Skip to main content

Faculty Insights

Why Deep Learning May Hold the Key for Early-Stage Entrepreneurs

A new paper by WSB’s Hart Posen suggests that entrepreneurial entry and exit patterns may have as much to do with learning as they do behavioral bias

By Clare Becker

October 7, 2019

Professor Hart Posen speaking during a faculty panel
Professor Hart Posen speaks on entrepreneurship and other topics during a faculty panel at the Wisconsin School of Business. Photo by Paul L. Newby II

If you ever find yourself jotting down a creative thought or idea before you go to sleep or on the back of a napkin, you’re not alone. Many of us have them—and realistically, most of us won’t have the time or inclination to pursue them.

But for Hart Posen, a professor in the Department of Management and Human Resources and the Richard G. and Julie J. Diermeier Professor in Business at the Wisconsin School of Business, examining how and why entrepreneurs enter the market is as important as the actual innovations themselves.

In a recent paper published in Organization Science, Posen and his co-authors try to understand why so many entrepreneurial ventures end up failing. They find that efforts to enhance nascent entrepreneurs’ learning about the merits of their ideas may help circumvent some of the common mistakes entrepreneurs make early in their journey.

Bucking the standard view

Posen describes the commonly held view of idea generation like this: Entrepreneur A has an idea to start a company. She makes a decision about whether to invest, perhaps deciding to enter the market. If things do not go as well as she hoped, Entrepreneur A may eventually decide to exit the market.

Add to that two behavioral phenomenon the entrepreneurial literature classifies as major mistakes entrepreneurs make: excess entry (too many individuals entering the market when they shouldn’t) and delayed exit (individuals who stay in the market too long when they should fold).

Hart Posen
Hart Posen is the Richard G. and Julie J. Diermeier Professor in Business at the Wisconsin School of Business.

Traditionally, Posen says, these excess entry and delayed exit patterns have been attributed to overconfidence bias—when individuals are overly confident in the value of their ideas.

“The observation is always, ‘Why are they entering the market, when perhaps they should not? Well, they’re overconfident. Why are they staying in the market when they should have exited long ago? They’re clearly overconfident,’” he says. Posen argues that blaming these issues on behavioral biases such as overconfidence may be a mistake.

Entrepreneur B, on the other hand, is what Posen’s study posits. Entrepreneur B comes up with an idea while working a day job or separate venture. Entrepreneur B embarks on a learning process, maybe several months or a year, to ascertain if the idea is solid. At this point, Entrepreneur B decides whether to invest and enter the market. After entering, Entrepreneur B continues to entertain feedback on the venture, considering whether to stay or whether to exit.

This learning process is not well-represented in the academic literature on entrepreneurship, Posen explains. Overconfidence bias does exist with some of these exit and entry decisions, but Entrepreneur A’s journey is described as a static, one-size-fits-all trajectory. It makes more sense to think about the journey as a series of decisions entrepreneurs make at each juncture as they continue to learn more about the merits of their ideas.

“What we’re showing here is that you don’t need the presence of a behavioral bias to observe these patterns of excess entry and delayed exit. You can observe the same patterns if there’s a continuum of learning and decision-making, and sometimes, entrepreneurs don’t spend enough time and effort on the learning process.”

A deeper level of learning

Deep industry knowledge comes from the experience and insights gleaned from working inside an industry. Posen, who was an entrepreneur himself for over a decade before he entered graduate school, says he had deep knowledge from experience founding a national chain of discount stores, but lacked that same advantage when he entered the market with a frozen dessert venture. The latter was more complex than he ever could have guessed at the outset—from calibrating freezers to understanding food science and distribution—and he “didn’t even know what questions to ask.” He eventually decided to discontinue the venture.

“To understand why so many new ventures fail, we need to understand this process of learning over time. It starts with an idea and moves through phases, a pre-entry learning phase and a post-entry learning phase,” he says. “The challenges of learning and decision-making are immense since the signals, feedback, and information are often far from clear. The challenge also changes across these phases because feedback in the form of actual sales in the post-entry phase is less ambiguous than what one learns with prototypes and focus groups pre-entry.”

I see a lot of my own mistakes as an entrepreneur as coming from not recognizing the importance of learning dynamics. I always look back and see how little I knew. Maybe that’s why I do this. I want to understand what I didn’t know.

Professor Hart Posen

Deep industry knowledge may facilitate entrepreneurs’ learning, allowing them to move faster and more efficiently down the learning curve, Posen believes, and with probably fewer mistakes made when entering the market. When errors do occur, they’re not as likely to be fatal.

“I see a lot of my own mistakes as an entrepreneur as coming from not recognizing the importance of learning dynamics. I always look back and see how little I knew. Maybe that’s why I do this. I want to understand what I didn’t know.”

And deep knowledge in one industry can assist entrepreneurs in their next venture.

“I’m always asking students who want to be entrepreneurs, ‘Where did you work before?’ Because the evidence suggests that most ideas come out of entrepreneurs’ deep knowledge and experience in their industries or with their technologies,” Posen says.

“There is very interesting research on how the auto industry evolved to be centered in southern Michigan,” he explains. “The answer is that, for example, Henry Ford lived there and started Ford Motor Company and David Buick lived there and started Buick Motor Company. A number of people, many of whom already worked at Ford, Buick, and others, looked at what they were doing and said, ‘I’ve got a great idea about how to do this differently or better.’ Now they’ve got deep understanding and they’re moving through that learning process, right? They’re not waking up in bed at night with an epiphany about making cars; they’ve already spent a great deal of time conceiving and learning about the merits of their idea before they jumped in to start their own firms.”

Implications for policy-making

Posen’s findings hold implications for public policy-making, a chance for entrepreneurs to learn more about entering the field before diving in the deep end.

Governments at all levels seek to provide incentives and assistance that promote entrepreneurship. How can they be most helpful? If the driver of entrepreneurial mistakes is behavioral bias, then policy implications are somewhat limited—it is not easy to make someone unbiased. If, as Posen suggests, “mistakes are driven by the amount and quality of learning pre- and post-entry, then we can intervene completely differently,” he explains. “We can create incentives to engage in a longer period of pre-entry learning about the merits of entering the market, facilitating feedback along the way.”

Policy aside, Posen says the study also helps inform his students who want to be entrepreneurs. “I think seeing this as a learning process is very important. If you want to be an entrepreneur in a particular part of the economy, go work there and through deep and extensive learning efforts, your initial idea will become better, your decision about whether it is worthwhile or not to enter the market will also be better.”

 Read the paper: “The Impact of Learning and Overconfidence on Entrepreneurial Entry and Exit,” published in Organization Science.

 Hart Posen is the Richard G. and Julie J. Diermeier Professor in Business and a professor in the Department of Management and Human Resources at the Wisconsin School of Business.