Human beings don’t like uncertainty. But under the right conditions, people can become more risk tolerant. Consider the famous experiments published by Amos Tversky and Daniel Kahneman in 1986; participants were asked to choose between:

A) A sure gain of $30
B) An 80% chance to win $45 and a 20% chance of winning nothing

An overwhelming majority, 78%, chose A, favoring certainty—even though B has a higher expected value ($36). But that changed when participants were presented with a different scenario:

A) A 25% chance to win $30
B) A 20% chance to win $45

In that case, 58% chose B. When both outcomes were uncertain, their appetite for risk increased.

The experiment demonstrated that certainty suppresses risk-taking, but uncertainty can encourage it. And, it turns out, this behavioral insight can also be applied to the way companies manage innovation.

Large incumbents tend to avoid risk

In large, established companies, innovation is often skewed toward the core. Many express a desire for a 70/20/10 innovation mix (70% sustaining, 20% breakout, 10% disruptive), but real-world pipelines often resemble 100/0/0—with virtually no projects far from the core.

Why? Because as project risk increases, most companies pull back. Promising ideas are often killed before they’re even fully explored because they are perceived to lack ROI or strategic fit within existing systems.

Using two innovation models can overcome risk aversion

But what if companies adopted two distinct innovation systems—one for the core and one for growth beyond it—each with its own structure, resources, and performance expectations? The two systems would look like the following:

A sustaining innovation system. Designed for incremental improvements close to the core business, this innovation system would be best for enhancing known products, channels, and business models, and it would have the following traits:

• Consumer focus: Addresses clearly understood, currently met, or under-met needs through refinement and optimization of existing offerings
• Risk profile: Low risk, measurable returns
• Process: Traditional stage-gate models
• Expertise: Deep domain experts
• Role of AI/technology: Used primarily for automation, predictive analytics, and process optimization—helping to streamline operations, reduce costs, and incrementally improve customer experiences
• Key Performance Indicators (KPIs): ROI-driven metrics

A breakout/disruptive innovation system. Tailored for exploration and transformative growth, it would be best suited to boldly pursuing new markets, technology, or consumer models.

• Consumer focus: Seeks to uncover and address unmet or unarticulated needs, often in white spaces or emerging behavioral shifts
• Risk profile: High risk, uncertain outcomes
• Process: Agile, iterative, fast-learning loops
• Expertise: Growth mindsets, non-traditional collaborators
• Role of AI/technology: Enables rapid prototyping, customer insight mining, market sensing, and adaptive experimentation, accelerating the discovery of new value propositions and business models
• KPIs: Learning velocity, long-term potential, strategic options

What the best innovators do differently

For some companies, this dual model approach is already in place. Our survey and interviews with 20 companies in Fast Company’s list of 50 Most Innovative Companies found that 79% already employ separate operating models—organizations' structure, metrics, KPIs, and ways of working—for disruptive innovation and sustaining efforts. Another 69% said that innovation is extremely important to the company's culture and daily operations.

When we asked leaders in these companies about the future structure of innovation, approximately twice as many predicted that innovation would become more centralized (56% of the total) as opposed to decentralized (25%), possibly reflecting a need for better governance, accountability, and strategic alignment amid complexity.

The tilt toward centralization—despite more democratized innovation tools—may reflect a need to better prioritize, focus, and fund innovation at scale, especially when stakes are high.

A more realistic path to innovation

Winning at innovation is not about making wild bets—it’s about designing the conditions that make bold ideas possible. Distinct operating systems with appropriate processes, governance, and KPIs allow companies to pursue efficiency and exploration simultaneously.

Rather than treating innovation as a one-size-fits-all process, leading organizations recognize that different kinds of innovation need different systems—and they structure accordingly.