Innovation Dies in Silos: Spreading Experimental Intelligence Beyond the First 16 Weeks of Experimenting
Once you've run your first experimental cycle, your first 16 weeks of genuine learning, the real question emerges: How do you spread this capability without killing it through bureaucracy? The answer lies in creating what we might call distributed experimental capacity that thinks for itself while serving the whole organization.
Here's how to build on momentum for spreading experimental intelligence beyond local settings:
Secure visible leadership commitment and curiosity, not just authority
Executives must model curiosity to experiment and celebrate learning from failure. When Microsoft CEO Satya Nadella took over, he famously shifted the culture from "know-it-alls" to "learn-it-alls." He started leadership meetings by sharing his own learning experiments and what had failed that week. Within three years, Microsoft's market capitalization more than doubled.
The provocative truth: When your CFO publicly shares what didn't work in her experiment, she gives every team permission to optimize for learning rather than looking successful. Without this, your experimental spaces will become theatrical performances of innovation rather than genuine discovery.
Provide the Right Toolkit
Equip teams with resources for testing, analysis, and feedback. Adobe's Kickbox program literally provides a physical red box containing a $1,000 prepaid credit card, instructions for running experiments, and tools for testing ideas. Since launching in 2014, over 1,000 Adobe employees have used Kickbox, resulting in products like Adobe Portfolio and Adobe Spark. The program demonstrates that investing in capabilities, not just ideas, drives innovation outcomes. This isn't about ideas. It's about capabilities.
The approach emphasizes continuous improvement through small experiments and rapid learning, leveraging internal capabilities like talent, relationships, and workflows. Each local experiment needs its own resources to sense and respond to its environment. Centralized control of all resources creates dependency, not autonomy.
The uncomfortable question: If you don't trust teams with $1,000 and three weeks, why do you trust them with customer relationships worth millions?
Balance Freedom with Guardrails
Here's where most organizations get it backwards. They either impose suffocating controls or declare "radical freedom" and create chaos. Both approaches kill experimentation.
Experimentation without structure can turn into chaos. Think of governance not as a constraint, but as the framework that makes bold exploration possible. Google's "70-20-10" rule allocated 70% of resources to core business, 20% to adjacent innovations, and 10% to radical experiments. This structure gave teams freedom within defined boundaries while ensuring the company maintained strategic focus. Clear boundaries paradoxically enable greater creativity.
Make Knowledge Sharing Reflexive, Not Optional
Innovation dies in silos. Organizations must break free from rigid structures that trap them by creating nervous systems for knowledge flow.
Pixar's daily "dailies" meetings bring together animators, directors, writers, and technicians to review work-in-progress footage every single day. This creates a culture where showing unfinished work isn't scary, it's expected. Ideas cross-pollinate constantly. Pixar's hefty revenues and numerous Academy Awards, in part, are a result of their making knowledge sharing reflexive, not optional.
Protect Against Short-Termism
This approach delivers meaningful results in months rather than years, but "meaningful" doesn't always mean "monetized." Sometimes the most valuable outcome is learning what won't work before you've bet the company on it.
The most valuable insights often emerge through multiple iterations and thoughtful reflection. Amazon Web Services started as an internal experiment to solve Amazon's own infrastructure challenges in 2002. It didn't launch publicly until 2006. It didn't become profitable until 2009. Today, it generates over $90 billion in annual revenue and accounts for the majority of Amazon’s operating profit. If Amazon had demanded immediate ROI, cloud computing might have been abandoned after year two.
Spreading Knowledge: From Local Experiments to Transformation
The paradox of successful experimenting is this: you can't become one through traditional transformation. You can only become one by distributing the capacity to sense, learn, and adapt throughout your organization, by creating local experiments that think for themselves while serving a shared purpose.
This means:
Starting small with genuine experiments, not massive programs
Measuring learning velocity, not just business outcomes
Celebrating spectacular failures that yield valuable insights
Giving teams real resources and real autonomy
Creating reflexive knowledge sharing, not episodic reporting
Protecting long-term learning from short-term pressure
The Uncomfortable Question You Need to Answer
If your organization's success depends on learning faster than the environment is changing, and it does, then the question isn't whether to build experimental capacity.
The question is: Can you tolerate the discomfort of local experiments that think for themselves?
Because here's what that looks like in practice:
Teams making decisions you wouldn't have made
Experiments failing in ways you didn't anticipate
Knowledge emerging from unexpected places
Authority becoming less about control and more about coordination
Success that can't be attributed to a single leader or initiative
This is what experimental organizations look like. They're messy. They're uncertain. They learn continuously.
And they survive.
