CFOs are all too familiar with the experience: someone at the company has a Big Idea for an innovative new product or service, or has identified the next great growth opportunity. Money is earmarked and a project is launched with great fanfare. But two years later, what looked dazzling in a PowerPoint deck now looks ill-conceived if not completely doomed. Yet ego, corporate politics, and a desire to recoup some kind of return on investment almost guarantees that good money will be thrown after bad.
Rita Gunther McGrath, professor of business strategy at Columbia University, says there is a better way. In fact, she co-wrote the book on it (Discovery-Driven Growth: A Breakthrough Process to Reduce Risk and Seize Opportunity, Harvard Business Press, 2009). Her basic message sounds simple enough: innovative growth can be attained by establishing and systematically testing project assumptions at progressively more challenging checkpoints, learning from the outcomes, and changing direction as appropriate. "If the results live up to your assumptions, you release more funds until you get to the next checkpoint," McGrath says. "If they don't, you alter the project or kill it. The important thing is that you're embedding learning into the planning process."
When McGrath explained the approach at CFO's Core Concerns conference in Baltimore this summer, "it resonated with me," says Gregg Olson, senior vice president and CFO of Trail Blazers Inc., which owns the Portland Trail Blazers NBA basketball team and the team's related merchandising enterprises. "We had just launched this new product — live, online streaming of our games — and had made certain financial assumptions. Unfortunately, we had no template to test the accuracy or efficacy of these assumptions. Her methodology would have given us more confidence around pricing."
Olson is far from alone among CFOs seeking a more systematic way to fund growth strategies. Many companies are knee deep in cash and equity capital (see "Time to Get Off Your Cash?" July/August) and would like to put it to work. But in a new, more risk-averse climate, the fear of failure is palpable. McGrath presents an alternative — "fail fast and fail cheap."
"If you fail fast you can take the risk of failing more," agrees Brian Kimmel, CFO of the National Association of Convenience Stores. "By establishing assumptions and then checking them along the way to see if they're right or wrong, you can kill the project before it breaks the bank. You've now preserved capital for the next growth initiative."
That's not how traditional ventures are funded. "Once a project acquires a life of its own, people are afraid to kill it," says Kimmel. "They've got too much skin in the game."
Or, these days, too little. "CFOs are balking at proposals that might, in fact, drive growth, because they don't see enough evidence that the idea will work," says Kevin Bolen, a partner at Innosight, a consulting firm that leverages McGrath's ideas in engagements. "But if someone says, 'Can I have $10,000 to test these assumptions I have about this interesting idea?' a great project that otherwise might not get funded gets off the ground." (As one example, see "And One that Worked" at the end of this article.)
"The dilemma that CFOs face is that you have to make investments in new initiatives with less information than you would like," says Paul Huck, senior vice president and CFO of Air Products and Chemicals Inc., an $8.4 billion international seller of gases and chemicals for industrial applications that has applied McGrath's ideas to its growth ventures. "By using discovery-driven planning, you look at the available options and the decisions you make along the initiative's journey." McGrath adds that, "A CFO's role is to architect rich discussions around what has been learned and what that implies for the next stage of investment."
Innosight client Procter & Gamble followed this "test and learn" approach as it developed Align, a product for the relief of irritable bowel syndrome. "The team first had to determine if their formula worked to relieve symptoms, since there's no sense spending money on naming or packaging [a product] until its efficacy is proven," Bolen says.
When the efficacy was established, the team addressed the next "deal-killing" issue: Would consumers follow the necessary 30-day regimen? "Limited trials with doctors and online sales validated this," Bolen says. "The team continued to test critical assumptions, learn more [it found, for example, that a planned tie-in to Metamucil, a well-known product, would not resonate well with users of the new product, and dropped the idea], and redirect the approach as needed." The product is now available at retail chains nationally.
One important attribute of discovery-driven planning is that it provides companies with a way to overcome a common pitfall of most corporate cultures: the tendency for champions of ideas to insist that they are right and that success hinges on upping the ante. "My definition of learning is that you don't know something until you do," says McGrath. "And you won't learn until you have a structure that permits the flow and acquisition of knowledge" one step at a time.
