How to Think About Experiments in Organizations

Often, we think of this process as a form of experimentation. But is the term “experiment” apt in the corporate setting? Traditionally, when using the scientific method, an experiment must be repeatable—that is, the act of doing one experiment doesn’t alter the context in which the second identical experiment is conducted. Moreover, many scientific experiments can be done by a single person—the iconic lab-coat-wearing scientist, activating chemical reagents and logging the results.

As Tuck professor Ron Adner makes clear in a new paper with co-author Daniel Levinthal, experiments in firms are a different sort of beast. Perhaps the most important distinction is that corporate experiments are usually strategic: they are designed to inform a future course of action in a dynamic context, not just arrive at an objective truth. This can make them non-repeatable because “experimenting with strategies often changes how the world—consumers, competitors, ecosystem partners—reacts to future actions,” Adner writes. Corporate strategy experiments are also, by their nature, more complex, since they frequently require multiple people to collaborate and buy into the project.

In “Strategy Experiments in Nonexperimental Settings: Challenges of Theory, Inference, and Persuasion in Business Strategy,” which was published recently in Strategy Science, Adner and Levinthal break down the corporate experiment into two main categories: Individual action versus joint action, and repeatable or nonrepeatable. Since each category has two elements, there are four total combinations. Learn more about each below. 

As Adner advises, “Not all experiments are created equal. Some deserve more support, some warrant taking more risk. Conducting an experiment can inadvertently change not just risk, but also potential return. Experiments are valuable because they can unlock action. But as a strategy professor, my bias is to be thoughtful in advance.”

Individual Undertaking of Repeatable Experiments

Essence: A single decision-maker explores ideas in a controlled environment where each experiment is self-contained and does not affect future opportunities. Because actions do not change the surrounding context, failed experiments carry little risk. Learning is structured and cumulative, relying on iterative trial-and-error with high clarity, as the underlying conditions remain stable.

Example: Thomas Edison’s methodical testing of lightbulb filaments—he ran thousands of experiments without worrying about broader consequences. Today, tech firms engage in similar low-risk exploration through A/B testing or usability trials, refining product features prior to market exposure. These activities often occur within R&D settings, where failures are expected and do not carry reputational costs.

Ecosystem Implications: Because these experiments occur internally and remain largely invisible to the outside world, there is minimal impact on customers, competitors, or partners. There is no need for external alignment or stakeholder persuasion. Firms retain full flexibility to revise or reverse course without consequence, making this category ideal for early-stage development and technical feasibility testing.

Individual Undertaking of Nonrepeatable Experiments

Essence: Here, an individual actor takes bold, high-stakes action that changes the strategic landscape. The environment is no longer stable—actions reshape what is possible in the future, for both the firm and others. Because the same experiment can’t be rerun under the original conditions, learning is less about statistical inference and more about theory-driven speculation.

Example: Apple’s launch of the iPhone not only changed its own trajectory but also shifted customer expectations and redefined the smartphone category. Similarly, a high-profile firm introducing a radically different pricing model or making a large, public-facing acquisition can't simply reverse course without cost. The strategic move is a point of no return.

Ecosystem Implications: Such actions have ripple effects across markets. They influence customer sentiment, invite competitive responses, and shape the expectations of ecosystem partners. If the experiment fails, it can tarnish credibility and limit future opportunities—not only for the firm, but for others pursuing similar paths. Conversely, a successful bold move redefines norms and sets new baselines for industry behavior.

Joint Actions and Repeatable Experiments

Essence: This involves experiments that are still technically repeatable, but which require collaboration among multiple actors. Even though the environment is stable, success hinges on persuading others—within or outside the organization—to support and participate in the initiative. As such, the challenge shifts from technical feasibility to coalition building and alignment of interests.

Example: When Edison launched the Pearl Street Station, the first commercial powerplant in the U.S., he had to secure approval from local government, coordinate with engineers, and gain financial backing—all without formal authority over these players. Modern examples include cross-functional strategic initiatives or early-stage ecosystem launches, where even a pilot effort depends on internal or partner buy-in, often before tangible results exist.

Ecosystem Implications: Persuasion becomes a critical enabler of progress. Internal misalignment can delay or derail otherwise sound initiatives, and external partners must be convinced not just of the technical plan but of its strategic value. While the environment allows for repeated experimentation, social and political dynamics make each test dependent on consensus, trust, and shared interpretation of results.

Joint Undertakings of Nonrepeatable Acts

Essence: This combines the challenges of joint action with the complexity of irreversible, high-impact moves. The business environment is dynamic and influenced by each strategic step, making learning ambiguous and planning difficult. Initiatives in this space require coalitions to take and sustain action even as conditions shift—and often, to pivot together when things don’t go as expected.

Example: In the COVID-19 vaccine rollout, governments, pharmaceutical companies, regulators, and logistics partners had to align quickly and act in concert, with each step reshaping the public health and regulatory environment. Similarly, the failure of Project Better Place impacted not only that venture but also diminished interest and investment in similar EV ecosystem models. In both cases, strategy couldn’t be separated from context- shaping outcomes.

Ecosystem Implications: Persuading others is not a one-time effort—it must be sustained throughout the lifecycle of the initiative. Failures don’t just delay progress; they can unravel coalitions and taint future efforts across the ecosystem. A successful pivot, in this setting, is not just a revised strategy—it reflects a recommitment of trust, energy, and conviction across a group of actors who must believe in a new shared vision, despite past setbacks and a now-changed context.