by Jake Knapp
“This is a good discussion, but there’s still a lot to cover today. Let’s have the Decider make the call so we can move on.”
And:
“Let’s just trust the Decider on this one.”
Smaller details—such as design or wording—can be pushed off until Thursday:
“Let’s leave it up to whoever makes this part of the prototype tomorrow.”
If anyone, even the Decider him- or herself, starts to invent solutions on the spot, ask that person to wait until after the sprint to explore new ideas:
“It seems like we’re coming up with new ideas right now. These ideas are really interesting, and I think you should make note of them so they don’t get lost—but to get the sprint finished, we have to focus on the good ideas we already have.”
That last one is especially tough. Nobody loves stamping on inspiration, and those new ideas might appear stronger than the ones in your sketches. Remember that most ideas sound better in the abstract, so they may not be that good. But even if one of those new ideas is the best idea ever, you don’t have time to back up in the process.
Your winning sketches deserve a chance to be tested. If those new ideas and improvements are truly worthwhile, they’ll be there next week.
Thursday
On Wednesday, you and your team created a storyboard. On Thursday, you’ll adopt a “fake it” philosophy to turn that storyboard into a realistic prototype. In the next chapters, we’ll explain the mindset, strategy, and tools that make it possible to build that prototype in just seven hours.
13
Fake It
A square-jawed cowboy stands outside a saloon. He adjusts his hat and squints across the dusty street, where five men in black suits sit on horseback, rifles clutched in their hands. Farther down the street, the townspeople huddle near the general store. A tumbleweed blows by. Nobody speaks, but everybody knows: There’s about to be some trouble in this town.
If you’ve ever watched an old Western movie, you’re probably familiar with this scene. Good guys in white hats, bad guys in black, plenty of melodrama. The town is often the most realistic part of the film: clapboard buildings, wooden boardwalks, and saloons with swinging doors.
Of course, those Old West scenes were never quite as real as they appeared. Sometimes, the director found an existing location that looked about right: an abandoned ghost town or a picturesque Italian village. But most films were shot on a set on some Hollywood backlot. That saloon behind the cowboy? Just a façade—an exterior wall with nothing behind it.
It makes no difference to the audience. For the few minutes we see the town, we get lost in the story. It all appears real. Whether it’s a façade or a ghost town, the illusion works.
• • •
Thursday is about illusion. You’ve got an idea for a great solution. Instead of taking weeks, months, or, heck, even years building that solution, you’re going to fake it. In one day, you’ll make a prototype that appears real, just like that Old West façade. And on Friday, your customers—like a movie audience—will forget their surroundings and just react.
Façades are easier to build than you might think. Let’s say you’re working on a project that will take a hundred days. And let’s say that 90 percent real is real enough to test. Simple math says it’ll take ninety days to get to that 90 percent real level, so you should be ready to test in about three months. But we’ve found that if you only build a façade, you can get to 90 percent on day one.
“Whoa, pardner,” you’re thinking. As of Thursday morning, you’ll have nothing but whiteboard drawings and paper sketches. Do we expect you to create a realistic prototype in just one day? Isn’t that impossible? It would be impossible, except that you’ve already done the difficult part on Monday, Tuesday, and Wednesday. The storyboard removes all guesswork about what to include. The solution sketches are packed with specific text and details. And you have the perfect team, with all the right skills to create your prototype.
Sure, you could take a longer time to build a more perfect prototype—but doing so would only slow down the learning process. That may not matter if you’re on the right path, but let’s face it—not every idea is a winner. Whether you’re taking a risk on a bold idea, or you’re just not sure, it’s better to find out early. Wasting time on the wrong thing is a major bummer.
But perhaps the biggest problem is that the longer you spend working on something—whether it’s a prototype or a real product—the more attached you’ll become, and the less likely you’ll be to take negative test results to heart. After one day, you’re receptive to feedback. After three months, you’re committed.
At the beginning, you’re in the sweet spot of all these charts (which, to be fair, we made up). You’re not attached to your ideas yet, so if they don’t test well, you’ll be flexible enough to fix or cut them. You’re in the perfect position to take advantage of that fast curve to 90 percent real, if you limit yourselves to building a façade. No plumbing, no wiring, no structural engineering. Just a façade.
The prototype mindset
Building a façade may be uncomfortable for you and your team. To prototype your solution, you’ll need a temporary change of philosophy: from perfect to just enough, from long-term quality to temporary simulation. We call this philosophy the “prototype mindset,” and it’s made up of four simple principles.
1. You Can Prototype Anything
This statement might sound corny, but here it is. You have to believe. If you go into Thursday with optimism and a conviction that there is some way to prototype and test your product, you will find a way. In the next chapter, we’ll talk about specific methods for prototyping hardware, software, and services. Those methods may work for you, or you may have to be resourceful and invent your own. But if you stay optimistic and adopt the prototype mindset, there is almost always a way.
