In our work, we often exit the research phase of projects with an immense load of valuable data and analytics. However, coming to an agreement on the insights and directions to pursue next is never easy. Parallel Clustering, a technique developed at Jump and recently contributed to the new book, Exposing the Magic of Design, allows teams to reach a consensus quickly when the clock is ticking.

Published as a contribution to the book, Exposing the Magic of Design by John Kolko. January 2011

Imagine you’re working on a team-based research project. You’ve already amassed a large set of data and done your analysis. You now have a number of interesting insights and numerous directions that you could follow. Unfortunately, everyone on the team has a different perspective on what’s most important. And the clock is ticking… your team needs to converge on some winning concepts sooner rather than later. How do you synthesize multiple perspectives into a set of clear, prioritized directions, while keeping true to what’s most important? More importantly, how do you prescribe where to go next, as opposed to simply describing the next steps? For some, Parallel Clustering is the answer.

Key Benefits

A synthesis approach developed at Jump Associates, Parallel Clustering takes numerous insights and turns them into clear, prioritized directions. For teams, it looks and feels like a unique mix of familiar activities: clustering, Mindmapping, and idea ranking. Yet, Parallel Clustering is different. Unlike a simple clustering exercise, Parallel Clustering quickly reveals multiple alternatives for grouping information. Different from a final ranking or selection activity, it builds robust themes, rather then simply filtering ideas out. Distinct from personal Mindmapping (creating a network-like diagram used to represent the insights and ideas that an individual believes to be linked to one another) – shared maps make everyone’s thinking transparent to the entire group. These three disciplines, as brought together under the umbrella of Parallel Clustering, lead to numerous benefits. These include the ability to:

Manage complexity.

Information overload can be a challenge for any synthesis activity. Parallel Clustering forces each team member to quickly articulate the top ideas/themes/needs/etc that they view as most important. This activity helps to cut through all of the clutter and very quickly understand what’s most important.

Foster collaboration.

Most teams are made up of individuals with various personalities and working styles. Some folks are quiet. Others can dominate a group. And some people simply need more time than others to process what they think. Making everyone’s thought processes explicit within a group can be a challenge. Parallel Clustering gives everyone the time and space they need to share what they think is most important and why. This helps teams to tap into unique individual perspectives and truly harness the team’s wide array of experiences and skills.

See new possibilities.

Even when teams take the time to listen to one another, it doesn’t mean that everyone hears the same thing. It can be difficult to truly understand others’ nuances. Since these various flavors can add richness and depth to insights, systems must be put into place to capture the details. Creating a shared recording or Mindmap helps to reveal both the similarities and differences in people’s framing of topics. By physically clustering similar or overlapping ideas in proximity to one another, you will quickly reveal what people agree on, as well as where there are points of difference that need to be discussed.

Make trade-off decisions.

At its essence, synthesis is about combining two or more entities to form something new. And when it’s time for teams to make decisions about which direction to pursue, there are many ways to gain clarity. Visually representing the different ways folks are prioritizing identical data points helps to foster a rich discussion on the must-haves, the nice-to-haves, and the different paths for getting there.

Move quickly.

When deadlines approach and the path forward is unclear, how do you make decisions? With a little bit of foresight and planning, the entire Parallel Clustering activity can be completed in a few hours. Get the team together in a room. Force people to “think with their gut.” Prioritize what’s most important. Get all the ideas out. Share as a group. Put these together, and you’ll move from chaos to clarity with great speed.

Common Problems

At Jump, we continually challenge ourselves to learn and improve our craft. As a result of this effort, we stumbled upon the method of Parallel Clustering. We offer these indicators so that others might learn from our mistakes. Here are a few early warning signs that your synthesis process might benefit from Parallel Clustering:

Letting one voice dominate.

A strong voice can dominate any team and lead to groupthink. Watch for individuals who make decisions for the group without gathering input from all valuable parties.

Going vanilla.

The process of building consensus can sometimes drive all the good stuff out. Don’t lose the nuggets of insight that make your ideas great.

Too much talk.

Team conversations are typically just that – lots of talk! Mix it up by writing everyone’s comments on a white board. Draw sketches. Capture ideas. Make connections and links. This helps everyone to feel heard, track the conversation and identify patterns.

Case Study

Project Background – Understanding the BRICS.

In 2005, Jump Associates partnered with a large, international client to help them explore their emerging markets strategy. The project name, BRICS, is an acronym first coined by Goldman Sachs in 2001 that refers to the developing economies of Brazil, Russia, India, China and South Africa. The goal of our work was to understand broad cultural themes in order to define near-term business opportunities in these countries. The Jump team was made up of four hybrid thinkers, who closely collaborated with our clients to better understand the day-to-day needs for small to medium-sized business people, and the cultural drivers that are different than those in the United States. The project data came primarily from secondary research and expert interviews.

