In the startup community, it’s common to hear variations of the mantra, ‘Ship early, ship often!’ Startups come up with a new idea, slim it down to a minimum shippable product, launch it, analyse the data, and then iterate. On the surface, this seems like an excellent way to build great products. But, having worked with many startups at Google Ventures, it has become clear that this methodology is riddled with problems:
– Testing hypotheses by shipping products is time-consuming and expensive. Even at a nimble startup, shipping a major product in a month is considered really fast. And inevitably, some will then fail.
– Launches are hard to abort. Once you invest time in engineering a solution, it is difficult to step back to reconsider a risky product rationally.
– Analysing post-launch data is a messy business. Even if you can decipher if a product is successful or not, it’s very difficult to interpret why that’s the case.
– Bad features are hard to roll back – at least a few users will embrace them. Even if you successfully remove them, code remnants will litter your work for years.
– Iteration is often de-prioritised in the face of new challenges, so people frequently leave products to fester instead of improving them.
At Google Ventures, we shortcut the typical ‘ship early, ship often’ cycle in a way that keeps the benefits and mitigates the problems: we skip over the build and launch phases and rely on prototyping with user research. In five days or less, a small team ideates, prototypes and tests significant hypotheses. At the end of a week, it is clear which ideas work, which need improvement, and which are a waste of design time. We call this a ‘design sprint’, and there are few keys to the process.
GET THE RIGHT PEOPLE
We try to get between five and 12 people in the room. They’re a combination of the people you’d expect (design, product, engineering) and people you might not, like your CEO, business development and customer service reps. Your business development person knows what it takes to close a deal, your frontline customer service people hear about inefficiencies in your product, and your CEO can see the big picture in the marketplace. If possible, we gather everyone in a secluded war room for the duration of the sprint so we can avoid distractions.
CREATE TIME PRESSURE
One of the reasons that startups establish a shipping culture is to create time pressure – it’s easier for people to rally towards a date. In a design sprint we try to create the same sense of urgency … on steroids. On Monday, we schedule five user studies to take place on Friday. At this point, we don’t even know what problem we’re going to tackle but a bomb is going to go off on Friday and the fuse is lit, so we have to hustle.
CHOOSE A SMART PROBLEM
With the fuse lit, it’s time to choose what problem to solve. We sketch out the user story like a comic strip that illustrates the product flow and look for big question marks like, ‘People will probably be confused by X’. Areas of great risk are also excellent ones to prototype – for example, ‘If people don’t comprehend our pricing structure, we’ll go out of business’. By the end of the first day, we have a nice meaty problem in hand.
Shortcut the typical ‘ship early, ship often’ cycle in a way that keeps the benefits and mitigates the problems: skip over the build and launch phases of the process, and rely on prototyping with user research instead
SKETCH THE RIGHT WAY
Now that we’ve identified a problem, we want to explore as many solutions as possible. We start on paper, since it’s a medium anyone can use. Everyone, including the non-designers, quietly sketches possible solutions. There is no group brainstorming. Loud voices do not prevail, and we avoid compromised group think solutions.
We start with very rough private sketching to keep people from zeroing in on their first idea. The final sketches are still rough, but detailed: no one draws like Da Vinci, but copywriting is realistic, user interfaces are plausible, and graphics are indicative. Multi-step sketches are encouraged. Drawing flows (‘The user does this, then that happens’) are better than isolated screens.
DECIDE THE RIGHT WAY
Teams commonly make decisions by arguing over the best solution. We avoid this because loud voices often win. Instead, each person considers each sketch and then votes by placing a small sticker next to the ideas they think are worth exploring in a prototype. After a couple of rounds of voting we have a heat map of the best ideas.
can easily fake interactions, and because it’s at a level of detail where you won’t waste time polishing the visual style to perfection. For mobile, we frequently use Flinto (flinto.com) or similar storyboarding-like tools. By the end of the day, we will often have created 10 to 20 screen ‘apps’. Always keep in mind you’re trying to suspend disbelief, so users are reacting, not commenting.
The user studies are the most critical component of the sprint process – but they aren’t sophisticated. You don’t need fancy equipment or a professional researcher to get a lot of value from a study. We schedule four to five people who fit the profile of potential customers, bring them to a quiet room at our office that is outfitted with a simple webcam, and have them click or tap through the prototype while talking out loud. At the same time, the other design sprinters observe from another room and take detailed notes. At the end of the day, we analyse our notes and look for patterns.
With the prototype and user study done, it is possible to take the lessons learned and build the minimum shippable product with clarity. Alternatively, if the sprint revealed large holes in the product direction we can cut our losses early and start over with a new hypothesis.
A lower-risk path to launch
Employing a prototyping mentality is all about mitigating risk. At small startups, building the wrong product wastes time on a short runway. At large companies, it costs money and can harm a brand. In both cases, it incurs significant opportunity cost. As a side effect, teams are happier working this way. More people get involved in the ideation phase, everyone gets exposure to real customers, and the team gets a rush from achieving an extraordinary amount in just a week. If you’re interested in running a design sprint within your business, our handbook should give you enough detail to try one yourself: gv.com/design-sprint.