In this tutorial you’re going to learn how to use the KJ method.

The KJ, as it’s often called, was created by Jiro Kawakita as an exercise to help teams organize data and get to group consensus quickly. It can be used in a number of ways, but it’s particularly useful for projects guaranteed to have strong opinions or debate.

Example: Imagine you’re leading the redesign of your company’s intranet. You’ve gathered 12 people who represent the different departments in your company. From past experience you know there are a few people who tend to dominate the conversation, like Dan, from Engineering.

In a KJ exercise, participants are not allowed to talk until the very end, which leads to our Big Idea: To get honest feedback, limit the impact of peer pressure and group politics.

Not being allowed to talk helps make this a quick exercise. The average time for a KJ is between 30 and 45 minutes. It also limits the ability for any individual or group to control the conversation. Again, Dan. I mean that guy’s such an [bleep]. Always has to be the smartest one in the room.

Anyway, the first step, choosing a facilitator and the group is already done. If you can’t get everyone in the same room you can use an online tool such as Miro to do it remotely.

The second step is to define a key question, yours is, “What did we learn from the usability study of our intranet?” Everyone on the team has thoroughly reviewed the results of the study before the meeting, and now it’s time to get group consensus on the top priorities.

In step 3, the facilitator reminds everyone that there is no talking allowed and then each participant puts an answer to the question on a sticky note and then puts these on the wall.

Here are some examples of what might come up:

• The search results always show outdated material first
• The time to render a page on mobile is 5 seconds, resulting in user frustration
• Important content is hidden inside PDF documents

KJs work extremely well when you’re working with facts. Here, it’s the data from the usability study, but it’s also okay for people to add their opinions if they don’t see them represented in the data.

Participants should be looking at each others contributions as they’re being added, this helps prevent duplicates and will also spark ideas.

Once there are no more items to add, the next step is to collaboratively organize the items into similar groups, again with no talking.

Examples of group names might include:

• Bad search results
• Out of date content
• Confusing navigation
• Poor mobile performance

These are just four examples of groups, but it’s typical to have up to a dozen.

Now disagreement is going to happen in this phase and that’s OK. For example, “Poor mobile performance” and “Outdated Servers” are competing names. Between you and me, that person who added that was totally Dan. “Outdated Servers” — his reason for everything.

Anyway, in this case, it’s fine for the group to have two names because this will get resolved in the final steps.

Speaking of which, voting is the next step. Everyone votes on the group names they think are best and also, each person writes down the three groups they think are most important in relation to the original question and then ranks them 1, 2 and 3.

So a single vote might might look like this:

1. Confusing navigation
2. Out of date content
3. Bad search results

Once everyone is done, the facilitator tallies the votes and makes the results available for all to see.

And now the last step is the discussion phase and finally the no talking rule is lifted. But the beauty of the KJ is that it’s already clear what the group thinks are the most important priorities and this makes the discussion so much easier.

Remember the conflict between the “Poor mobile performance” and “Outdated Servers” names? Well, it turns out that your team’s consensus was that this entire group was a fairly low priority. So the specific disagreement about the group name just doesn’t merit a lot of discussion.

What does get discussed are things such as whether it makes sense to combine two groups, whether it makes sense to duplicate an item and put it into another group. If there’s a lot of activity here, you can always recast another vote and see if that changes things. In the end though, the final ranked groups are turned into a summary statement that everyone agrees with.

This might be, “We learned from the usability study that our top three priorities should be improving the search engine, improving navigation, and performing a content audit.”

Oh and one more thing, it turns out that updating the intranet’s server will help solve some usability problems after all, I guess Dan actually had a point. Dan.

Thanks for watching. Be sure to check out the rest of the series as well as our entire course catalog at Gymnasium.