How do academic ideas make it into practice? When faculty members consult with industry, the path between research and practice is direct.
By Kelly J. Andrews
“Innovation is a new match between a solution and need, delivered to the customer,” said Karl T. Ulrich, the CIBC Professor and professor of operations and information management. “It begins with a sensed opportunity—a hypothesis that value can be created.”
The occasion was the Impact Conference on Managing Innovation Systems held on October 27, 2006. Ulrich, the conference co-director, was addressing a high-powered group of researchers and executives convened at the top of Jon M. Huntsman Hall. While Ulrich was referring to new product development cycles, his words apply equally well to the research conducted by Wharton’s faculty.
Faculty research is also a match between a need—the questions raised by industry—and a solution. The professors here work with more than 1,000 businesses, government agencies, and nonprofit organizations around the world, in cluding most of the Fortune 500 companies.
The answers they find become business practice around the world.
In this article, Wharton Alumni Magazine looks at a few examples of the real-world impact of close research collaborations between professors and corporate partners in October 2006.
Research Centers as Hubs for Industry Collaboration
“As a research center we are guided by the concerns of our 17 member companies,” said Marketing Professor George Day, co-director of the Impact Conference and of the Mack Center for Technological Innovation, which produced it. “We created a list of priorities by consulting with them, asking what are their problems. We are currently funding 36 faculty members throughout the School, who do the research, and then communicate the results through conferences.”
The Mack Center is one of 24 research centers and initiatives at Wharton. Each is a direct connection to industry through the participation of supporting members, as well as a nexus of interdisciplinary thought. Their work generates courses, academic programs, community outreach, published research, and partnerships among academics, government, and industry.
Explained Day, “Wharton is fortunate to have outstanding academic departments with vertical expertise. The research centers are the horizontal linkages that tie them together.”
“We are all active consultants and teach on programs for executives, so we have a good sense of industry practices and issues,” he continued. “We wouldn’t succeed if we didn’t deliver value to our partner companies.”
Each year, each Wharton center conducts its own conferences and members meetings, about a dozen of which are designated Impact Conferences for their broader application. The 2006-2007 series kicked off with “Inclusion in the Business of Sports” in September 2006 and will continue with “International Financial Integration” in May 2007. In between, the conferences will cover topics from environmental management to household investment portfolios to urban economic development.
New Models for Innovations at Merck
The varied research that comes out of the centers highlighted in these conferences is funded by member companies and rooted in the business problems they face. Sometimes it is even produced in concert with members, as faculty consulting and research projects.
That’s how Christian Terwiesch, Wharton associate professor of operations and information management (OPIM), ended up at the Impact Conference co-presenting with Brian Kelly, a 1997 alumnus of the Penn Engineering Executive Masters in Technology Management program and vice president, business integration, at Merck & Co.
The project began three years ago through a conversation between Kelly and Terwiesch’s colleague and frequent co-author, Karl Ulrich. The professor asked a question about Merck’s development pipeline that Kelly couldn’t answer, and the collaboration began.
The relationship, which included a research grant and access to data from Merck, has yielded fascinating results, soon to be published in Management Science. The paper by OPIM doctoral candidate Karan Girotra, Terwiesch, and Ulrich shows that when large, publicly traded pharmaceutical firms experience a failure in Phase III drug development (the last stage before official drug approval), there is an average decline in firm value of $551.9 million. Substantial fixed costs ? time, manpower, and resources—mean that a late-stage failure of a compound can be devastating to a company.
Case in point from a Merck competitor: In December 2006, when Pfizer experienced late-stage failure for its most promising heart drug, the company’s valuation plummeted.
Merck faced a similar situation in the early 2000s. After experiencing wave after wave of scientific and commercial success in the 1980s and 1990s, the pipeline began to dry up. Success had created overconfidence that manifested in inflated estimates of probability of success (POS).
In the boom decades, “most things we tried, worked,” said Kelly. “The drugs had an impact on patients and were commercially successful, and this reinforced a bullish attitude. When we struggled in the 2000s, we realized that we were way too optimistic.”
For drugs in various stages of development, general odds for reaching the market are well known: 20% for drugs in Phase I, 30% for Phase II, 60% for Phase III, and 80% for new drug applications (NDAs).
However, those general odds are aggregates. Increasing probability of success for each compound at each phase would have multiplying effect on Merck’s valuation.
“If you ask most people what’s the difference between a 60% and 70% chance of success, they can’t tell you,” explained Terwiesch. “But if you look at the expected payoff, there is actually a 50% difference between the two. It’s a big implication. So increasing the accuracy of projections is critical.”
