Discovery-Driven Planning for Disruptive Markets
Plan to learn and iterate, not to execute a preconceived strategy
One of Christensen's most practically important insights is that markets for disruptive technologies are fundamentally unknowable in advance. The techniques that work brilliantly for researching sustaining innovations, including customer surveys, focus groups, market sizing, and financial modeling, fail catastrophically when applied to disruptive technologies. This is because customers themselves do not know how they will use a product that creates new capabilities, and traditional market research asks customers to evaluate products against needs they already understand.
Christensen coined the term 'agnostic marketing' to describe the appropriate stance: explicitly assume that no one, not the company, not the customers, not the experts, can know in advance whether, how, or in what quantities a disruptive product will be used. Business plans for disruptive innovations should therefore be plans for learning, not plans for execution. The initial strategy is almost certainly wrong, and the company must preserve enough resources and organizational credibility to pivot when a better direction emerges through real-world experimentation.
The data supporting this approach is compelling. The most respected industry forecasters could predict sustaining technology markets within 7-8 percent accuracy but missed disruptive technology markets by 35 to 550 percent. Honda's wildly successful entry into North American motorcycles was not the result of deliberate strategy but of accidental discovery when employees noticed demand for the small Supercub bikes they were riding for personal transportation.
- Markets that do not exist cannot be analyzed with conventional market research techniques
- The initial application for a disruptive technology is almost never the one that ultimately drives the largest market
- Experts in existing industries have poor track records predicting disruptive markets because they apply sustaining-technology mental models
- Companies must create market knowledge through expeditionary forays rather than through analysis of existing data
- Unanticipated successes are often more strategically important than unanticipated failures, yet most planning systems focus on closing the gap between plan and reality only when results fall short
- Assume your initial strategy is wrongBegin with the explicit assumption that your first idea about the market, the product, and the customer is probably incorrect. This is not pessimism but realism grounded in historical evidence. Budget resources and organizational credibility for at least two or three iterations.Pro tipHP's Kittyhawk managers reflected that their biggest mistake was acting as if their forecasts were right rather than as if they were wrongWarningDo not bet all resources on the first attempt. Apple bet heavily on Newton's PDA-specific features, leaving nothing to redirect when the actual market turned out to be different.
- Get into the market with real products quicklyLaunch a simple, inexpensive version of the product as rapidly as possible. Sell it to real people for real money. The only useful information about disruptive markets comes from actual market interaction, not from surveys, focus groups, or expert panels.Pro tipWatch how customers actually use the product rather than asking them how they would use it. Honda discovered the dirt bike market by observing neighborhood interest in their employees' personal Supercubs, not through market research.
- Pay attention to unanticipated successesConfigure your management systems to flag and escalate unexpected positive results, not just shortfalls against plan. The most important market signals for disruptive technologies often come from unanticipated applications that nobody planned for.Pro tipIntuit discovered that many Quicken personal finance users were actually using it for small business bookkeeping. This unanticipated success led to Quickbooks, which captured 70 percent of the small business accounting market within two years.
- Design the product for flexibility, not optimizationDo not over-engineer the product for a specific application that may not materialize. Keep it simple and adaptable. Invest in features that maintain flexibility rather than committing to expensive, application-specific capabilities.WarningHP designed the Kittyhawk with an expensive automotive-grade shock sensor specifically for the PDA market. When the actual demand came from video game makers wanting a $50 drive, HP could not redesign downward and had to abandon the product.
- Iterate rapidly based on market learningWhen the market reveals a better direction than your initial hypothesis, pivot quickly. Preserve enough resources and credibility to redirect. Fail early, fail cheaply, and learn fast.Pro tipFrame failures as tuition payments for market education. Each failed experiment eliminates uncertainty and narrows the search for the right product-market fit.
Honda entered the U.S. motorcycle market intending to sell large, powerful motorcycles to compete with Harley-Davidson. The venture was failing badly when employees noticed that people kept asking about the small Supercub bikes they were riding for personal transportation. After years of resistance, Honda finally began marketing Supercubs through sporting goods retailers rather than motorcycle dealers.
Intel developed the microprocessor as a sustaining technology for calculators, not computers. When a Japanese calculator company that had commissioned the chip went bankrupt, Intel was left with inventory it tried to sell to other calculator makers. The personal computer application was discovered through experimentation, not through market research or strategic planning.
This framework crystallized from multiple case studies in which elaborate planning failed for disruptive products while accidental discovery succeeded. HP's Kittyhawk drive was designed through exhaustive market research for the PDA market, which never materialized. Meanwhile, the actual early markets for tiny drives, including Japanese word processors and video game systems, were completely unforeseen. Honda planned to sell large motorcycles in the U.S. and failed; its blockbuster Supercub business emerged from employees riding their personal bikes around Los Angeles. Intel's most profitable product, the microprocessor, was initially designed as a sustaining technology for calculators; its use in personal computers was an unanticipated market discovered through experimentation.