Enterprise Design Sprints

The design sprint


by Richard Banfield

With a few minor exceptions, there are five phases to every design sprint. As mentioned in chapters 1 and 2, completing all five phases in five days is the best approach to delivering design-sprint value. In this chapter, I’ll explain the logic behind each phase and why the exercises work best in this order:

  1. Understand
  2. Diverge
  3. Converge
  4. Build
  5. Test

Phase one: Understand


The first day of the design sprint is about reducing the noise of assumptions and establishing a clear signal for why we should be addressing this particular problem. The team will review the background research, identify gaps in knowledge, and expose the riskiest assumptions.

Introductions, agenda review, and roles (20 mins)

My recommendation is for the facilitator to introduce team members to one another well before the sprint starts. A video conference works well for this, allowing everyone involved to see who will be there and start getting comfortable together. For larger organizations with distributed teams, it’s likely people will be meeting each other face-to-face for the first time on day one. So plan for a little extra time for socialization and a round of quick introductions. Use name tags if necessary.

It’s also a good idea to get everyone to share their concerns from the start. Use the Hopes vs. Fears exercise to provide an opportunity to get personal expectations out in the open. Once this is done, the facilitator should assign roles and walk everyone through the agenda.

List assumptions and facts (30 mins)

The first exercise of the day is to list the assumptions on a whiteboard, or equivalent space that everyone can see. The facilitator asks questions like: What do we assume about the customer? What about the current buying experience do we assume is working for the user? Are we sure that the customer can articulate our product’s value? Have we asked customers what they want?

The facilitator uses these questions as prompts for conversation between team members, ensuring that everyone has a chance to provide suggestions or questions of their own.

The assumption board will be referenced and updated throughout the design sprint, so place it somewhere visible. Next to each assumption, write down how the assumption can be tested, and what a validated or invalidated test result would be. Although this process is done primarily during the Understand phase, continue adding assumptions and associated tests as the team discovers them over the course of the sprint.

Assumption Test with… Validated if…
Customers want a shorter checkout process Prototype and interview TBD
Customers understand the value of this feature Interview TBD

Enterprises can harbor institutional-level assumption based on years of habits and even successes. But these assumptions can be extremely dangerous in today’s rapidly changing business environments, especially if they’re incorrect or founded on outdated information.

Review background research and findings (1–2 hours)

It’s important to collect, organize, and distribute background research several days before a design sprint starts. But it’s difficult to ensure the team will review it before arriving. So it’s best to go through the research on day one. For complex problems, reviewing the research will take up a good part of the day. But it must be done. Without background knowledge, the team will be poorly equipped to work on the exercises that follow.

The aim of the research review is to make sure the team has a firm grasp on the business challenges, the customer’s current expectations and the value proposition offered by the business or product. This understanding will make comparisons to competitive options more fact-based and less likely to be influenced by opinion or presumption. (If there is no product yet, and this design sprint is being used to discover a new product opportunity, you might not have a clear value proposition yet.)

Needs, wants, and desires, or real pain points (1 hour)

Identifying the needs of your customer is probably the most important exercise of the day. Sorting between what the user needs, versus what they want, will help the team better understand the problem. For example, users may need to get from location A to location B but how they choose to do that might differ. One user may want to ride a bicycle, while another wants to drive a luxury car. The need is fulfilled in both instances but in very different ways.

User’s journey or critical path (45 mins)

Plotting the user’s journey allows everyone to see where the contact points are between customer and product. The facilitator asks the participants to map the steps a customer takes while interacting with a product. Each participant contributes by adding, editing, or clarifying activities, like “customer searches for lighting solution on shopping site” or “company sends a notification to the user’s smartphone/app.”

By the end of the exercise, the finished product will look much like a children’s treasure map. Loosely mapped touchpoints with brief descriptions of what happens at each point is enough fidelity. High fidelity illustrations won’t add any additional value to this exercise. So just keep it simple.

I find this to be one of the most interesting exercises in the design sprint. How a company visually describes the customer journey explains a lot about the team’s understanding of the customer’s needs. The more the team is aligned around how the customer navigates the journey, the more understanding they have of the problem. As a facilitator, if you notice a lot of disagreement about the journey touchpoints, it’s worth going back and discussing the assumptions driving those interactions.

Develop ‘problem statement’ (45 mins)

Once the background research has been done and the needs of the customer have been established, it’s important to figure out what the problem is that needs solving. Identifying the problem and writing it in statement format also doubles as a vision of the future. Think of the customer problem and the product vision being two sides of the same coin.

My advice is for each person on the team to write their own version of the problem statement using the format below, and then compare versions with the rest of the team. Having discussed the variations as a group, the facilitator can then write a final version on the whiteboard.