Too often in business, he said, when someone says a project has an 80% probability of success, it’s less of a prediction than an indication of how much they want to do the project. And unless the accuracy of predictions is tracked, it can’t be improved.
Terwiesch and Kelly reported that their methodology produced a substantial increase to the accuracy of Merck’s predictions.
Of course, predictions alone don’t create drugs that work ? they merely improve the efficiency of the pipeline.
“A key premise is that the safety and efficacy and biopharmaceutic properties of any compound are intrinsic,” said Kelly. “There is nothing we can do to change those properties. The role of development is to reveal them. Success is based on revealing those properties as soon as possible to assess whether to take the compound to the next stage of testing. Appropriate POS estimates are critical to inform how much it is worth to reveal the properties.”
Doing so efficiently means that sometimes potentially effective compounds are rejected.
“If I have an imperfect but sufficiently predictive test after Phase One, we’re going to use it,” said Kelly. “If you want to improve POS, you must be willing to throw out some compounds that are good with an imperfect test if that means that the drugs that make it through have a higher chance of succeeding. It will drop your cost structure.”
While choosing between leads can be a thorny problem in product development, that’s less of a concern for Merck. In pharmaceuticals, a queue is beneficial.
At earliest stage, 5,000 to 10,000 compounds are evaluated against a “target”—the clinical indication. In the last stage, Product Development, one to three compounds are evaluated and zero to two compounds emerge as NDAs. “Some years the successful products are zero, and all the money developed to research and development goes to nothing. We can make good decisions, and have bad outcomes.”
Effective R&D strategy can make a big difference. When pharmaceutical companies have developed a backup—a redundant compound that could potentially serve the same indication—they can decrease the impact of a late-stage failure. But whether that’s the right call or not depends on the POS for each of the drugs.
“If I can only launch three things and five things are competing and several are redundant, how do I choose?” said Kelly. “Do I launch three of the same or diversify? If you can’t estimate dependent POS, you have no ability to make decisions.”
“One of the lessons here,” said Terwiesch, “is that you can’t value a drug in a vacuum.” The value depends in large part upon the company’s strategy and portfolio. “Net present value won’t do it,” he continued. Instead, you need to look at how the drug fits into the overall pipeline. He and his colleagues are working on models that will value the whole pipeline, taking into consideration the balance of risks involved as well as the utilization of knowledge resources.
The outcomes of the Merck/Wharton collaboration are already finding a wider audience. The research will be included in Ulrich and Terwiesch’s upcoming book, Managing Innovation Systems. And earlier in 2006 Wharton launched a week-long Strategic R&D Management program for executives under academic director Terwiesch, who developed the curriculum in collaboration with Ulrich and INSEAD colleagues. The course has already been offered through Executive Education at Wharton, INSEAD, and Merck itself.
Putting Theory to Work at Dupont
Executive Education is not only a product of industry collaboration—it can also be the originator. The next pair of copresenters at the Impact Conference on innovation actually began their relationship in an ExecEd classroom.
Ian C. MacMillan, director of Wharton’s Sol C. Snider Entrepreneurial Center and The Dhirubhai Ambani Professor of Innovation and Entrepreneurship, has worked with John Ranieri, now vice president and general manager of DuPont Bio-Based Materials since 1996. The two met when Ranieri was a visiting fellow as part of Wharton’s Advanced Management Program, the intensive five-week Executive Education program offered to highly qualified senior managers.
Ranieri has utilized and developed entrepreneurial management structures that more effectively develop and lead transformative growth opportunities in an environment where information, speed, and adaptability are of critical importance. As an example, a real options framework developed by Ian MacMillan was applied to more effectively assess and maximize a portfolio’s valuation. The idea is that a small upfront commitment for the initial stages of new product development is a “real option,” similar to a financial option. Investors have a choice to make larger investments in later stages if the technology successfully passes early tests. The system was further elaborated for greater adaptability, to maximize capital leverage, and to develop partnerships that are accretive in value.
Ranieri further allocated his investments in an “opportunity portfolio”—a concept introduced by MacMillan and Columbia’s Rita Gunther McGrath, GrW’93, in their book, The Entrepreneurial Mindset. Picture a matrix with two axes: technological uncertainty and market uncertainty. Projects with low uncertainty in both dimensions are Enhancements—new products that generate incremental revenue by improving or finding new applications for existing products. Those that have high uncertainty in both dimensions are Stepping Stones—high risk but with the potential to transform both technology and the marketplace.
Organizing investments in this manner and using options- based management and financial systems assures that both winners are identified in a timely fashion, while projects that are not market-valid or technologically feasible are stopped before they become too costly.
