How Forgetting Leads to Innovation

Those who most readily capture imaginations are entrepreneurs-innovators with a vision who create something from nothing. Large corporations, by contrast, are easily cast as the villains. They thwart aspiring entrepre­neurs in every possible way-by blocking distribution channels, by filing lawsuits, or even by giv­ing away something of similar value for free.

What prevents corporations from winning with offense? After all, chief executives of large organizations have an enormous advantage in the innovation game. Their organizations contain mammoth resources such as established brands, technical expertise and strong cus­tomer relationships, just to name a few. These are advan­tages independent entrepreneurs can only dream of. Yet large organizations often struggle to innovate.

Since 2000, we’ve studied best practices for manag­ing high-growth-potential new businesses within estab­lished organizations. We compiled detailed multi-year case companies, including Corning, The New York Times and Hasbro. In every case, leaders were well aware of the opportunity for the new company (“NewCo”) to leverage valuable resources from the core business (“CoreCo”).

Indeed, NewCo must be able to borrow, but borrow­ing alone is not sufficient. Equally important is a com­peting need for NewCo to forget. The new business unit must forget because NewCo is fundamentally different from CoreCo, either because it has a different customer base, a different value proposition, a different value chain, a grounding in a different set of core competen­cies, or all of the above. It also carries much greater lev­els of uncertainty.

These differences in business definition are easy to understand. Every executive we interviewed could ex­plain in great detail how NewCo’s business model was different from CoreCo’s. Nonetheless, to NewCo’s detri­ment, CoreCo’s behaviors often persisted. There were powerful forces at work.

Consider the case of Corning Microarray Technologies (CMT), a biotechnology venture launched by Corning in 1998. At the time, biotech­nology researchers had an aggressive agenda to conduct millions of exper­iments on DNA-a brute force approach to unlocking the secrets of the genetic code. Automation was attractive. If machines could multiply the rate of experimentation and com­puters could sift through the results, the time to new revolutions in med­ical therapies could be reduced.

Sales of laboratory equipment were booming. The piece of experimental apparatus that presented a business opportunity to Corning was the DNA microarray-nothing more than a rec­tangular glass slide, a few inches long, with hundreds of microscopic samples of DNA adhered to it. With micro-arrays plus several other pieces of equipment, researchers could con­duct hundreds of experiments at once. But there was no microarray on the market that fully pleased researchers, and many chose to “print” their own microarrays-a great waste of time.

The microarray opportunity attract­ed Corning because it had world-class expertise in the manufacture of spe­cialty glass. There was a tremendous opportunity to borrow. But what did Corning need to forget?

Reinforcing the Past

Corning’s business model was consistent across almost all of its markets. Corning sold components to industrial manufactur­ers and almost always emphasized high quality and high reliability above other product attributes. The compa­ny held strong intellectual property rights on its products and therefore faced limited competition.

The microarray business, however, was different. CMT needed to sell to senior laboratory administrators, an unfamiliar customer. It needed to place greatest emphasis on cost and convenience, not the highest possible quality. It operated in an unfamiliar, emerging scientific field in which patent protection was highly unlike­ly. It required expertise in the advanced life sciences as much as expertise in glass manufacturing. And, it was much less predictable than the other busi­nesses in Corning’s portfolio.

Despite these well-understood dif­ferences, Corning chose to follow the organizational model that had proven effective in its other lines of business. In part, it did so out of convenience- organizational change creates head­aches. In part, it did so because managers did not see just how pow­erful organizational memory can be.

Mirroring the other business units, CMT included its own manufactur­ing, sales and marketing functions, and linked up with Corning’s central­ized R&D groups. CMT also adopted Corning’s detailed and disciplined five-stage innovation process to move from idea to market-along with mile­stones that were rooted in decades of experience.

There was one variation from Corn­ing norms. Corning departed from its practice of promoting primarily from within by filling a number of technical positions with outside experts in ad­vanced life sciences.

Although CMT achieved some ear­ly successes, it soon faced several unanticipated difficulties. For exam­ple, CMT reinforced its past by rely­ing too heavily on its existing experts in the physical sciences, not enough on its newly hired experts in the advanced life sciences. The latter disagreed, rightly in retrospect, with decisions to delay product launch­es to achieve quality standards that were higher than necessary, but they did not have much influence.

