Why Entrepreneurs Should Think Like Scientists

There’s no surefire strategy for building a thriving start-up. Entrepreneurs have launched successful businesses using frameworks like the Business Model Canvas, the Balanced Scorecard, and more. But in a recent study involving young ventures in Europe, one technique consistently boosted performance. Firms employing the scientific method—the centuries-old discipline of formulating, testing, and tweaking hypotheses—generated more revenue than companies in a control group did and were more likely to pivot away from unviable ideas, a routine necessity for early-stage start-ups. And among the top revenue generators in the experiment, the differences were especially pronounced. Among the top 25%, start-ups using the scientific method generated €28,000 more than the control companies did over the course of the experiment, on average, and among the top 5%, they made €492,000 more.

“We started this research by asking ourselves, Why do so many start-ups fail?” says Alfonso Gambardella, a professor of corporate management at Bocconi University who is one of the study’s authors. “After discussing the problem, we realized that many entrepreneurs were not very good at making predictions about their business. And this hurt them when it was time to make important decisions.”

For the study, Gambardella and colleagues recruited 759 start-ups by advertising on social media, in newsletters and magazines, and at events. The entrepreneurs running them received training in various strategic frameworks as well as in evidence-gathering techniques such as qualitative interviews, surveys, and A/B testing. The key difference between the control group and the treatment group was that the treatment group was also taught to employ the scientific method when applying the frameworks and techniques. That required founders to develop theories about elements of their proposed businesses and then test them through rigorous experiments. The control group’s founders, on the other hand, were free to apply any frameworks and techniques in the way they wanted.

Over a span of several months, large teams of research assistants collected data on the companies and interviewed the founders regularly by phone. One of the start-ups they followed was Mimoto, an electric-moped sharing service. It set up an experiment to test its hypothesis that college students hustling between classes would rather ride mopeds than walk. After placing dozens of mopeds near an urban campus, Mimoto realized that young people didn’t use them any more than older people did. Rather, people of all ages—especially those who had unpredictable commuting patterns—picked up the mopeds as they passed through the campus. The entrepreneurs went back to the theory stage and, on the basis of the data gathered during the experiment, determined that young professionals, particularly lawyers, were the most likely to need mopeds, because they were constantly rushing to and from meetings with clients. The company then pivoted to focus on that demographic.

Mimoto’s experiences support Gambardella’s contention that where founders start out is not really that important because they will test and test to determine what is possible. Experimentation allows them to adjust, change, or terminate their approaches much sooner than they would have without using the scientific method. Pivoting is common for early-stage start-ups. YouTube started as a dating site before pivoting to video. PayPal was originally an internet security company. The key to a successful pivot is focusing not on your original theories but on the answers you receive from your experiments. They should provide insight into customer demand, industry pain points, and ultimately, viability.

“Some of the firms pivoted during our study, and that’s not necessarily a bad thing because most pivoted very early,” Gambardella says.

Embracing the scientific method isn’t difficult. Entrepreneurs can easily apply it by following these steps:

First, state your main hypothesis. Mimoto’s founder believed that college kids would want to rent mopeds to get to classes faster. Only by articulating that hypothesis clearly and then observing how mopeds placed around campus were used, and by whom, could the entrepreneur have realized that he was wrong.

Begin with the idea that your intuition tells you is most plausible. “You have to start somewhere,” Gambardella says. “But you should have other potential theories you can use if you determine that your first theory is incorrect.”

Define the critical dependencies underlying the hypotheses and investigate them. Leonardo Del Vecchio, the founder of Luxottica, a maker of eyeglasses, did this when he came up with a theory about how to grow the company (although it wasn’t a start-up at the time). He posited that eyeglasses could be more than a means of vision correction—they could be a form of fashion. He believed that people would pay extra for stylish glasses and that some people would purchase more than one pair and create a wardrobe of eyewear. Before going to market with his hypothesis, he wanted to explore two uncertainties: Could Luxottica even produce fashionable eyewear? And could it move from a market controlled by independent retailers to a more direct-to-consumer approach? Del Vecchio believed both were possible.