Selecting Methodologies – Using Parallel Clustering to quickly prioritize themes.

One of our deliverables for the BRICS project was a set of actionable themes based on our research. After weeks of analysis we had numerous insights, but were lacking a cohesive story. In fact, we had too many insights, too many interesting directions, and a lack of clear focus on what was most important. We had already spent time on clustering activities, such as sorting Post-it Notes. But that only got us so far. It was time to turn insights into action. We decided to organize a Parallel Clustering activity to help us extract and prioritize our top five themes.

Although our client didn’t directly participate in the Parallel Clustering activity, we were able to share our process with them after the fact during discussions. Having documentation in place to show them how we got from point A to point B was helpful in providing transparency and getting buy-in on the final themes.

Understanding the Process – A step-by-step guide to moving forward.

1. Take time to reflect.
On the BRICS project, our first step was to give each team member a short amount of time to reflect on the insights that they thought were most important. We all went our separate ways to gather our thoughts and came back with a top 10 list. The short amount of time forced us to “think with our gut” as we say at Jump. The team had been living and breathing this content for seven weeks and this exercise forced us to quickly synthesis what was already formulating in the backs of our brains for the entire project. (Activity Type: Individual / Time: 15 minutes)

2. Get out all the possibilities.
Once we finished our lists, we came back together as a group. Collectively, we developed a Mindmap that contained all of the possible theme areas. One by one, each of us took a turn sharing our lists. We dedicated one individual to facilitate and record the conversation. As each person shared their top insights, the recorder captured what they said. While we built a shared Mindmap on an 11×17 sheet of paper, other projects at Jump have captured similar discussions on a whiteboard or wall with Post-it Notes. Any of these formats will work.

To create a useful Mindmap, the recorder captured like topics near one another on the sheet of paper. When acting as a recorder, it’s important not to impart personal bias into the Mindmap. It’s dangerous to assume that certain ideas are linked. Rather, ask questions and get confirmation from participants that their ideas are represented accurately. This can be a challenge; especially when the group is trying to move quickly.

On the BRICS project, if someone thought that their ideas were indeed like something else that was already on paper, then the insight was recorded in close proximity. These slight variations ended up being important to capture, not just for content variation, but because they went a long way towards helping everyone on the team feel heard. In many cases, several of us had the exact same insight on our lists. In these cases, we only needed to write it down once.

We concluded the group worksession by making photocopies of our Mindmap, thereby generating identical sets of the data for folks to work with. Groups who choose to Mindmap on a whiteboard rather than on an 11×17 sheet of paper typically take a photo of the whiteboard that they then print out for the team. (Activity Type: Group / Time: 60 minutes / For more info see Figure 1)

3. Make your own clusters.
Next, we all went to work clustering and prioritizing as individuals. To begin, we set a target number for how many clusters we wanted. We chose five to seven. We then took our Mindmaps and circled connections between the insights that each of us, as individuals, saw as having the closest relationships with each other. We each used a different color marker to highlight our top themes. The different colors later helped us to track who said what. While this step sounds simple, individuals are often most challenged at this point. This is the place where folks are really forced to be definitive about how they defined a particular space. (Activity Type: Individual / Time: 15 minutes / For more info see Figures 2 & 3)

4. Compare outputs.
We came back together to compare our results. Armed with our own Mindmaps taped side-by-side on the wall, we compared what we saw. Some clusters were repeated across multiple maps, and were therefore easy to prioritize as final themes. Most clusters had parts and pieces that overlapped across different teammates, but were not exactly the same cut. These differences revealed valuable individual takes on themes. Many comments like “oh, I never thought of it that way” were heard. These differences pointed to interesting nuances and valuable points of view. As we discussed further, we took notes to capture the insights, meaning and decisions we made along the way. (Activity Type: Group / Time: 30 minutes)

5. Define meaningful themes.
The final step of Parallel Clustering is to make final decisions on what cut of themes to go with. Although this might seem challenging, if done right, it should be fairly painless. On BRICS, we had already spent the entire afternoon thinking about it, both individually and as a group. After much discussion, we had a good idea where everyone else in the room stood. We knew what was important, what wasn’t, and different ways to cut the data. Through sharing and discussion, in about 30 minutes we combined all of our thoughts into a final Mindmap which highlighted our final themes. (Activity Type: Group / Time: 30 minutes / For more info see Figure 4)

This article was previously published in Exposing the Magic of Design, a new book by John Kolko. To learn more about the complete work, please click here.

Related posts:

1. Insights from CES 2011
2. How to Find Insights from Research
3. Insights from Design Council UK Visit