How the World’s Top Companies Use Experimentation to Outlearn Uncertainty

Leaders at Airbnb wondered whether listings with professional photographs might perform better than those using user-uploaded images. Rather than relying on instinct or anecdote, they ran a controlled experiment: some listings were assigned professional photography, while others retained user-generated photos. The results were striking. Listings with professional photos received more than twice as many bookings and earned hosts over $1,000 more per month. What began as a simple test ultimately led Airbnb to launch a full-scale photography program, transforming how hosts presented their properties and how customers experienced the platform.

This is experimentation in action: a disciplined approach to uncertainty that allows organizations to uncover insights they might never reach through planning alone. 

Booking.com reportedly runs over 25,000 experiments each year, a practice that has helped transform it from a small startup into a global travel powerhouse. According to Lukas Vermeer, its director of experimentation, Booking.com runs more than 1,000 experiments simultaneously, often tailoring tests to individual website visitors. These are primarily A/B tests, in which two alternatives are assessed side by side to determine which performs better. Over time, this approach allows the company to optimize entire customer journeys, refining everything from search results to booking flows based on real-world behavior rather than assumptions. 

What these companies demonstrate is that sustained experimentation fundamentally changes how organizations learn. 

Why experimentation matters more than ever

Building a culture of experimentation creates the conditions for unexpected opportunities to surface and be exploited. It encourages organizations to move beyond incremental improvement toward breakthrough innovation, while also improving internal processes and engagement. Employees in experimental cultures tend to be more curious, more resilient and more willing to challenge the status quo. 

Creating this culture starts with the leaders. For experimentation to take root, leaders must be willing to redefine what success and failure mean. Instead of treating failure as something to be avoided or punished, leaders need to frame it as an essential part of learning. This shift enables a growth mindset in which teams are encouraged to generate ideas, test them quickly and scale what works. Crucially, leadership teams must model this behavior themselves. When leaders visibly test, learn and adapt, experimentation becomes embedded in the organization’s DNA rather than confined to innovation labs or product teams. 

Empowering employees to test and learn

A true culture of experimentation empowers employees at every level to test hypotheses and iterate continuously. That requires time, tools and psychological safety. Providing dedicated time for experimentation sends a powerful signal. 3M famously allowed its researchers to spend 15 percent of their time exploring scientific topics or personal interests, regardless of immediate commercial relevance. The policy led to numerous innovations, including the invention of Post-It Notes. 

Google adopted a similar philosophy, allowing employees to spend 20 percent of their time on side projects. While not every experiment succeeded, the approach produced significant breakthroughs like Gmail and AdSense. By making experimentation an expected part of the job, companies like Google and 3M normalized creative exploration and reduced the fear associated with trying something new. 

Amazon has taken a related but distinct approach, fostering a culture of “many small bets.” Rather than seeking uncertainty upfront, Amazon continually tests new products, processes and business models, accepting that most experiments will fail, but that a few will deliver outsized returns. 

Leaders don’t need to replicate these models exactly. Even modest steps, such as allocating one day per month for experimentation, offering workshops or providing small seed budgets, can be enough to spark momentum. 

Making data the backbone of learning

Experiment without measurement is just trial and error. Effective experimentation depends on data. Leaders should encourage teams to document their experiments clearly: what hypothesis was tested, what data was collected and what was learned. Results, positive or negative, should be shared openly to maximize organizational learning. Over time, this creates a shared language or evidence and reduces reliance on opinion-driven decision-making. 

As Adam Savage, the special effects designer and co-host of Mythbusters, has said: “In the spirit of science, there really is no such this as a ‘failed experiment.’ Any test that yields valid data is a valid test.” the essence of this approach is learning: rapid experimentation is vital for outpacing competitors, far more so than simply being right. 

Reducing fear through structure and play

Many organizations struggle with experimentation due to fear—specifically, fear of failure. Psychologists describe loss aversion as our tendency to fear losses more than we value gains. In business, this often shows up as risk avoidance, perfectionism and decision paralysis. Leaders must actively normalize failure as a learning mechanism and a key part of progress. Amazon founder Jeff Bezos captured this succinctly when he said, “If you know it’s going to work, it’s not an experiment.” Booking.com’s Lukas Vermeer echoes this philosophy, emphasizing that experiments exist to discover what works, not to prove someone right. 

Some organizations have gone further by gamifying experimentation. Platforms such as LabQuest have integrated points, badges and leaderboards into testing and user research, turning participation into a game. This approach has reportedly increased engagement and improved data quality, with significantly higher participation rates and more actionable insights compared to traditional methods. Gamification reduces the emotional stakes of failure and reframes experimentation as something engaging rather than intimidating. 

A simple framework leaders can use

One practical framework for experimentation is the Build-Measure-Learn-Loop, popularized by Eric Ries in The Lean Startup. It begins with a clear hypothesis: We believe that changing X will improve Y. Teams then run a small, fast, low-cost test, measure the results using relevant metrics and decide whether to scale, refine or abandon the idea. 

This loop isn’t limited to product development. HR teams can experiment with new onboarding processes. Marketing teams can trial messaging variations. Even finance teams can explore alternative budgeting allocation models. When every initiative is treated as a learning opportunity rather than a final verdict, organizations become more adaptive and resilient. 

Steven Bartlett, founder of Social Chain and host of The Diary of a CEO podcast, underscores the role leadership plays in this process. “Get your team to conduct fast, fearless experiments—more often,” he advises. Bartlett has described how his social team reports weekly on the tests they’ve run, reinforcing that experimentation is a core expectation. As he puts it, whether teams behave this way ultimately comes “down to the leadership.” 

Thriving through uncertainty

In a world changing at unprecedented speed, relying solely on past data and established models is increasingly risky. Markets shift, customer expectations evolve and competitive advantages erode quickly. Experimentation offers a way forward, not by eliminating uncertainty but by learning within it. 

High-performing companies test, learn and adapt in real time. For leaders, the lesson is clear: the ability to foster experimentation is no longer optional. It is a core capability for navigating unpredictability and uncovering unexpected solutions. 

The Art of Unexpected Solutions: Using Lateral Thinking to Find Breakthroughs by Paul Sloane was published on the January 3, 2026, by Kogan Page, priced £14.99.