April 14, 2015
The gap between lab coat and suit is famously large. Scientists are often lost in accounting spreadsheets and investors are often frustrated by scientific jargon.
Through running workshops with thousands of scientists and inventors, MaRS facilitators have found that this gap is often bridged with the theory of the lean startup. As theories of change, the lean startup process and the scientific process are remarkably similar. Lean entrepreneurs, like scientists, perform experiments to learn more. While scientists are learning about the world itself, entrepreneurs are learning about their products, customers and markets.
The lean startup movement essentially uses the scientific process to disprove assumptions until a product-market fit is found. Here are several principles from the scientific process that apply directly to the lean startup movement.
The facts about market size (or usability, or sales cycles) that allow entrepreneurs to have faith in their startups are often just assumptions. Entrepreneurs need to reframe these conclusions as hypotheses that must be tested. This reframing requires the entrepreneur to accept the possibility that he or she could be wrong, which is not necessarily an easy thing to do.
Karl Popper was the first to clearly define one of the hallmarks of a good scientific theory: falsifiability. It’s a fancy word that means that a hypothesis is only valuable if it can be proven wrong.
“People will like this app” is a weak hypothesis because it can’t clearly be proven false. It’s too wooly. “Twenty percent of people contacted will pay $1.99 for this app” is a better hypothesis because it can clearly be proven false.
Decide on a way to test your hypothesis as cheaply and as quickly as possible. When conducting A/B tests or cohort tests, entrepreneurs should only change one variable at a time. If your landing page has three changes on it and you lose half your traffic, you won’t know which of the three changes was responsible. Change one thing at a time. Measure. Repeat.
“The only problem with experiments is not when they go wrong, but when you can’t end them.” This is quote from Scott Berkun when he was at WordPress, but it would be equally at home in a science textbook. Experiments should have clear start and stop times and a clear go/no-go threshold of success. You also need to decide what metrics to measure before you start.
Philosopher Thomas Kuhn was famous for his theory of paradigm shifts. Older scientists often cling to old theories by ignoring data from younger scientists, until the young scientists gain more experience and take over. This is similar to Clayton Christensen’s theory of disruptive innovation. Large tech companies often ignore startups because they (initially) serve a different market segment. Market leaders can be suddenly thrown from their perch, seemingly overnight.
Was there data from your experiment you ignored because you didn’t like it? Successful scientists try to identify their biases and assumptions before they generate data. Entrepreneurs are similarly prey to confirmation bias, where they discount people who disagree with them and pay attention to the people who agree. In reality, they are both data points that should be treated with detached interest.
Scientific theories are changing all of the time and are never static. Science is never “done”; it is constantly evolving. Similarly, business models ebb and flow with markets and segments and should be constantly tested for integrity. Go where the data takes you and don’t become complacent (it won’t last long).