To create a problem statement, replace the words in parentheses with your own vision of a solution:

Clearly identifying the problem is important, so don’t be afraid to rewrite this statement a few times or add a longer explanation if it helps with understanding.

You’ll also notice the last two sentences of the statement project what the outcome of a solution might be. It’s unlikely you’ll have a clear solution in mind, so focus instead on the outcome you’re seeking to create. For example, if we were to use a rideshare transportation example, we might say: “We envision a world where owning a car may no longer be a liability. We’re bringing this world about through smartphone access to a new kind of shared transportation solutions.”

As important as it is to determine if there is a problem, it is just as important to understand if that problem is solvable and whether it needs to be solved. The problem statement is the first step to answering the question: What is this product, and is it useful?

Retrospective (15–20 mins)

Before phase 1 is completed, it’s important for the team to circle up and discuss the day’s work, and plan ahead for the next day. I like to ask the team questions and look for alignment on the answers. The questions are a summary of the day’s work: Who is the ultimate user of the product or feature? Under what conditions would a user engage with this product or feature? What pain point do they have that will be addressed by this product or feature? What triggers, internal motivations or external pressures are involved in them using the product or feature? What outcome does the user expect from the encounter with the product or feature?

It’s possible that the answers to these questions won’t be crystal clear to begin with, that’s okay. Discussing them and aligning the team around what needs to be done in the phases that follow is more important than concrete answers.

Phase two: diverge


The goal of this phase is to generate possibilities. This follows directly from the Understand phase, where our goal was to understand the terrain and identify the problem worth solving. For this phase to be effective I recommend having the same mindset that improv actors embrace: build on the previous person’s idea. Instead of judging ideas, we’ll find the best solutions when we have an open mind and encourage crazy ideas.

Mind mapping (20 mins)

We start with mind mapping to warm up creativity and prepare the group for the exercises that follow. This is an individual exercise, so each person will do their own mind map. The facilitator reads out the previous day’s problem statement and then sets the timer for the group to write down ideas.

Using a blank sheet of paper the participants write down ideas that might be solutions to the problem. Just like improv, by asking, “yes, and…,” each idea will generate another idea. Participants add each new idea to the previous idea with a connecting line. The result will look resemble an alien spider, with connected ideas radiating out from a starting point.

Crazy eights (30–40 mins)

In a design sprint, we approach ideation in layers. The mind mapping gets the creative ideas out of the participants’ heads, and then the Crazy Eights and How Might We exercises allow us to go deeper. The time limit allows participants to explore ideas but deliberately doesn’t give them enough time to over-analyze their solutions. The objective is to generate ideas, not eliminate them because of judgments like, “Oh, that’ll never work.”

To do the Crazy Eight exercise hand out blank paper and have each participant fold the paper in half, then half again. This will create eight panels (front and back) on the sheet of paper. Give each person eight minutes to sketch eight different solutions, about one minute for each. Quick and dirty sketches are perfect. Once everyone is done, repeat the exercise. Repetition reinforces the “yes, and…” mindset and pushes participants to come up with new ideas that they may have never considered in the first round.

How Might We (30 mins)

How Might We is an innovation exercise used by Google, Facebook, Procter & Gamble, and design studio IDEO. The question, “How might we?” is another extension of the improv idea and pushes participants to think about how they could bring their solutions to life.

To execute this exercise, ask participants to write answers to the question as it relates to each of their solutions. For example, “How might we hail a taxi using a person’s GPS location?” Answers should be a short sentence or a sketch. Tim Brown, CEO of IDEO, says the How Might We technique works best with ideas that are ambitious, yet also achievable. Brown says it doesn’t work as well with problems that are too broad.

Storyboarding (30 mins)

The storyboarding exercise is designed to take the ideas generated by Crazy Eights and expanded by the How Might We questions and develop them further. Each participant starts with a blank sheet and adds three Post-it notes down the side of the page. They should choose the most promising solutions to storyboard. Each Post-it note is one frame in the storyboard. The top note represents the current state of the customer; the second note is how the customer would experience the new solution; the bottom note shows the outcome created by the new experience.

Think of this as a storyboard for a short film. Each note should be used to draw the action. Use the space on the paper to the side of the Post-it to write a brief explanation. Each frame should be self-explanatory. If a participant needs to explain what’s going on in a frame, ask them to redraw the action. Once everyone is done, hang the storyboards up in the shared space.

Silent critique (30–40 mins)

Once the ideation exercises are complete, the group will shift gears from non-judgemental creativity to individual critique. The purpose of doing this silently is so all voices get expressed, not just the senior leaders or influencers.