Or, as Gerald J. Kochanski, CFO and treasurer of Nephros Inc., a privately held manufacturer of medical equipment, puts it, "Stuff happens, but few companies have a process in place to learn from it. Instead, they base everything on the premise that if you develop a well-defined recipe, all you need to do is put in the right ingredients and it will come out as planned. Life rarely happens that way, so why should a multi-million-dollar venture?"
Then, why don't companies cook up a better way? "Part of it is ego, some of it is arrogance, and the rest is the political nature of organizations," Kochanski says. "People are afraid they'll get their necks snapped off for saying, 'We got to this checkpoint and things didn't turn out the way we'd hoped.'"
Pulling the plug is never easy, but it's exactly what Air Products and Chemicals did to a planned venture in the automotive-coatings market, once it applied discovery-driven principles. "We needed to test how big the market would be before we'd invest in it, and this gave us a template for doing that," says Ron Pierantozzi, the company's former director of new-business development. "The licensing investment was significant and we didn't want to risk it until we were sure. We made some assumptions about the market and our projected revenues and then tested them. They failed, and we quickly killed the project."
Getting the most from a new discipline, however, requires, well, discipline. While Air Products CFO Huck praises McGrath's methodology for requiring project sponsors to provide "more of the information needed to make informed judgments about investments in new projects than conventional tools do," others note that some companies may not be ready to make such a leap. "I don't think many organizations have the discipline to document their assumptions and then review them at periodic intervals," says Gary Cokins, manager of performance management solutions at SAS.
But more companies may now be motivated to try. When the Trail Blazers were assessing streaming video, "We had no data to draw from, nothing to compare our expectations to," Olson says. "Had we established checkpoints, we could have tested our assumptions and adjusted as needed, instead of simply setting a price that seemed reasonable." Now that he has a year's worth of data to draw on, Olson plans to use McGrath's method going forward.
McGrath sums it up like this: "It pays to design a relatively expensive test" that can validate, redirect, or reveal the futility of a much more expensive investment.
Russ Banham is a contributing editor of CFO.
Three that Failed…
Big bets placed by big companies that proved to be big disappointments. Could "discovery-driven growth" have changed the outcome?
1. Revlon's Vital Radiance line of cosmetics for older women. The problem, the company didn't market the program appropriately, failing, for example, to take into account the fact that older women "don't like to be reminded of [their age] at the cosmetics counter," says Rita Gunther McGrath, professor of business strategy at Columbia University.
2. Michelin's PAX run-flat tire was "a great idea," McGrath says, "except that the company didn't see the implications of a requirement that the undercarriage of the car be redesigned, that customers would have to take their cars to a specially equipped service station to get the tires changed at a cost of more than $1,000 each, and that the tires made the cars heavier, thus requiring more gas."
3. AOL's $850 million acquisition of social-networking site Bebo. AOL "didn't realize how difficult it is to make money via ad placements on social networks," McGrath says. "It also failed to appreciate that Bebo's traffic growth was flattening, and it didn't give enough consideration to the possibility that Bebo's most talented employees would make a hasty exit postacquisition."
…And One that Worked
Sometimes a picture is worth 1,000 square feet of retail space.
Innosight partner Kevin Bolen says that one key aspect of discovery-driven growth is to "be open to surprises and then learn from them." He cites a client in the apparel business that wanted to improve the product-selection process. The traditional approach would be to come up with an idea, develop samples, and then do a retail trial — an expensive proposition. Instead, "we did 'shopalongs' to understand customers' frustration with the shopping process. We then brainstormed 10 potential solutions and tested them with photo mock-ups to gauge customers' initial reactions."
Two of the solutions were well received, so the project moved to the next checkpoint, which entailed using some empty office space for several days, where the company set up a mock retail environment to show prototypes to a variety of consumers. Alterations were made based on those responses, which gave the client the confidence to go to the next checkpoint — the traditional retail trial. So the usual starting point became the third step in a lower-cost, low-risk model that took only an extra two weeks and, had it failed, would hardly have spelled disaster. — R.B.