Also, CMT adopted Corning’s long-successful routines for rapid­ly improving new manufacturing processes. But these failed, con­founded by peculiarities of DNA fluid and inconsistencies among suppliers of DNA. One day the process would appear to be working fine, but the next day a mysterious new problem would arise.

In addition, following a deeply ingrained discipline of accountabil­ity, CMT adhered to a speculative plan for too long. After a few delays, CMT started missing deadlines.

CMT’s leaders felt enormous pressure, because falling short of plan was seen as failing. CMT’s loss­es put a drag on the profitability of the business unit to which it report­ed, further heightening tensions. Rather than reexamine fundamen­tal choices, CMT leaders viewed their struggles as minor setbacks, and they simply urged their team to work harder to get back to plan.

As CMT fell further behind plan, the cohesiveness of the leadership team faltered. Demonstrating a common pattern associated with for­getting struggles, CMT began re­solving conflicts by reverting to what had worked in the past. Eventually, CMT reached a crisis point, and the senior management team consid­ered major changes.

Everyone involved in CMT understood the differences between Corning’s existing businesses and the new business. But forgetting is not a matter of intellectual recogni­tion alone. Organizations have memories that are more powerful than individuals.

Overcoming Memory

CMT needed to find a way to overcome the following sources of organizational memory:

Collective instincts. High-growth­potential but highly uncertain ven­tures place managers under a great deal of stress, and present a far more ambiguous environment than most managers in established corpora­tions are accustomed to. Under con­ditions of stress and ambiguity, all people naturally gravitate toward the familiar-that is, rely on instinct- without even being aware of it. When an entire management team shares the same experiences and the same success stories, instincts are reinforced. CMT’s management team came of age within Corning.

Organizations have memories that are more powerful than individuals.

Reporting structures. NewCo often reports to functional or gener­al managers within CoreCo. An alternative, having NewCo and CoreCo appear as peers on the orga­nizational chart, seems awkward, since NewCo may be only one-hun­dredth the size of CoreCo. But hav­ing NewCo report within CoreCo is worse than awkward. Doing so vests power in a leader that is dedi­cated to excellence in the existing business. It creates one more rein forcer of organizational memory. CMT reported directly to an estab­lished business unit.

Relationships between managers. Relationships inside organizations develop in ways that are consistent with what makes the organization successful. Move the same man­agers to the new business, and the relationships persist. Established patterns of interaction continue. For example, once a balance of power and authority between two individ­uals is agreed upon, even implicit­ly, it is very hard to change it. Once a pattern for dividing and conquer­ing tasks is established, it is not easy to change it. Once people become accustomed to calling on specific individuals for technical advice, they will likely continue to draw on that network, even when the busi­ness challenge changes. Corning needed to shake up CMT to encourage managers to build new relationships consistent with its new business.

Trusted Measures of Performance. Once there is a history of evaluating business performance based on set standards, it is difficult to change it. Doing so seems unfair and inconsis­tent. It seems to be conferring spe­cial treatment to the new kid on the block, NewCo, and can be resented by the senior team. But emphasiz­ing CoreCo’s performance measures within NewCo serves to reinforce CoreCo’s business model, easily leading to misperceptions and mis­steps. CMT leaders hoped to achieve standards of quality that were proven to be achievable for other products, but not when work­ing with DNA.

Planning systems. Large, estab­lished companies deliver consistent earnings to investors when they are disciplined about holding managers accountable to plan. Planning process­es are designed to create and reinforce this discipline, highlighting outcomes against predictions with the expecta­tion that shortfalls are due to manage­ment underperformance. This mind-set is damaging for NewCo, however, because NewCo needs a learning agenda. Plans should identify the most critical unknowns, and the methods for resolving them. By contrast, once CMT made a commitment, it pro­ceeded as though its plan was achiev­able as written, and redoubled efforts to execute when reevaluating funda­mental assumptions was called for instead.

