Run low-cost experiments on intersectional ideas

Test cross-domain ideas quickly and cheaply before investing in them fully.

Why it works

Intersectional ideas carry higher uncertainty than within-domain improvements — there is less established knowledge about whether they will work. This uncertainty makes large commitments risky. Low-cost experiments — prototypes, pilots, small bets — generate information about viability at a cost that the expected learning justifies, regardless of whether the idea succeeds. This converts high-variance intersectional bets into manageable portfolio exploration.

How to do it

1. For any intersectional idea, identify the smallest test that would confirm or disconfirm the core assumption.
2. Run that test before building anything further.
3. Set a clear decision criterion in advance: "If X happens, we continue; if Y, we stop."
4. Treat the result as information, not success or failure — update your approach based on what you learned.

Evidence

The lean startup methodology and design thinking literature both provide theoretical and practical support for low-cost experimentation as a strategy for navigating high-uncertainty innovation. This is organizational consensus with strong face validity. Lynn, Morone, and Paulson’s study of discontinuous innovations documents this as a "probe and learn" process, in which firms bring an early version to market as a series of low-cost experiments and let what they learn redirect the product. (mechanistic)

Minimum viable tests can be misleading if the test conditions don’t capture the relevant dynamics of the real application — small tests sometimes fail to surface the properties that matter at scale.

Sources

• Lynn, G. S., Morone, J. G., & Paulson, A. S. (1996). Marketing and Discontinuous Innovation: The Probe and Learn Process. California Management Review, 38(3), 8–37.

Common mistake

Designing a test that is so hedged or small that a positive result doesn’t actually reduce uncertainty about the real question — which consumes time and resources without generating usable information.