Make sure all the storyboards are displayed on the wall, provide each person with several colored stickers (dot stickers) and ask participants to vote for the ideas they like best. Each person can use his or her entire allotment of stickers on one idea or distribute them in whatever way they see fit, even voting for their own ideas. The result will look a little like a series of heatmaps. The higher density dots indicate the most popular ideas.

Group critique (45 mins)

Don’t do a group critique until the silent critique is complete. The group critique is exactly that, a chance for the entire team to discuss the ideas on the storyboards. The facilitator will gather everyone around each storyboard and ask them what they like about it. It’s essential that everyone gets to share what they like about each idea. The emphasis should be on the positives. In the next phase, participants will have a chance to think about the negatives. Note takers will capture this qualitative feedback.

Retrospective (15–20 mins)

This activity will be essentially the same each day–circle up, discuss the day’s work, and plan ahead for the next day.


Phase three: Converge


The purpose of the Converge phase is to reduce the potential solutions to a single version that will be tested. To make this convergence possible, the facilitator needs to ensure the assumptions identified in the Understand phase are being considered. Only the idea that addresses the riskiest assumptions and is aligned with the problem statement should move to testing. Converging is hard work. The team will be responsible for choosing some ideas over others. This means making tough decisions, so let participants know in advance to be prepared to let go of some favorites.

Identify conflicts (20–30 mins)

The entire group will be involved in this exercise as they seek to identify storyboards that are so similar that they can be merged. Approaches that conflict shine light on what choices there are in solving the problem. Talk through the different approaches and decide which is the best to continue with.

Review assumptions (10 mins)

As mentioned, the alignment between a potential solution and the original assumptions is important. During each phase, the facilitator should remind the team of the assumptions. Gather the group around the assumptions board created on day one and discuss how the selected storyboard will address the assumptions. If the assumptions tests need to be adjusted or changed, do that now.

Assumption Test with… Validated if…
Customers want a shorter checkout process Prototype and interview Customers choose the prototype over the current checkout process
Customers understand the value of this feature Interview Customer can clearly articulate the value without prompting

Review parking lot or back burner board (10 mins)

If ideas were created in the Diverge stage that should be preserved, add them to the parking lot or back burner board. Record the best ideas and then move on. Don’t get too distracted by these sideline ideas.

Sketching (30–40 mins)

Once the reviews are complete, the teams will start with three quick rounds of sketching. Once participants have selected a single storyboard/idea to pursue from the Diverge phase, individuals must sketch out what the solution might be and then share it with 1–2 other people for feedback.

These small teams must then make tradeoffs to combine their solutions into one. Small teams repeat the process one more time by sharing solutions across the broader team and creating a single version of the solution. By stepping this out into a multi-pronged approach, participants are less likely to be overwhelmed by a single big decision. It also gives them a way to clearly articulate the reasons behind the design decisions they made earlier in the day.

Whiteboard the final storyboard (1–2 hours)

The final storyboard exercise is really important because it forms the foundation of the build and tests. The facilitator can lead this exercise by dividing the group into different roles. Some people will sketch while others rewrite the descriptions in detail. Don’t focus on design details, that will be the focus of the next phase.

The storyboard should be created in a way that the entire team can see it. The whiteboard is ideal. I’ve seen some cases where the facilitator does all the sketching at the whiteboard while the other members provide inputs. I don’t recommend this approach because it allows participants to sit back and watch the facilitator do most of the work.

Retrospective (15–20 mins)

Phase Four: Build (prototype)


In enterprise environments, I prefer calling this phase “Build” rather than “Prototype.” The reason is not all solutions will be prototypes in the traditional sense. Some of the solutions I’ve participated in have been things like sales scripts or service interaction models. “Build” is a more inclusive term that’s less intimidating for non-designers. However, if you’re reading this book, there’s a good chance you’ll be designing digital solutions, and in that case, “Prototype” should work just fine.

During this phase, the primary activity is going to be designing and creating something that can be tested. If the group includes designers, then use these skilled people to create the screens, pages, or features you’ll be testing. If you have developers on the team, you can also create HTML/CSS prototypes that will live on the web. The additional fidelity and interaction of a coded prototype means you’ll have a smoother user experience, but it’s not required.

If you don’t have developers on your design-sprint team, don’t panic. Paper prototypes are generally enough fidelity for testing your assumptions. The advantage of paper prototypes is they can be created quickly, inexpensively, and changes are often as simple as redrawing a screen.

InVision is made for prototyping and is my go-to tool for this exercise. Using the templates in InVision, even non-designers can create workflows using design outputs from applications like Sketch, Keynote, or Photoshop. Even images from a smartphone camera can serve as the screens or pages of the app or website you’re designing.

