The Innovator’s Dilemma

A few enthusiastic users can help reveal an opportunity. They cannot, on their own, establish a viable market. Meanwhile, existing products can consume the time needed to find out.

Two difficulties—not one diagnosis

Christensen’s disruptive-innovation theory explains how established businesses can overlook low-end or new-market opportunities because existing customers and financial priorities favour current products. Not every unsuccessful launch is a case of disruption.

For any new offer, distinguish evidence of demand from capacity to deliver and learn. A protected team cannot make an unwanted product viable; a promising market cannot compensate for a product that never becomes usable.

Lead users: insight, then wider validation

In von Hippel’s research, lead users experience needs ahead of a wider market trend and expect substantial benefit from solving them. They are not simply demanding customers or early adopters. His method includes testing lead-user concepts with more typical users.

Use their insight to form a hypothesis, then investigate:

• Need: who else has this problem, how serious is it, and what do they do now?
• Buying and adoption: who uses, decides and pays? What price, switching costs, implementation and purchasing barriers matter?
• Viability: can we reach enough suitable customers and deliver value at a sustainable cost?

“That sounds useful” is weaker evidence than a credible commitment: for example, a paid pilot with agreed outcomes and a realistic route to purchase. No single test guarantees demand.

A worked root-cause hypothesis

Reported pattern: several new-product attempts did not become commercially successful; development relied on a small group of users, wider demand was not checked first, and existing products kept people busy.

Possible causal paths—not established findings:

• Selected users’ enthusiasm → assumed wider demand → substantial development before testing buyer commitment → an offer whose repeatable market remained uncertain.
• Existing-product commitments → fragmented time and delayed learning or delivery → slower progress on the new offer → continued dependence on existing products.

Possible underlying conditions: no explicit owner or decision criteria for market validation; no agreed limit on investment before evidence; and no feasible arrangement protecting learning time. Check these conditions rather than blaming the users or the team.

Review each attempt separately: what was known when investment was approved, who was consulted, what buyers actually committed to, why prospects declined, and how planned development time compared with time actually available. Distinguish no viable market from a market never properly tested, and either from delivery, pricing, timing or sales-channel problems.

Change the learning and allocation system

1. Own the investigation. Name someone responsible for testing need and buyer commitment—not merely collecting feature requests.
2. Protect a small learning budget. Agree which existing commitments stop, wait, transfer or receive cover. A separate team may help, but is not the only workable arrangement.
3. Test before scaling. Use interviews about actual behaviour, prototypes and bounded pilots with prospects beyond the original collaborators. For future markets, test the trend and timing as well as current demand.
4. Make staged decisions. Agree evidence, cost and time limits for continuing, changing direction or stopping. Include delivery and support economics, not just customer interest.

Use the Three Key Questions

• Right problem? Is demand missing, delivery constrained, or both? What evidence distinguishes the explanations?
• Right intervention? Combine a proportionate market test with realistic capacity and decision rights.
• Right measures? Track qualified prospects, credible purchase commitments, adoption, customer outcomes, delivery cost and learning time actually protected. Check effects on existing customers and cash.

What would count as progress?

Set context-specific review criteria before each test. Record reasons for rejection and compare results with the original assumptions. Early evidence may be qualitative or sparse; use time-series analysis only when the data and question suit it.

No chart pattern uniquely diagnoses the Innovator’s Dilemma. A detected change does not establish its cause, and no detected change does not prove that a product has no market.

Return to “Where might the difficulty lie?” →
Investigate causes with evidence →