MacMillan and McGrath advocate establishing a strategy, structuring an opportunity portfolio that supports it, then employing adaptive execution (i.e., learning along the way through the use of real options).
In 2002 Ranieri, who holds a PhD in medical sciences, joined the unit of DuPont charged with creating greener energy sources and materials using biology, chemistry, materials science, and engineering. Two years earlier, DuPont Chairman & CEO Chad Holliday had set a target for DuPont to derive 25 percent of its revenue from non-depletable resources by 2010—a lofty goal.
Ranieri mapped DuPont’s projects into opportunity matrices, and the product development cycle for Bio-Based products was aligned to create a portfolio with allocations in each sector of the opportunity matrix. By understanding the overall mix, the business created new product targets that were both high value-in-use and reduced the associated environmental impact.
Instead of evaluating projects across the board, each was evaluated within its sector. Instead of embarking on massive but risky projects, DuPont used relatively small investments to gain knowledge, reduce uncertainty, and target capital investments intelligently.
Five years later, DuPont Bio-Based Materials is reaping the benefits with a dozen Bio-Based Materials business opportunities ready for commercialization in the next five to seven years.
“Real options framework changes the dynamics of the team and the questions asked,” said Ranieri. “For example, here’s a question that was not obvious several years ago in the biofuels market: ¡®Is there something else we could make that is superior and could be transformative? How can we have both value-added products and reduce the environmental footprint at the same time?’ It turned out that these qualities weren’t mutually exclusive. We opened up new large-market opportunities with the technology base, and as we learned, we found surprises along the way that weren’t factored in the original valuations. We won both ways, feeding success and stopping failure simultaneously.”
As an example, the answers to those questions resulted in significant new product opportunities—an integrated process that produces cellulosic ethanol from parts of the corn that were formerly waste products and partnered with Broin, and a further partnership with BP to develop biobutanol, which has advantaged performance as a fuel compared to ethanol.
“Being able to ask the right questions at the front-end of innovation is not obvious nor easy,” he said. “But when you can more effectively learn and adapt, that’s how you get the right answers that create significant value and transform markets.”
How a Vanguard Partnership Creates a Unique Knowledge
In October 2006, the Pension Research Council (PRC) published a new working paper on trading behavior in pension plans. Just another academic research paper? Not in this case—for it arose out of a unique multi-year partnership forged between the Wharton Pension Research Council and Vanguard Group, an investment management company and one of the Center’s Senior Partners.
This partnership gave PRC Director Olivia Mitchell, International Foundation of Employee Benefit Plans Professor and professor of insurance and risk management, and Takeshi Yamaguchi, a doctoral student in insurance and risk management, an opportunity to conduct real-world empirical research on the behavior of 401(k) plan participants. It also gave them two co-authors—Gary R. Mottola, a researcher at the Vanguard Center for Retirement Research with a PhD in social psychology, and Stephen P. Utkus, WG’84, the founding director and principal of the Vanguard Center whose deep interest in investor decision-making began during his years at Wharton.
While Vanguard was a long-time member of PRC’s 24-company advisory board, a ground-breaking research collaboration grew out of an Impact Conference on behavioral finance, which both Mitchell and Utkus co-organized in 2003.
“Our conference looked at the psychological and socio- logical, as well as economic determinants on behavior,” said Mitchell. “Yet coming out of the conference, we recognize that little in behavioral finance focused on real-world investment decision-making in retirement plans. There was information from other fields—not pensions—that people traded too much and were unduly influenced by what their coworkers told them around the water cooler—not the determinants that financial planners would even consider in the decision-making process. We wanted to see how these apply to pensions.”
Vanguard was the perfect partner. “Research collaborations with industry depend on an individual who wants to get involved and who champions the project,” Mitchell explained. “Steve was unique in understanding how academic research applied to his company.”
The Vanguard Center for Retirement Research was created in 2001 and has been led by Utkus since its inception. It was established as part of the Institutional Investor Group at Vanguard overseen by managing director Bill McNabb, WG’83.
“In the early days of formulating our research agenda, Bill and I were talking about the question of 401(k) trading,” stated Utkus. “We knew that only a few participants traded, but some did so actively, and we were wondering whether they were actually making money or possibly undermining themselves and their retirement security by trading excessively. Over time I attended several of Olivia’s PRC conferences and also became involved with the PRC advisory board.”
When the 2003 behavioral finance conference revealed a gap in the research, the researchers of Vanguard and PRC decided to pool research efforts to look at 401(k) trading.
