In the year 2000, Nokia predicted a great surge of 3G cellular technology.
By 2002, the company believed there would be 300 million 3G handsets in operation, and all of them would be using location-based services, secure mobile payment systems, and a vast library of mobile applications. In anticipation of this expectation, Nokia invested substantial resources and energy into R&D, hoping to serve a booming demand for 3G phones. But the demand arrived late—not until 2008 did the 3G handset tally reach 300 million. Nokia’s phones weren’t the hold-up. The problem was that many of the other 3G technologies weren’t in place, so that the value proposition promised to consumers took years longer than expected to materialize. Nokia succeeded in making its handset, but still lost its bet because of other external pieces crucial to its success.
Ron Adner, the David T. McLaughlin D’54 T’55 Professor at Tuck, detailed Nokia’s innovation blind spots in his acclaimed 2012 book The Wide Lens: What Successful Innovators See That Others Miss. After reading the book, Adner’s colleague Dan Feiler, associate professor and Raether T’73 faculty fellow, who studies the psychology of judgment and decision making, realized their research had something in common. Both professors are interested in the challenges of interdependence as it relates to risk. Adner approaches it from a macro level, focusing on innovation strategy when multiple firms collaborate. Feiler approaches it from a micro level, studying how bias and overconfidence impact managers’ decision making.
Feiler was aware of research in behavioral psychology demonstrating a systematic overestimation when people mentally compute the likelihoods of conjunctive events. They suspected a different bias was affecting managers thinking through the probability of multiple partners successfully contributing to a shared goal. After all, when a company like Nokia invests a fortune in research and development, managers aren’t doing these calculations in their head. There had to be something else going on. So Feiler and Adner teamed up to figure out if there was an additional source of bias that the literature had not yet considered. As Feiler explains, “when it comes to conjunctive projects, merely thinking through to the likelihood of each needed part makes us feel more optimistic about the aggregate chance of success even when the aggregate chance is known. So it turns out that people feel better about an X% chance of overall project success when that X% is known to be the product of various parts that are all individually quite likely.”
The chief challenge of collaboration is that it creates a dependence on parties over which a firm has no control.
These findings have recently resulted in a paper by Adner and Feiler, published in the journal Organization Science, titled “Interdependence, Perception, and Investment Choices: An Experimental Approach to Decision Making in Innovation Ecosystems.”
Understanding interdependence in the business world is becoming more important, as collaboration among firms is growing in popularity. Why? “It’s because as technologies have evolved, collaboration has become easier,” Adner says. “And at the same time, we’re seeing greater pressure towards collaboration because it allows for more advanced and richer value creation: we can do things together that neither one of us can do alone.”
While the benefits of collaboration are easy to see—exciting new products—its dangers are often underappreciated. The chief challenge of collaboration is that it creates a dependence on parties over which a firm has no control. Nokia, for example, relied on software developers to build 3G apps, but it couldn’t direct their work. This means firms need to be shrewd about measuring the likelihoods that partnerships will coalesce towards a desired outcome. Gauging this type of probability is a two-step process: assessing the individual likelihood of each component, and then using those figures to calculate the likelihood of the overall project. For instance, in a six-party project where each party has a 75 percent chance of completion, compounding each 25 percent chance of failure results in an 18 percent overall chance of success.
In a series of five experiments around this basic theme, using business executives, managers, and undergraduates as subjects, Adner and Feiler uncovered a subconscious bias towards overestimating the chances that a multi-party project will succeed. In one experiment, they studied the decision to “green-light” a conjunctive project and whether it depends on how the risks are presented: separately for each part, or in aggregate for the whole project. It did matter. “Participants reported a higher willingness-to-pay for the opportunity when presented with separate probabilities than when presented with the aggregate probability,” they write.
We’re not saying this bias can’t be managed, but it can only be managed if you understand it’s there. Otherwise, your organization is at greater risk of having to live with the consequences of these biased decisions.
In another experiment, they found that simply exposing subjects to the higher individual probabilities generated an over-optimism in the overall project, compared to subjects who merely saw the overall probability alone. For example, an 18 percent chance of success, on its own, doesn’t seem very good. But the subjects’ assessment of that 18 percent chance brightened after seeing the higher likelihoods of the individual pieces. “After reviewing the relatively good chances for the individual parts, you get some spillover of that positivity into how people feel about the overall project, even when the aggregate chance of success is calculated,” Feiler says. “We call it an intuitive confidence spillover.” One potential solution to this bias is to shield the decision maker from the individual probabilities, so she can make a clear-eyed appraisal of the project as a whole.
“We’re not saying this bias can’t be managed,” Adner noted. “But it can only be managed if you understand it’s there. Otherwise, your organization is at greater risk of having to live with the consequences of these biased decisions.”
Adner and Feiler will continue their joint research in this area, exploring the new questions that this research raises. They discussed these questions with the MBA students in Adner’s Research-to-Practice Seminar, Strategy in Innovation Ecosystems, which Feiler joined for discussion of this paper. “The classroom discussion of the paper was really rich. The students reinvigorated our interest in one idea in particular—a bias towards overvaluing certainty for individual parts within conjunctive projects,” says Feiler. Adner continues, “We had done an early experiment along these lines, but had let it fall by the wayside. The classroom discussion brought it back into focus, and gave ideas for additional experiments, which could be the heart of a new paper. This is part of what makes the class so great—it completes the research-to-practice cycle by taking teaching-to-research.”
This article was published in print in the summer 2019 issue of Tuck Today.