To resolve those uncertainties, Luxottica partnered with fashion houses, such as Armani, to produce designs, which immediately increased demand. And the company started to invest in direct marketing to test its ability to sell to consumers rather than to optometrists. Today EssilorLuxottica, as it’s currently known, is the world’s largest eyewear company, valued at more than $95 billion.

You can do unstructured brainstorming to identify the uncertainties your hypothesis involves, but you should focus on the top few—let’s say three—factors in your experiments.

Refine your theories and test them again—but more broadly. When Mimoto found that its theory wasn’t feasible, it developed a new theory involving young professionals. But it didn’t assume the new theory would work in the real world just because it had worked in an experiment. Mimoto tested it by placing dozens of mopeds in urban environments. It reapplied the scientific method to its new theory until it was confident it had found a viable solution.

“Entrepreneurs who adopt the scientific method are not visionaries. They’re aware that their theories may be wrong. And when they’re wrong, they modify their theories, which they test again. If no theories are successful, they terminate the start-up,” says Gambardella.

Even if you end up shutting down a business, don’t worry. “Failing a few times is a recipe for future success,” Gambardella says.

About the research: “A Scientific Approach to Entrepreneurial Decision-Making: Large-Scale Replication and Extension,” by Arnaldo Camuffo et al. (Strategic Management Journal, 2024)

“We Would Have Ended Up With a Product That Wasn’t Viable”

Cosimo Cecchini is a cofounder of Osense, a start-up focused on technology for sustainability that participated in Gambardella’s research. He recently spoke with HBR about the study and what he learned using the scientific method. Edited excerpts of that conversation follow.

Why did you join the research project?

Our first start-up idea was for a peer-to-peer product-rental service. Our mission was to reduce consumption to improve sustainability. Companies would be forced to produce fewer products, and their carbon footprint would shrink. Joining the study was a great opportunity for us to learn how to test our hypothesis, which basically assumed people would use this platform and that it would help reduce climate change.

But you were wrong?

We used the scientific method to try to validate our idea that people would rather rent than buy products and that the owners of the products would make money from the rentals. To do that, we conducted field interviews asking whether people would be inclined to reuse other people’s goods and whether people had goods they wanted to rent out. We conducted approximately 25 interviews, but after only five, we realized this wasn’t going to work.

What would have happened if you hadn’t used the scientific method?

We probably would have developed our idea about peer-to-peer rentals further and built a prototype. It would have been costly and time-consuming. And we would have ended up with a product that wasn’t viable.

The only positive feedback we received during the interviews was that people were interested in products that promote sustainability. So we came up with a new hypothesis around that idea.

Which was?

A platform to encourage the adoption of electric vehicles by persuading rental-car companies to offer them and connecting those companies with renters. It was a completely different idea, but it revolved around sustainability, so we decided to try it. We used the same approach, a field study, to determine whether the new idea had any viability. We spoke with rental car companies and their customers about their pain points. We asked whether a platform like the one we proposed would generate additional interest in electric cars. Only 20% of the people we spoke with offered any positive feedback. We were hoping for 60%.

Was that superfrustrating?

Yes, but we were happy to learn these lessons about our hypotheses early on. Almost immediately after giving up on the second idea, I stumbled upon a McKinsey report that said limiting Scope 3 carbon emissions—which aren’t produced directly by a company but come indirectly from things in its value chain—was the biggest challenge for sustainability-minded corporations. When we investigated Scope 3 emissions, we learned that employee commutes, business travel, and logistics are massive contributors. We asked ourselves, “Can we develop a tracking system that would give companies real-time access to this data?”

We interviewed sustainability managers. After three field interviews, we knew we were on the right path. After 10 interviews, nine of the conversations had been entirely positive.

What were the next steps that you took?

We had to prototype the idea and test it to determine whether it was feasible in a real-world setting. We ran a six-month pilot with a company that immediately became our first client once the pilot was complete.

And is your idea still viable?

Yes, but we will continue to test it and test it as we progress.

Copyright 2024 Harvard Business School Publishing Corporation. Distributed by The New York Times Syndicate.

Explore AAPL Membership benefits.

Recommended Resource

Fast Forward Thinking: 40 Rules for Entrepreneurs and Investors in Medical, Science, and Biotech by Luis Pareras, MD, PhD.