to send all participants an email (or reach out to them in some other way) a week or two prior to the card sort as a heads up. By alerting colleagues in advance, before sending the card sort invitation, you may be able to discover important questions beforehand, find out which potential participants will be out of the office during the card sort period, and offer yet one more opportunity to engage with your users.

STEP 7: Close card sort & analyze results 

You can leave the card sort open for anywhere from two days to one week from the time you invite participants. Unlike large employee satisfaction surveys, a card sorting exercise is brief and can be quite fascinating to participants — almost like a puzzle.

If you invite colleagues in distant time zones to participate, ensure they have at least two full business days from the time they receive the invitation to complete the exercise.

As you begin analyzing the card sorting results, keep in mind that you are looking for both general insights and specific ideas for how to craft the top two levels of the new intranet’s navigation.

The card sort exercise addresses two key questions:

  • How do users group sample content?
  • What labels do users write to describe groups of content?

These are the questions your analysis should focus on. First look at what content was frequently grouped together. These groupings can form the basis for your first and second levels of intranet navigation.

Next, look at the phrases people typed in to describe the card groups they created. What terms popped up the most? What themes appeared in the group labels? Strong themes provide insights into users’ mental models and can be used as the core of the new site’s navigation.

Example of data visualization & analysis from Optimal Sort
Example of automatic card sorting data visualization & analysis from Optimal Sort

Not every insight you gain from card sorting will offer a concrete path forward. Instead, the results may highlight unclear trends that you must further explore in task testing. Or, perhaps you’ll discover that people associate a specific piece of content with two different departments, which will require cross-linking the content no matter where it ends up living on the site.

For example: one common scenario is that employees see “paychecks” as tied equally to “Finance” and “HR.” This may lead you to use task testing to validate a proposed “Time & Pay” section of the intranet which would house all content and links related to paychecks.

A final word on card sorting

Intranet card sorting can be fun, quick, and useful, but becoming an expert can take some time. If you’ve never run a card sort before, take an experimental approach and assume you’ll be learning all along the way. You could even start with a small card sort around a specific collection of content (like HR material) in order to develop your card sorting analysis skills.

Any card sorting is better than none, so even if you stumble through it a bit, the resulting insights and user engagement will be well worth it. Good luck!

Comments6

Join The Discussion

  1. Alex Manchester

    Thanks for the mention, Ephraim! Another sterling effort from you guys on this article.

    It is my strongly-held opinion that using online card sorting *only* is not particularly helpful.

    If it is a struggle to travel everywhere (as it can be in the continental Americas!), I would then argue for getting people to come to you, or create a tutorial/guide/host some video conference meetings to discuss the activity, then send out ‘train the trainer’ packs of physical cards and help people to conduct them in their offices. That’s always my first few recommendations if we can’t do the sorts ourselves in remote offices.

    Only as a last resort would I use solely online sorts for data collection – as said in my article there’s just so much more to the conversations and tidbits you pick up than just the sort data.

  2. Ephraim Freed

    Thanks for posting these thoughtful comments Alex.

    You make good points and your approach aligns with a more ethnographical process, rather than just a pure information management philosophy.

    The reality, though, is that many intranet teams are under the gun and have limited time and resources. Even the excellent “train the trainer” approach you’ve mentioned requires a lot of time and planning.

    If people can follow your advice, that’s excellent, but if they can’t then only using online card sorting is still much better than doing none at all.

  3. Alex Manchester

    “If people can follow your advice, that’s excellent, but if they can’t then only using online card sorting is still much better than doing none at all.”

    Yes, agree. “Some user testing is better than none” is absolutely true.

  4. Andrew Wright

    While I think card sorting can be useful, I think there are a couple of pitfalls to be wary of.

    Many intranet systems these days provide meta data functionality which makes it possible to automatically create multiple navigation paths to the same content. For example, an Expense Claim Form can sit under a heading called Forms and ALSO the headings of Finance, HR or Popular Documents. There is no reason that content should belong to only one heading. In fact, good intranet sites will provide different ways for their staff to find content.

    Often in card sorting exercises, people can get flustered because they feel they can only put the card under one heading, when in fact they want to put in more than one heading.

    To avoid this situation (and the associated confusion) and to get the most value out of card sorting, I suggest the following:
    * Before the card sorting exercise, ‘Pre-sort’ the cards into categories eg. put all the procedures together or all the forms together. Then use these groups for your card sorting.
    * Explain to the participants that they will be asked to group the same cards a number of times – not just once
    * The first time around, get them to group the cards according to their own ‘free form’ categories
    * For subesequent card sorts, ask them to group the cards according to a specific requirement. For example:
    – group forms by which are the most popular
    – group forms by when they are used
    – group forms by which department owns the form
    – group forms by which role uses the form the most

    As you can imagine, this approach will generate different outcomes from the same set of cards. And each of these outcomes can actually be implemented with the use of meta data.

    Here are a couple of articles with more information:

    Card Sorting Doesn’t Cut the Custard
    http://www.zefamedia.com/websites/card-sorting-doesnt-cut-the-custard/

    Contextual integration: how it can transform your intranet
    http://cibasolutions.typepad.com/wic/2011/03/contextual-integration-how-it-can-transform-your-intranet.html

  5. Ephraim Freed

    Thanks for sharing your thoughts Andrew.

    It’s good for folks to know that there are several ways to gather the needed information for building a user-friendly intranet IA.

    We have found that card sorting provides a decent amount of information about secondary relationships. While most users may group a specific card with a consistent set of other cards, some will group that card with different sets of cards.

    We use the secondary groupings to provide a lot of the information you mention. It sounds like both methods can work well.

    Your post on contextual integration is a very good read, especially for folks who are implementing traditional CMS software for their intranets. Thanks for posting it!

  6. Steve

    Another budget-friendly online Card Sorting tool: http://www.usabilitest.com/card-sorting

Comments are closed.