Why do startups fail and can we create an AI formula to avoid failure? Interview with Tom Eisenmann, Professor at Harvard Business School


It’s a well-known fact that 90% of startups fail. We all hear lack of funding as one of the biggest reasons startups fail. But have you ever wondered if research has been done on why a startup fails? What are the structural reasons that lead to boot failure? What can startups do to avoid the road to death? These questions led me to interview Thomas R. Eisenmann, Howard H. Stevenson Professor of Business Administration at Harvard Business School, Peter O. Crisp President, Harvard Innovation Labs, and Faculty Co-Chair of the HBS Rock Center for Entrepreneurship.

Being a young entrepreneur, my goal in this research is to help other entrepreneurs avoid known pitfalls. That’s why I was intrigued by Professor Tom Eisenmann’s book “Why Startups Fail: A New Roadmap for Entrepreneurial Success”.

The article below is based on my interview with Professor Tom Eisenmann.

After reading about the six reasons startups fail outlined in Prof. Eisenmann’s book, I was curious to hear his thoughts on the top reason among those six. Professor Eisenmann was quick to name the main reason as diving too fast without studying the market for the product. He explained that many startups jump too fast to create a product without doing market analysis. This wastes capital and human resources for at least four months. As Professor Eisenmann explained, my mind raced to examples of several startups I’ve seen in Silicon Valley that eventually change direction. While I admire startups that are nimble and move their focus area adjacently or orthogonally, I agree with Professor Eisenmann that greater diligence in initial market research could mitigate a lot of suffering that startups and their investors go through when they don’t find a market-fitting product eventually. The professor called it the false start and found it analogous to how an athlete jumps too early to gain an advantage over his competitors but ends up losing the race.

During the discussion, we did not stop there on the subject of the false start. Professor Eisenmann looked at the human psychology behind the false start. He mentioned that given the natural impulse of entrepreneurs, speed is their motto. As entrepreneurs want to get started in building the business and engineers love designing products, it is natural for startups to conduct customer interviews and keep progressing with their startups. But that still doesn’t justify the lack of research startups have to do before defining what the product should look like.

Has Professor Eisenmann implemented his methods to prevent startup failure? The professor gave a humble response saying he failed to provide a diving plug to an entrepreneur whose business was on the brink of failure. But in general, Professor Eisenmann received positive feedback on the book from entrepreneurs who thanked him and made changes to their business plans.

Growing up in Silicon Valley, I’ve always heard of successful entrepreneurs as demigods. Therefore, I was intrigued by the professor’s book, where he calls bold vision as the reason for startup failure. So I asked him about this contradiction. The professor explained that entrepreneurship is about telling stories. It is therefore important for entrepreneurs to have a big vision in order to be able to convince strategic partners, investors and potential employees to believe in their vision. But the problem arises when the vision is felt so strongly by the entrepreneur that he fails to see the signals from the universe that the vision is flawed and needs to be rethought. It is a natural and defensive ego response of any human being to ignore such signals. We teach entrepreneurs to stay the course and believe in their point of view and approach to building a solution. Thus, being able to filter out the signals that suggest change becomes very difficult for an entrepreneur who naturally learns to be stubborn in his approach. Being persistent is very important for an entrepreneur, so he feels he should try one more feature to convince the client and offer more investors to seek funding despite all the rejections. But the problem becomes when the entrepreneur has gone too far. This is where the big vision becomes a cause of failure for the startup.

Professor Eisenmann explained how hyper-speed could be catastrophic for startups. Entrepreneurs love growth. Investors love growth. In fact, Y-Combinator says that growth is at the heart of a startup. Thus, it is possible to grow too quickly in a direction that may be chronically unprofitable. Unless the next wave of customers has the same needs as the first, it becomes more difficult to acquire them. You may need to market harder, cut prices, and do more to stay on the growth trajectory. You need to hire, stay in control, and scale. It’s all done in a startup that has none of that, to begin with. If you go too fast, you get into the flow where customers are less likely to stick around or raise awareness of the product. The foundations of the operation begin to shake. Corporate culture can take a serious decline. The first group of employees may believe in the long-term vision, but the next hires may not – conflict between “new guard” and “old guard”. The startup can go south very quickly, especially if you’ve burned through the initial cash. Before you know it, your startup is gone.

After talking with Professor Eisenmann about the causes of startup failure, I was curious if we could create a mathematical formula that could predict startup success using attributes of entrepreneurs and investors. Using econometric and statistical models, it seems quite feasible to predict a startup’s success by pouring in data and measuring factors like founders’ experience, target market, competitors, and more. Professor Eisenmann tried this. In the appendix to his book, he explains his survey of 470 startup founders of primarily software-based companies who launched startups between 2015 and 2018 and raised at least half a million dollars but no more than 3 million. He asked these founders 50 questions about who they were and how they handled the first year of their business. By following the companies through 2020, the Professor had a few years to see how the company was doing and whether the next round was a downturn or an upturn. Certain variables are predictive of success based on statistics. But his findings showed that the things people assume to be highly predictive of a startup’s success aren’t necessarily so, such as where the founders went to school, their MBAs, their gender, age or background. experience as a serial founder. None of these factors really moved the needle. Each factor appeared significant when considered in isolation. But when you put all of the founding attributes into a multivariate model, no single factor turned out to be hugely significant. The only statistically significant predictor was the number of late-stage rivals the startup faced. You don’t want to start a ride-sharing service if Uber and Lyft are out there. Overall, at best, you can only explain part of the variance in startup results by adding all the variables. The R-squared (reflecting the variance) is quite low.

Can artificial intelligence (AI) predict the success rate? Prof. Eisenmann doesn’t think so and explained that the problem with training AI models lies in the use of historical seed data. Historically, a considerable amount of startup data is collected from a certain type of founder due to gender bias or bias towards underrepresented minorities. AI software might recommend finding a 25-year-old South Asian, white, or East Asian computer science student as a founder. But that goes against other studies that have shown that older founders are more successful as entrepreneurs due to richer networks and experience. It takes huge sacrifices to be an entrepreneur. When you’re 20 and unmarried, you can probably afford to take more risks. But as a 45-year-old man with family responsibilities, you wouldn’t take the plunge unless the startup idea had legs. So, probably, the success of older entrepreneurs relies on appropriate risk aversion.

Do you think ethics plays a role in the failure of startups? Professor Eisenmann said “no doubt about it”. The slippery slope is always there as an entrepreneur. Entrepreneurs at every stage of the process must make ethical decisions about the representations they make. Usually they hardly sell anything, just a vision. So, with continued pressure, it’s tempting to distort the progress they’re making and punch above their weight. The pressure is all the greater when you are short of money. Therefore, lying becomes an occupational hazard. “Fake it till you make it” is often talked about in Silicon Valley. But there are ethical ways to fake it. People who coach entrepreneurship are much more sensitive to ethics today than they were before. College people teach entrepreneurial ethics.

All in all, I enjoyed learning about startup failure during the discussion with Professor Tom Eisenmann. While learning about a structural framework and deeper thought process, I felt his book did a great job of laying out the root reasons for startup failure. I hope this interview and the book will be useful to founders entering the fascinating world of entrepreneurship. Go big but think first!


Comments are closed.