Academic Research with Practical Application
A series of working papers have resulted from the collaboration—with more to come. For example, the first paper, for which Mitchell served as lead author, formed a profile of 401(k) participants and their trading behavior: “We find that most 401(k) plan participants are characterized by profound inertia. Almost all participants (80%) initiate no trades, and an additional 11% makes only a single trade, in a two-year period.” They found that the most active traders are affluent, older, with higher incomes, and longer job tenure who use the Internet for trades, hold a larger number of investment options, and are more likely to hold active equity funds rather than index or lifecycle funds.
Most recently, the October 2006 paper elaborated with a surprising result, finding that the risk-adjusted returns of traders are no different than those of nontraders. Even more intriguing, the researchers found contradictory effects from trading. As they concluded: “Certain types of trading such as periodic rebalancing are beneficial, while high-turnover trading is costly. Interestingly, those who hold only balanced or lifecycle funds, whom we call passive rebalancers, earn the highest riskadjusted returns.”
The research collaboration between the Pension Research Council and Vanguard is multi-faceted. A related study with Mitchell, Utkus, and Wharton doctoral student Tongxuan (Stella) Yang, looked at 401(k) plan design. It resulted in a 2005 paper published in Restructuring Retirement Risks by Oxford University Press. The team was trying to measure the effectiveness of employer matching in encouraging employees to contribute to defined contribution plans.
Explained Mitchell. “We were interested in the question of whether and how companies can help employees save for retirement. Employers often provide matches in the 401(k) setting but lack a clear idea of what impact the matches might have. For instance, when there’s a larger match.”
She continued, “Do more people join the plan, and do they contribute more?
To examine this question, they needed firm-specific data.
The reason is that tax law constrains how much workers can contribute to their pensions, with complicated rules that are linked to what other employees in their company do. For instance, when low-paid employees don’t join the plan, or don’t save much, that limits how much other employees can save. So we wanted to know what everyone in the plan was doing, to understand how these interactive constraints work and how companies can encourage retirement saving more effectively. Working with Vanguard helped us get a much better picture of what real-world employees and employers do.”
The results were surprising—and a little depressing. The authors concluded that company matches do induce people to contribute but the effects are small. In the typical 401(k) plan, only 10% of non-highly-compensated workers are induced to save more by match incentives. The authors suggest that automatic employer contributions, regardless of employee participation, can be a more cost-effective and influential means of boosting retirement preparedness.
“This is just the tip of the iceberg—we have more papers coming out,” Mitchell said.
Two and a half years into the project, Utkus said they’ll keep going as long as they have questions. He explained that creating new knowledge is part of Vanguard’s culture. The mission is to help plan sponsors and individual investors make better decisions, and the results are widely applicable. After all, who doesn’t want to retire in comfort?
“When we do a formal working paper with Olivia or Takeshi, a lot of the math or statistics will be lost on our clients because they don’t have PhDs,” Utkus said. “They don’t need to. But the introduction, the conclusions, the topline results will be very practical. We’re doing high-quality, independent academic research, but on the other hand it has results people can relate to.”
Describing her own criteria for getting involved in any project, Mitchell insisted, “I’m a researcher first and foremost, so my condition for getting involved in any project is that it be publishable. We take pains to make sure that we retain publication rights. We work with whoever we’re working to make sure we have the right interpretation and caught the nuances that matter.”
Bristol-Myers Squibbs Looks at New Product Diffusion
Often the process through which arcane research questions result in accepted business practice is more convoluted than the partner research at Merck, DuPont, and Vanguard. First a concept is picked up by an innovative, influential individual who takes a graduate class, attends a conference, or reads an academic journal.
The individual puts the idea into action, and successful results create believers. Management teams seek to replicate the results across departments—and some team members even take the idea to new employers. As the influence spreads, winning concepts are identified by market research firms, which disseminate the idea from client to client, and by the business press, which writes about the trends. The final stage of acceptance is when the idea becomes a standard business practice, a metric demanded by directors or a function embedded in commercial software.
That’s what happened with Christophe Van den Bulte, an associate professor of marketing who has extensively researched social networks and the power of influencers.
He experienced that power himself recently. One afternoon he was contacted by a director at McKinsey, who wanted to discuss an unpublished working paper Van den Bulte had posted on his faculty web page. Two weeks later he had an email from Boston Consulting Group, and not long after that, from Arthur D. Little. The link to the unpublished paper had been found by influencers at each company.
“It’s a trickledown process,” he said. “Influential people with technical knowledge like to think at high level of abstraction, so they read academic journals and paper. Then they interpret what they find and pass on the conclusion to their peers and staffs.”
It’s not a coincidence that Van den Bulte both researches social networks and benefits from them.