Norms about status. Most organiza­tions naturally confer status upon cer­tain areas of expertise. At a consumer products company, it may be market­ing experts. At a commodities trading company, it may be financial experts. Such implicit social hierarchies are not easily disrupted. Without careful orchestration, norms of organizational status persist. NewCo may end up dominated by people with expertise that is valuable for CoreCo, but less so for NewCo. CMT’s management team was full of experts in the materials sci­ences; the advanced life sciences were underrepresented.

Redesigning CMT

Corning’s struggle to forget was not unusual. We observed similar struggles at every com­pany we studied. Recognizing the dif­ficulties, senior management made numerous changes, starting by ap­pointing a new general manager. They also appointed two outside bio­technology experts to crucial leader­ship positions within CMT. In addition, they changed the reporting structure so that the head of CMT re­ported to the corporate level, separat­ed CMT’s research and development functions from Corning’s centralized R&D groups (and the established processes within them), and adopted a learning-oriented approach to plan­ning. CMT also co-located all of its staff to enable development of a more cohesive culture.

All of these changes took many months, but the impact was remark­able. Because Corning appointed a single leader responsible for the entire business and had him report directly to the senior staff, the pressures and norms of the traditional core business were avoided. Because CMT hired external experts in biotechnology to influential positions within CMT, there were consistent voices establish­ing reasonable standards for quality. Because CMT was no longer tightly integrated with centralized R&D groups, it was freed from a structured product development process and made progress by following a more iterative path. Because the manage­ment team treated projections appro­priately -as though they were guesses, not as though they were a non-nego­tiable basis for judging performance- the team regularly questioned the fundamental assumptions underlying its business and learned quickly.

The team made rapid progress. CMT launched its first microarray product in 2000, and customers rated it a “home run.”
(Adapted by permission from Ten Rules for Strategic Innovators: From Idea to Execution (Harvard Business School, 2005) by Vijay Govindarajan and Chris Trimble, from the Tuck School of Business, Dartmouth College.)

Erasing Corporate Memory

THE AUTHORS, based on their research, concluded that success in forgetting is more likely when com­panies make the following choices:

  • Hire Outsiders. Only outsiders placed in influ­ential positions effectively question orthodoxy every day.
  • Have NewCo report to an unconventionally high point. NewCo’s supervisors must be able to think long term. They must not be fully absorbed in CoreCo’s daily business imperatives.
  • Rebuild relationships. Working relationships between major functions such as marketing and manufacturing must be reconsidered and rebuilt from scratch.
  • Build a new dashboard. CoreCo metrics likely have little to do with NewCo’s performance.
  • Focus on learning. Projections for new busi­nesses are just guesses. Evaluate NewCo’s leader based on his/her quality of decision making and speed of learning.
  • Create a new culture. Examine deeply held beliefs about what makes CoreCo a success, and recognize that NewCo may need a culture grounded in entirely differ­ent success factors.

Common Mistakes When Setting Up a New Company

—— Design: Similar to “CoreCo,” and closely integrated.

—— Problem: “NewCo” can borrow, but cannot forget. NewCo inevitably tries to apply CoreCo’s success formula in an environment where it is unlikely to work.

—— Design: NewCo is distinct from CoreCo, and isolated from it.

—— Problem: NewCo can forget, but cannot borrow. It is unable to benefit from CoreCo’s valuable assets.

—— Design: Distinct from CoreCo, and linked.

—— Problem: Senior management team not engaged in ensuring link is healthy. Cooperation disintegrates. Effort to borrow fails.

—— Design: NewCo distinct and linked in one or two high-leverage areas. Senior executives engaged in keeping links healthy and productive.

—— Result: NewCo forgets and borrows.

Chief Executive magazine (published since 1977) is the definitive source that CEOs turn to for insight and ideas that help increase their effectiveness and grow their business. Chief Executive Group also produces e-newsletters and online content at and manages Chief Executive Network and other executive peer groups, as well as conferences and roundtables that enable top corporate officers to discuss key subjects and share their experiences within a community of peers. Chief Executive facilitates the annual “CEO of the Year,” a prestigious honor bestowed upon an outstanding corporate leader, nominated and selected by a group of peers, and is known throughout the U.S. and elsewhere for its annual ranking of Best & Worst States for Business. Visit for more information.