Phase five: Test


The final phase of the design sprint is to test the original assumptions, validate or invalidate the problem statement, and extract knowledge about customer’s preferences. The output will be the insights collected from customer or prospective-customer interviews.

While design sprints are structured to generate more qualitative than quantitative insights, both are still considered important. In the Understand phase, when we are formulating our problem statement we’re reviewing or collecting insights by discovering the problems customers are thinking about. In the Test phase, we want to validate (or invalidate) the problem statement. We do this by conducting just enough interviews that we can gather whether that problem is real or just someone’s perception.

The collective facts obtained from these interviews will enable you to make decisions based on objective observations. To do this you’ll need to recruit up to12 potential users and give them access to the prototype. Specifics of recruiting and testing are discussed in the sections below.

Recruiting for interviews

Bringing the users into the picture is often the most exciting part of the design sprint. Users are the fairy dust that you get to sprinkle on the design sprint because their feedback brings your prototypes to life. Once users start interacting with your prototypes, you’ll get the answers you were seeking. The fundamental research questions or problem statements you need to answer will tell you who you need to engage in the interviews. Douglas Ferguson suggests asking the following questions: Are you looking for a new or existing user? Are they people who fit your sprint target? Whom should we exclude?

On the other hand, the design sprint’s short time schedule means you will need to start recruit candidates before you even start the sprint. Recruiting interview candidates will be different for each design sprint so I encourage you to plan ahead. Recruiting several users in just a few days can also be challenging and stressful.

Without preparation, a team might find they have scheduled interviews with the wrong people. In a recent design sprint that was conducted to find solutions for LendingTree customers with low credit scores, the sprint team discovered that all the recruits had very high credit scores. As Ferguson cautions, “Be super careful that you are talking to the right people; take the time to prepare a proper screener.”

I’m not a fan of recruiting potential customers using Craigslist and the promise of $100 for their time. This approach attracts candidates who are more interested in earning $100 than they are in giving you feedback on a problem they care about.

Interviews (a few hours to an entire day)

The interview

Each team will have specific roles during the interview process. You’ll need at a minimum, one person to conduct the interview and one person to take notes, or be the observer.

Don’t expect good results if the interviewer is also expected to take notes. It’s difficult for an interviewer to be truly present if they are also trying to be a good notetaker. An interviewer should be focused on the interviewee and asking the right questions, while the notetaker will be taking objective notes about what they hear and see.

If you have a larger team you can create multiple interviewer and observer teams. Assign roles the day before during the Build phase so team members can prepare for what they need to do. The interviewer will prepare questions. Author and researcher, Steve Krug has created an extremely thorough list of scripts and suggested questions for interviewers.

The remote interview

I recommend you do all user interviews in real-time and, wherever possible, in person. However, when a remote interview is necessary, a video conference with screen-share functionality is best. This will allow the interviewer and notetaker to see the facial expressions and body language of the user.

Video conference tech also allows the interview team to record the conversation and review it again with the broader team. If an extended team or stakeholders will be joining via video conference, it’s necessary that they remain quiet and objective, and not interrupt the interviewer. Keep all feedback until the end when the user has left the room or video conference.

Don’t coach your candidates

As excited as you might be about your potential solutions, be careful not to sell your ideas to the interview candidates. It’s more important to get a neutral candidate than someone that is going to support your idea. You are not selling or pitching them anything.

It’s also important that you don’t coach them toward an answer you’re hoping to hear. If a user is struggling, don’t jump to their rescue and tell them what to do. Rather say something like, “It looks like you’re still thinking about the step, can you tell me what’s on your mind?”

This isn’t an exam

Take your time to make the interview candidates feel comfortable. Interviewees can often feel like they’re being tested or graded so explain to them that their honest feedback is the best feedback. There are no right or wrong answers, only their unfiltered responses are needed.

This works in reverse too. Interviewers may believe an answer is wrong because of their own personal perspectives or biases. If you’re the facilitator, remind interviewers that they must be mindful of their tone of voice, facial expressions, and responses to feedback. Turning up your nose at a user’s response because you don’t agree sends a strong message that you don’t approve.

Negative feedback is often the best feedback

Rejecting feedback, because it wasn’t what you expected or desired, isn’t going to help uncover answers. Remember, a design sprint is a process for generating understanding. If a user is struggling through a flow or says negative things about the solution, that’s new information you can use to improve your work.

Instead of defending your solution to the user or redirecting them along a path you’re interested in, ask questions like, “Tell me more about that” or “I’d love to know what you’re feeling” Steve Krug has once again provided an excellent list of questions that an interviewer can ask.

