To lead effectively through strategic uncertainty, harness the power of experimentation, writes Sharique Hasan

In today’s uncertain environment, strategy built on forecasts and historical patterns often loses reliability. Economic instability, geopolitical shifts, and accelerating technological transformation cloud the road ahead. The fog of the future obscures what once felt predictable.

Yet strategic planning often begins with the assumption that foresight can help us bring uncertainty under control. This may be true in stable environments, but today’s world is shaped by forms of change that do not follow predictable paths. Technological shifts, institutional disruptions and evolving norms frequently move faster than the frameworks used to anticipate them. This often means that standard methods for generating strategic clarity tend to lose their effectiveness, just when clarity is most important.

Today, organizations face a different task. Rather than attempting to forecast the future, they may need to develop strategies for working with ongoing uncertainty. One approach involves shifting the organization’s emphasis toward learning rather than planning. Formal systems built around effective learning can help organizations respond more intentionally to conditions they cannot predict. Organizations can observe changing patterns, generate outcomes and make adjustments without requiring a rigid view of the path ahead.

Within this type of framework, formal experimentation has a critical role to play. It creates a system for navigating ambiguity through structured inquiry and execution: allowing organizations to test their assumptions, gather evidence and refine their actions as their circumstances evolve. The experimental organization harnesses the power of experimentation to learn faster, adapt earlier and respond more strategically to a fast-changing world.

Causal inference in organizations

Learning from experience requires more than collecting data and measuring outcomes. It depends fundamentally on understanding what factors caused those outcomes. Causal inference helps distinguish between patterns shaped by underlying mechanisms and those that reflect noise. 

For organizations operating in complex environments, this distinction matters. Without it, decision-makers may act on patterns that appear meaningful but lack any empirical basis. Choices built on spurious patterns often prove unstable, difficult to interpret and costly. However, in the absence of causal evidence, strategy tends to follow narrative and intuition. These may align with prior beliefs and mental models, or offer tidy explanations, but they are often shaped more by inertia than by evidence. When organizations act on anecdotes or beliefs, they can do so with confidence and still fail. These missteps often reduce the chances of learning over time.

In contrast, experiments offer a more disciplined approach. An experiment begins with a hypothesis, imposes contrast between conditions, and relies on measurement to assess outcomes. Each aspect of the experimental process serves a purpose: to quantify effects, support inference, and enable the revision of beliefs based on new evidence. What results from rigorous experimentation is a clearer account of why some actions work and others do not – and what direction is the most promising. 

That’s why experimentation has become routine in many organizations. Many digital platforms have institutionalized large-scale testing to inform design and operations, while in policy and medicine, randomized trials shape decisions that impact the health and wellbeing of millions. These examples suggest that systematic causal inference is possible when it is embedded in practice within organizations.

From running an experiment to experimentation

Individual experiments can yield useful insights. However, their value increases when experimentation becomes part of the organization’s regular practice. One-off tests may clarify short-term questions, but they do little to drive institutional learning. In contrast, a systemic emphasis on experimentation integrates the testing of new ideas into ongoing decision-making. This approach treats experimentation as a core and strategic capability – one that supports continuous trials on all decisions, large and small. Over time, it fosters learning processes that are distributed, continuous and adaptive.

In experimental organizations, failure is neither avoided nor exceptional. The
purpose is to expand the organization’s ability to learn across variation. Value does not come from isolated wins, but in the patterns revealed across multiple trials – potentially thousands of them. As results accumulate, the organization builds a broader understanding of which conditions shape outcomes, and which assumptions need refinement.

Experimentation contributes to two aims: confirming prior expectations and uncovering new opportunities. Through structured testing, existing beliefs can be examined and refined. At the same time, unexpected outcomes reveal overlooked variables or open new directions. These uses of experimentation – validation and discovery – work in parallel. When sustained over time, they increase the range of what the organization can learn and incorporate into its models of action.

The impact of experimentation deepens when findings, both successes and failures, are documented. Records of past tests – what was tried, what happened, and in what setting – create an archive of practical knowledge. This provides a resource for recognizing patterns and generating new questions, and means future decisions can be based on more than immediate judgment.

As experimentation becomes more integrated into an organization’s strategy, the quality of decisions begins to shift. Judgments reflect tested knowledge. Actions align more closely with actual business conditions. Strategic learning becomes less reactive and more deliberate.