2. Prototypes Are Disposable
Don’t prototype anything you aren’t willing to throw away. Remember: This solution might not work. So don’t give in to the temptation of spending a few days or weeks getting your prototype ready. You’ll have diminishing returns on that extra work, and all the while, you’ll be falling deeper in love with a solution that could turn out to be a loser.
3. Build Just Enough to Learn, but Not More
The prototype is meant to answer questions, so keep it focused. You don’t need a fully functional product—you just need a real-looking façade to which customers can react.
4. The Prototype Must Appear Real
To get trustworthy results in your test on Friday, you can’t ask your customers to use their imaginations. You’ve got to show them something realistic. If you do, their reactions will be genuine.
How real is real enough? When you test your prototype on Friday, you’ll want your customers to react naturally and honestly. Show them something flimsy—a “paper prototype” made up of drawings, or a simplified wireframe of your design—and the illusion will break.
Once the illusion is broken, customers switch into feedback mode. They’ll try to be helpful and think up suggestions. In Friday’s test, customer reactions are solid gold, but their feedback is worth pennies on the dollar.
Goldilocks quality
This distinction between feedback and reaction is crucial. You want to create a prototype that evokes honest reactions from your customers. You want it to be as real as possible, while sticking to your one-day timeline. As our partner Daniel Burka says, the ideal prototype should be “Goldilocks quality.” If the quality is too low, people won’t believe the prototype is a real product. If the quality is too high, you’ll be working all night and you won’t finish. You need Goldilocks quality. Not too high, not too low, but just right.
Of course, “Goldilocks quality” looks different for each product. Next, we’ll show you some examples: five teams, prototyping everything from iPad apps to a medical clinic. As you read their stories, you’ll see how each team applies “Goldilocks quality” and the prototyp
e mindset to their unique challenge. We’ll begin with FitStar, a company that had to build an elaborate prototype—and do it without the most important person.
FITSTAR
Question:
How can we explain a new kind of fitness software?
Format:
Simulated App Store and iPad app.
Tools:
Keynote (presentation software), acting, iPhone videos, iPad.
“People are getting the wrong idea. They download the app. They try it. But they think it’s something else.”
Mike Maser sat back in a plastic chair in our San Francisco office. The brim of his baseball cap was shredded from years of wear, and his plaid lumberjack shirt was faded. Not the look we’d expected for a guy who palled around with professional athletes and spent half his time at video shoots in Los Angeles.
Mike was the CEO of a startup called FitStar. In 2013 and 2014, FitStar’s iPad apps would win Apple’s coveted Best of the Year awards. In the App Store, FitStar would rank among the top of the charts in the health category, and in 2015, the fitness technology company Fitbit would acquire the startup.
But this afternoon was before all of that, back in 2012, when nobody—except for Mike and his cofounder, Dave Grijalva—quite knew what FitStar was all about. GV had invested in the company, and Mike and Dave spent a week with us. The goal of our sprint: Find a better way to explain their new app.
Mike and Dave had a vision for bringing personal fitness training to the masses. Personal trainers are expensive and tricky to fit into busy schedules. “Most people just can’t do it,” said Mike. Thanks to Mike’s connections in the entertainment industry, FitStar had brought on the best personal trainer they could think of: Tony Gonzalez, a fitness guru and star player in the National Football League. They’d filmed hundreds of hours of Tony giving instruction for different exercises, at all kinds of ability levels. And Dave—a programmer with a background building video games—had created algorithms that stitched together video clips of Tony into custom workout sessions.
They had created an automated personal trainer. It could tailor each workout so it was suited to the customer’s fitness level and goals. As he or she progressed, the workouts would adapt and get more difficult. The app had just launched, but FitStar was waiting to promote it until they knew customers could understand how it worked.
So far, people were confused. The message about customization and personalized training wasn’t getting through. Most of their early customers thought it was just a workout video, like the old VHS tapes and DVDs hawked on television. “Once you have that mental model, it’s tough to break,” said Dave.
By Wednesday afternoon of their sprint, Mike and Dave had a slate of promising ideas for improving the initial experience of the iPad app, everything from better descriptions in the App Store to new animations between exercises.
Unfortunately, Mike’s favorite idea appeared impossible to prototype. He wanted to record videos of Tony Gonzalez asking the customer questions as he or she set up the app. In real life, when you start working with a personal trainer, the trainer is right there, talking with you. Mike was betting that a back-and-forth conversation would give FitStar the opportunity to explain—in Tony’s own words— how customized the software could be.
But Tony wasn’t in the sprint with us. He was on the other side of the country, playing football for the Atlanta Falcons. Plus, it would be impossible to build a new version of the iPad app in just one day. And even if we could, we would never be able to get it launched to the App Store in time for Friday’s test. We only had one day before the customers came in, and a seemingly impossible prototype to build.
