In the world of entrepreneurial incubation, design thinking, a user-centered way to conceive and create a successful product, is often compared and contrasted with the lean startup approach, which is more engineering-based and quantitative. The two methods are far from mutually exclusive, however, as both seek to effectively serve customers’ needs through a systematic, low-risk path to innovating in the face of uncertainty.
To explore the places where design thinking and lean startup intersect, we organized a conversation between two faculty members who help run the Stanford Graduate School of Business Startup Garage, where students design and test new ventures. Stefanos Zenios, the Stanford GSB Investment Group of Santa Barbara Professor of Entrepreneurship, joined lecturer Matthew Glickman, founder of BabyCenter and Merced Systems to discuss the science of startups.
Stefanos Zenios: Every startup is a series of hypotheses — about who’s a customer, what makes your product or service attractive to these customers, and so on. Lean startup provides a rigorous framework that you use to prove or disprove as many of these hypotheses as possible at as low a cost as possible. An interesting question is how do you generate the hypothesis? If you have not already identified the user need through your own pain point — having experienced the need firsthand — how would you discover that need? Design thinking gives you a methodology to do that. There’s an interplay. For example, in design thinking you develop a prototype that you use to get feedback — that’s very qualitative — and lean startup makes it more rigorous, so you don’t end up convincing yourself that the feedback is positive feedback.
Matt Glickman: I echo that. We start the class with design thinking, which is more inquiry-based and open-ended. It forces you to put yourself in the customer’s shoes. BabyCenter, which is still the leading site for pregnant women and new parents, was started by two guys without kids — so we needed to go out and try to understand what it’s like to be a pregnant woman. We talked to lots of people and started to intuit that the real problem was that they don’t know what’s coming up next in their pregnancy and their needs change frequently.There was no target customer who said, “Boy, wouldn’t it be great if I could get highly personalized week-by-week information!” But by developing intimacy with the customers, we were able to come up with the insight that with the internet you can personalize down to the week of pregnancy and give people exactly what they need to know. That ended up being the core insight that we built a whole business around. We developed a set of features, and it turned out that baby-naming and personalized information — two of the 20 things we tried — were the ones that really moved the needle. You want to build a “minimal viable product,” see what’s working, and double down on those.
Zenios: Matt makes a couple great points. Defining the customer problem comes from patient, sustained interactions with the customer over time, because you don’t know when that insight is going to come. When DoorDash started, the innovators didn’t know what problem they wanted to address, only that they wanted to help small business owners. [DoorDash, a Startup Garage success, is an online service that delivers meals directly from restaurants to customers. It was launched in 2013 by four Stanford students.] The first weekend of the class they spoke to 60 business owners of many kinds. One was a lady who had a small shop that was making macaroons. Her macaroons were extremely popular, and by talking to her and her customers, the team found out that the customers wanted to be able to place the order and have it delivered to their offices. That became the DoorDash idea — except in the end, they focused not on macaroons but restaurants.
Glickman: A lot of people come to the class with anxiety that they don’t have an idea yet. But we show that you have to go down a lot of alleys to see what underlying problem you’re trying to solve. And you do this experientially, sometimes by just picking a group you want to serve and learning what their problems are. The greatest correlation to success in the class is with how many interviews they do. MBA students are uncomfortable with it at first — they’d rather analyze competitors and work on spreadsheets and pricing. But in all startups, there’s a combination of thinking and doing. It’s like the Roosevelt quote [points to an inspirational Teddy Roosevelt poster in Zenios’ office]: You have to be in the arena, getting dust and sweat on yourself.
Zenios: For many students, the mind-set is “I want to make this hypothesis work,” whereas we want to teach the mind-set, “I want to discover if this hypothesis is not valid — and if it’s not valid, why it’s not valid.” And again, this is where the interplay between lean startup and design thinking is valuable. We had a class team whose original hypothesis was to put educational materials on the web that would teach people how to cook diabetic-healthy meals. The team’s first experiment was what we call a “concierge” version of that business — walking the customer through a planned service in person, before developing what will ultimately be a software implementation of that service. So the students ran an ad on Craigslist, announcing that that they were offering a class to diabetic patients on how to cook healthy meals, and at the end they’d give them a package of supplies for cooking for the next week. They did the training and gave them the package, and they asked, “Would you be willing to sign up a week from today to get another box of supplies?” “No.” “Why not?” “I like the social aspect of this, so I want to come back for additional recipes and meeting with these people.”
So the hypothesis changed because they had delivered the product manually — and through that personal connection with the diabetic patients, they heard, “No, I prefer something else.” The concept of the “concierge minimal viable product”— a product with enough features to gather meaningful feedback — was developed as way to discover a solution that’s less expensive than writing all the code for an actual software product. And from a design-thinking perspective, you’re the concierge, and when you deliver the service, you get to know your customer.
Glickman: All entrepreneurs say they’re customer-friendly, but businesses don’t usually get to know the customer this well. One idea that’s counterintuitive is to put in the minimum amount of effort to get the maximum learning. If you want to test delivering food to homes, how can you do that without spending three weeks actually delivering food to homes? One way is to build a web page to see if people will click on it. An even simpler way is to place keyword ads around “delivery.” As an entrepreneur, you repeat that process again and again, so doing it more cheaply allows you to do more iterations.
Zenios: This goes back to another element of lean startup: the notion of the pivot. If the data invalidates your hypothesis, you keep some elements fixed and change other elements. But how do you know what to pivot to? The guiding principle is deeper customer understanding, like in the example with the diabetes team. So that’s how design thinking and lean startup interact: The empathy you gain through design thinking helps you identify possible pivots.
Glickman: A common pivot we see is around health care: Students want to do something for patients, and as they develop their idea, they realize that it’s the insurance companies that pay for this service, so they need to pivot one part of the business by thinking about the needs of who’s paying. Sometimes they’ll end up deciding that they’re not interested in selling to health plans or that they don’t even want to be entrepreneurs, and those are important discoveries. They’ve still learned a systematic innovation process that they can apply elsewhere. We do see a lot of nonprofit teams, which is great, and I think there are applications in government, too. As an example, I would love for a team to look at ride-sharing — like what Lyft and Uber have done — and see how you use that to help solve public transportation problems in San Francisco. Perhaps you can integrate ride-sharing with bus systems.
Zenios: This process for gaining deep customer understanding helps large companies as well, because they thrive on processes. For example, Intuit has identified a small team of innovation catalysts — they go on “innovation safaris” with customers to do observation and prototyping and testing. The team then uses the feedback from these safaris to decide which features and which products to design and launch. Since starting to do this, the company has increased the number of new features added to products 100-fold per year. Citrix does something similar, integrating lean startup with design thinking. When I give these examples to executives, they say, “Thanks, you gave me a template.” It’s not creating an R&D lab — it’s a small group of people with day jobs who have the freedom to go on these innovation safaris.