Marketers at pharmaceutical company Bristol-Myers Squibb knew that opinion leaders can make or break the commercial success of a new drug. What they wanted to know was how to identify those leaders to deploy their representatives effectively rather than blanketing the entire medical community, and they hired pharmceutical marketing specialists Rivermark to help.
The citations in Malcolm Gladwell’s bestseller The Tipping Point led to Everett Rogers’ scholarly classic, Diffusions of Innovations. Those references led to University of Southern California public health professor Thomas Valente, who said he understood physician networks but not marketing. He pointed Rivermark to Van den Bulte.
Working with Rivermark and Valente, Van den Bulte began a small-scale analysis of physician networks in San Francisco. Using survey data from doctors nominating their most trusted contacts, he mapped connections between physicians.
Some of the physicians appeared to be the hub of wheels with many spokes—they were connected to other doctors who trusted them. Some physicians were isolated, and others were cross-connectors—individuals who linked up two networks that otherwise seemed to be almost separate.
Some of the hubs were unsurprising— they were renowned leaders at highly reputed university medical centers, and Bristol-Myers Squibb had long established relationships with them. But several highly connected physicians in one of the networks were not even on the radar for Bristol-Myers Squibb. And the researchers were surprised to discover how the two parallel networks were divided: by ethnicity. In one network, the physicians had Western surnames, and in the other, Asian ones. While ethnic enclaves are quite obvious in social life, the way that those social norms transfer to professional life was only obvious in retrospect.
In the next step Van den Bulte and his fellow researchers, including Wharton marketing colleague Raghuram Iyengar, are analyzing two separate sets of data. One, they are combining the network data with databases containing information on sales calls by pharma reps, on physicians’ prescription volume for the previous generation drug in this treatment category, and on the timing when each doctor wrote his or her first prescription for the new Bristol-Myers Squibb drug.
Two, they are overlaying demographic and professional information—where the doctors attended medical school, how long have they been in practice, whether they have a publication record—to find personal characteristics that may be used in other markets to predict the structure of the social network, including who the influential physicians are.
With the first phase completed in October 2006, the study is being extended to more cities in the U.S. and China to see whether the results hold up.
How research Becomes Practice
So again, how does the topic of Van den Bulte’s research— new product diffusion and social networks—relate to the spread of academic discoveries to business practices? Van den Bulte supplies a few examples from his own experience, recalling how Tom Valente, who recommended him for the Bristol-Myers Squibb job, knew him.
While still a doctoral student, Van den Bulte read a halfcentury- old sociology study on the diffusion of a new drug. He thought it would be interesting to re-analyze the data using modern techniques. A professor with whom Van den Bulte was taking a class at the time had just come back from a conference where, he happened to remember, someone had made a presentation using those very same old data. Van den Bulte checked the conference program and found his name: Thomas Valente. Voila—a cross connection.
In another example of how networks lead to the spread of academic research, Van den Bulte presented his ideas at the annual business-to-business marketing conference run by his former faculty adviser at Penn State’s Institute for the Study of Business Markets. In attendance was Dominique Hanssens, executive director of the Marketing Science Institute, a Cambridge, MA industry-sponsored nonprofit organization that supports academic research to develop and disseminate marketing knowledge. Hanssens drafted Van den Bulte to write the book on social network marketing for its Relevant Knowledge line, a series of accessible books that interpret academic theory for business practitioners.
The book’s publication will continue the dissemination process that is already in motion. In September 2006, Knowledge@Wharton published an article about his work, allowing Van den Bulte’s research to reach the biweekly journal’s worldwide mass audience of 500,000 subscribers across English, Spanish, Chinese, Indian, and Portuguese editions.
Having his ideas reach both academic and professional audiences is important to him. Echoing Mitchell, Van den Bulte said, “If I can’t publish my results, I’m not interested in consulting. A condition for my working on this that we could use it for research purposes.”
Fortunately, the firm found that reasonable. The manager in charge of the new product’s launch told Van den Bulte he wasn’t worried about exposing the results to competitors because academic research takes so long. “By the time the paper is out, either we will already be successful with the product, and we won’t care,” the product manager joked. “Or, the product will have failed, in which case we really won’t care.”
Getting Involved with Wharton Research
The Wharton Partnership is the Wharton School’s program for fostering industry/academic collaboration. The Partnership engages with more than 200 corporations and foundations.
The Partnership provides member organizations with a single entry point to the School. A relationship manager assists Partners in identifying and connecting with faculty who are researching topics of specific interest to them. Companies collaborating with research centers have early access to findings.
For more information, visit partnership.wharton.upenn.edu, call 215.898.5070, or e-mail firstname.lastname@example.org.