But all we had to do was fake it. On Thursday morning, we divvied up the prototyping tasks. Dave grabbed his laptop and started writing the script for Tony’s video introduction. Mike volunteered to stand in for Tony as the on-screen talent. He put on some workout clothes, set up an iPhone to record video, and read lines from the script.
What about the software? FitStar couldn’t reprogram and re-release their iPad app in time for the test. But we didn’t need a real app. We just needed something that looked like a real app. We remembered that you can run Keynote (Apple’s presentation software, like PowerPoint) on an iPad. A slideshow running full-screen would look just like an app. It could even play videos.
We divided the storyboard into sections—one for each of us to work on. Then, using both the storyboard and the solution sketches as our blueprint, we built the prototype, screen by screen. We found a template kit online with realistic iPad buttons and icons that we could copy over. We added photos and illustrations from the actual FitStar app to make it more realistic. We dropped the videos from Mike and Dave into the slides.
To complete the illusion, we added screenshots from the iPad App Store to the beginning of the slideshow. These screenshots showed FitStar’s app in the health category, and even showed the installation process. When we’d finished making all the slides, John took on the job of “stitcher”: going through the whole deck, making sure everything looked consistent from slide to slide.
By the end of the day, it looked just like real software—even though there was no software at all. FitStar’s prototype was just like one of those Old West façades: The illusion only worked for a few minutes, from a certain angle. But that was enough to answer Mike and Dave’s big question for the sprint: Can we better explain our app to new customers? After Thursday, FitStar was ready for their test.
Some solutions worked. The videos of Mike explaining the software were effective. Right away, customers could explain the app in their own words (“Kind of an automated personal trainer”), and they were willing to pay for it (“Can I sign up right now?”). Other solutions failed. After the introductory conversation, there was a clip of Dave wearing a lab coat. He introduced himself as “Doctor Algo Rhythm” and went on to explain how the software was programmed. But by that point, customers understood (“I get it.”), and they were ready to exercise. They found Doctor Algo unnecessary and even obnoxious (no knock to Dave’s acting).
For FitStar, success in the market depended on quality. But in their sprint, success depended only on being real enough to answer their key questions. They got the information they needed to identify the right solutions—and shut down the wrong ones—with a prototype that only took seven hours to build.
SLACK
Question:
What’s the best way to explain Slack to non-tech customers?
Format:
Two competing websites with interactive software.
Tools:
Keynote, InVision (prototyping software), the real Slack software, and some acting.
Slack had two competing ideas to prototype. First was “The Tenacious Tour,” a step-by-step explanation of the software. Just as with Blue Bottle Coffee, this idea could be faked with a series of slides that looked like a website. No sweat.
But the other idea, “Bot Team,” was tricky. It involved a team of computer-controlled “bots” who would send messages back and forth to one another, and even respond to messages typed by the user. To be realistic, the bots should respond to a variety of questions and comments from the customer, an experience impossible to fake with slides.
Merci had the solution: We could pretend to be computer-controlled characters ourselves. During the test, we’d send messages to the user, and—like bots—reply in a not-too-intelligent way. Of course, if the idea proved successful, Slack would have to write computer software to control the bots. They could never have teams of people sending messages to every customer who visited their website. They’d need a staff of thousands, or millions! But for our test, for just five customers, it would work.
FOUNDATION MEDICINE
Question:
What essential information do oncologists need to make treatment decisions?
Format:
Paper medical report with first page only.
Tools:
Keynote, real
istic test data, printer.
Earlier, you met Flatiron Health, a company dealing with the complexities of getting cancer patients into clinical trials. Boston-based Foundation Medicine, another GV investment, was working on a different problem in cancer care: using DNA analysis to suggest possible treatments for patients.
In 2012, Foundation Medicine had developed a test called FoundationOne. The company’s lab could analyze a single tissue sample and give doctors a report on every known genomic alteration associated with cancer, along with a list of potential treatments.
The test was groundbreaking. FoundationOne’s diagnostics provided a wealth of information, often leading to surprising treatment options. But all that information presented a challenge: It could be overwhelming, even to expert oncologists. In those early days, the FoundationOne results were delivered on paper, and the team at Foundation Medicine was determined to make those sheets of paper as easy to understand as possible. So they ran a sprint with us to try out some new ideas.
The team decided to focus on the front page of the paper report. It was, of course, the first thing a doctor would see when she reviewed the test results. But if the doctor was in a hurry—and oncologists usually are—it might be the only page she had time to process. Foundation Medicine wanted to deliver as much information as possible on the top sheet.
In our sprint with Foundation Medicine, we had three competing ideas for the test report. To implement these ideas would require months of work in the lab and a serious quality-assurance effort. After all, medical reports have to be 100 percent accurate. But for our prototype, we only needed to learn which approach was most promising. We didn’t need to meet the same accuracy standards that applied to the real report. And we didn’t need to alter anything about the lab analysis itself. All of that could come later. For now, our question was about those crucial minutes as the oncologist reviews the front page.