Experimentation works

There is increasingly strong evidence that structured experimentation is linked to performance gains. A recent study I conducted with Rembrand Koning of Harvard University Business School and Ronnie Chatterji at Duke’s Fuqua School of Business (2022) studied the performance trajectories of over 35,000 early-stage ventures. We found that firms using A/B testing grow faster and introduce new products quicker, leading to more rapid scaling (and faster failures, too).

Our findings suggest that experimentation functions as more than a technical tool. It enables a flexible approach to decision-making in settings defined by uncertainty and change. By generating and testing alternatives systematically, organizations can improve both learning speed and decision quality, even when information is incomplete.

Booking.com exemplifies the approach. It runs thousands of experiments at any given time, evaluating every idea, including those from executives, on the basis of measured results. The accumulated insights inform interface design refinements, pricing strategy, and user flow improvements. Yelp has also adopted structured testing, achieving improvements in various parts of its business. Microsoft conducts over 100,000 experiments per year, using a centralized platform to support decentralized innovation across products.

Part of experimentation’s value lies in how it provides a feedback loop. Tests produce early data that inform next steps, reducing exposure and allowing for directional changes before larger commitments are made. This removes the need for precise prediction and replaces it with structured iteration. Over time, it enables organizations to take bigger bets, because risks are mitigated through earlier experiments that give decision-makers a sense of what to expect.

The four pillars of experimental organizations

How can leaders embed experimentation in their organizations? There are four pillars to consider. 

Technology and technique Sustained experimentation requires infrastructure that supports both rigor and scale. Centralized platforms can standardize test design, execution and analysis, reducing redundancy and supporting a shared procedural language. The validity of experimental results depends on data that is reliable, relevant and structured in ways that support causal reasoning. Without this foundation, testing becomes either unreliable or inefficient.

Human capital Experimental capacity depends on having the right people. As testing becomes a core input to decision making, it calls for roles that integrate technical competence with conceptual understanding. Some employees generate ideas grounded in user experience or operational constraints. Others design the formal tests and interpret outcomes. These contributions often overlap, but the essential task is to support the
full range of skills that an experimental organization requires.

Organizational design Effective organizations enable experimentation across teams. Those closer to customers often hold the contextual knowledge needed to generate and refine hypotheses. Decentralization helps surface these insights. At the same time, shared standards and common evaluation criteria allow results to be compared and decisions to be grounded in fairness and consistency. The organization functions both as a collection of local inquiries and as a network for shared learning.

Leadership Leadership shapes the context in which experimentation operates. When leaders ask for evidence and support tests that fail, they signal that experimentation is part of serious decision making. This support becomes crucial when results challenge existing assumptions. If findings are ignored, the information they provide disappears. Leadership also ensures that testing remains focused on priorities that matter. When experiments shape real choices, they become part of how strategy is made.

Strategic discipline 

Experimentation delivers the most value when aligned with the strategic aims of the company. Local tests can generate useful findings, but their broader relevance depends on connection to meaningful strategic goals. A more deliberate strategy brings experimentation into contact with areas that affect future direction – testing new markets, discovering unfamiliar user needs, or trailing different business models. When tests are selected to explore and learn in these spaces, they help the organization extend its understanding of what is possible.

Strategic use of experimentation also requires clarity about purpose. Some tests confirm prior expectations; others uncover new possibilities. Experiments lose value when designed to reinforce existing assumptions. Well-constructed tests can surface outcomes that challenge current views, redirect attention or reveal overlooked variables. Surprise can be a powerful driver of insight.

For learning to take hold, new knowledge must be shared across the organization, moving beyond the team that produced it. Reviews, documentation, and shared reflection allow patterns to emerge across domains. In some cases, results from one setting may inform another. Equally important is the structure of incentives. When rewards emphasize outcomes alone, risk aversion may take hold.

As experimentation becomes more central to the organization, it begins to shape how strategy is formed. Instead of relying on static forecasts, decisions evolve from evidence gathered across a portfolio of bets. Resources shift based on observed effects, not hierarchical power. Over time, this process supports earlier adjustments and sharper alignment with context. Small tests generate insights that accumulate, producing a deeper and more flexible understanding of how the environment works. The result is not just better choices, but a strategic advantage grounded in the capacity to learn.

The power of experimentation 

In uncertain environments, strategy depends on methods that support learning. Experimentation provides a structured way to generate feedback, adjust direction and stay responsive. Well-designed tests in organizations where leaders are committed to continuous experimentation support strategic coherence through changing conditions. 

The advantage of experimentation lies in consistent learning. It is the capacity to refine action based on evidence gathered over time: small insights accumulate into more effective decisions. 

What sets successful organizations apart is not what they know at the start of the journey – but how they learn as conditions evolve.