Reviewing the interview sessions

Your recollections will be freshest immediately after an interview. However, it’s also common for memories—even fresh ones—to be filtered or obscured in some fashion. That’s why it’s so important to review notes, video, audio and observers comments to fill in the perceptual blanks to which we’re all prone and gain a broader perspective on the interview.

Consider, too, that the more interviews you do, the more likely your brain will filter the memories in order of last-to-first. The most recent interviews will be clearer than the interviews conducted earlier in the day. This is called the availability bias and happens when we overestimate the likelihood of something happening because a similar event has either happened recently or because we feel very emotional about a previous similar event. This can easily be overcome by reviewing interview notes and video.

Final retrospective (15–30 mins)

Setting the mind to the phase

Jenny Gove, Kai Haley, and Marta Rey Babarro from Google talk about the evolution of sprints and how they have impacted teams across Google.

Each phase’s name describes what the team will be doing, but it doesn’t describe how participants should think. Below is a list of character roles that will help you and your team bring the right mindset to each phase of the design sprint.

  • Understand – think like a scientist
  • Diverge – think like an artist
  • Converge – think like a detective
  • Build (or prototype) – think like an architect
  • Test – think like a journalist

Thinking like a detective

Think of your team as a group of forensic experts sifting through the clues looking for evidence. The question you need to answer in the Understand phase is: “What’s going on here?” Seen through the lens of the customer, this might sound like: “Here’s a problem I would love to have a solution for.” In this phase, it’s important to stay away from questions and conversations about how a solution might be delivered or what form it might take. We’re not concerned with that at this stage. It’s first important to know whether there is a case that needs solving.

Thinking like an artist

Your customers probably don’t know what the solution should be, or that they even need it. All they know is they have a problem or a pain that they are currently putting up with. As a result, the customer is a great resource for understanding the problem, but not for trying to find the solution.

When you move to the Diverge stage, you’ll switch from an analytical mindset to a creative mindset. Your goal will be to create as many possible solutions as time allows. Embracing a sense of openness and a lack of judgment will help you get into the right frame of mind.

Thinking like a scientist

Converging on the best idea, or ideas, requires the puzzle-solving mentality of a scientist. During the previous phase, the team will have developed lots of possible solutions. The team will be piecing together the elements that work while discarding the ideas that don’t support our problem space. Don’t be afraid to put ideas aside if they don’t fit the best profile. Your Parking Lot or Burner Board is for capturing these ideas that may be worth exploring in a future design sprint or discovery activity.

Thinking like an architect

Having done the necessary detective work, your best idea will now have made it to the Build phase. As is often the case with building projects, we start by influencing its design, but then its mere existence has the effect of influencing how we behave. This is exactly what happens when we start architecting our prototypes. Our initial ideas become refined and improved when we see the product, service, or ideas in action. This is sometimes described as “thinking by doing.”

Thinking like a journalist

If “thinking by doing” describes the architect’s mindset, then the journalist mindset might be described as “thinking by questioning.” Approach the problem as if you will be expected to provide sources, evidence, and a clear storyline. Think through the who, what, when, where, why and how the mantra of journalism. Consider questions like: How did we get to this point? Who is this about and who does it appeal to? Why were those people the target of this story? What makes them and this subject so interesting?


About the Authors

Richard Banfield
CEO, Fresh Tilled Soil

Richard recently published Enterprise Design Sprints, a collaboration with InVision. Before that he published Product Leadership: How Top Product Managers Launch Awesome Products and Build Successful Teams, which he co-authored with Nate Walkingshaw and Martin Eriksson. He also authored Design Leadership and Design Sprint: A Practical Guidebook for Building Great Digital Products, which he c-oauthored with CTodd Lombardo and Trace Wax.

Richard is the CEO and co-founder of Fresh Tilled Soil, and under his leadership, Fresh Tilled Soil has delivered UX and product design to 700+ clients across the world. Clients include American Express, FedEx, Keurig, Intel, Harvard University, GE, Walgreens, BBVA, Shopify, Titleist, Citrix, and Genetech. His colorful life experience includes being an officer in the army and a dive master on the remote Islamic Republic of the Comoros.

Currently listening to: Loving J.S. Ondara’s Tales of America. This young, Kenyan born musician has a healthy dose of Dylan with the soul of Rodriguez.
Currently inspired by: Paul Klee inspires me every day. Transformed himself a dozen times through diligent introspection and a willingness to abandon old habits for new learning.
Cultural thing I’m lovin’: Loving all the Yuval Noah Harari books, and inspired by the work of Japanese architect Tetsuya Nakazono.