From Skunkworks to Production, and Back
Inside the organizational tension between the established and the future
Any company that has survived for a decent number of years with a reasonable product-market fit eventually finds itself getting really good at squeezing what already sells while nervously thinking of what comes next. For survival, you have to keep one eye focused on the engine room of reliable products and another on the fuzzy cloud of emerging technologies, consumer behavior shifts, and entirely new markets.
This isn’t just a strategic challenge; it’s a structural and even emotional one. Because the mindset, pace, and risks involved in running a stable product line are completely different from those required to explore the unknown. Yet both have to coexist. That friction—between exploiting existing strengths and exploring radical new opportunities—is not a flaw in the system. It is the system. And managing it well is the difference between a company that endures and one that slowly fades.
This is where skunkworks makes a glorious entrance: small, semi-autonomous teams set up to chase ideas that the rest of the organization isn’t yet ready for. Skunkworks are designed to be fast, flexible, and a little weird. They're supposed to move beyond the gravitational pull of the core business, because if they don't, they'll get stuck optimizing spreadsheets instead of creating something new1.
But skunkworks can be dangerous too; first, because it surely needs money to keep going (and results are not immediate, with failure being common). And also because organizations often don’t think hard enough about what happens if the wild idea actually works. What if the prototype is so compelling that it demands a roadmap? This is where the tension shows up.
As I wrote before, a prototype is not a product. A prototype is an argument. A concept. A "what if" frozen in breadboards, code, and cardboard. It’s easy to mistake a working demo for a mature solution, but making something work once in a lab is a far cry from making it work reliably for customers across regions, use cases, and edge conditions, including thorough documentation, training, and post-sales support.
This gap between demo and production is where most innovations go to die. It’s also where the resentment often builds: the product team sees the skunkworks folks as disheveled dreamers; the skunkworks team sees the product org as risk-averse bureaucrats. No one wins.
Researchers Charles O’Reilly and Michael Tushman studied this tension in detail and came to a practical conclusion: the companies that managed it best were what they called ambidextrous organizations2. These companies kept their innovation units structurally separate from the core business—different pace, different KPIs, even somewhat different culture—but they tightly integrated them at the senior leadership level.
There’s a proper timing for it, and many get it wrong. Some may start worrying about the future too early (first get a product in the market that sells, then work on the next gen). Some may wait too long to innovate while incrementally polishing forever the cash cow. Some might push innovation too far, plainly ignore it, or merge it too early. Either way, the prototype can’t cross the gap. But when leaders act as a connective tissue—allocating capital, resources, and managing incentives—then innovation doesn’t just happen in a vacuum. It moves.
If you're trying to balance mature products and future bets, here's what the best organizations do:
They build bridges early. Product-minded thinkers are embedded in innovation teams from day one. They’re not there to kill ideas, but to help make them real.
They stagger integration. Don’t throw a prototype over the wall. Start the transition while the idea is still evolving. Let product, legal, ops, and support test assumptions early.
They change incentives. If innovation teams are rewarded for velocity and core teams are rewarded for stability, there will always be conflict. Align leadership incentives around shared outcomes, not isolated KPIs.
They embrace structural duality. Separation is healthy. You want skunkworks to be weird. You want product teams to be rigorous. What you don’t want is to pretend to be both.
They reframe the mission. You’re not building an eternal prototype. You’re building the future of the company. And that means acting like the future needs to scale.
Every company that survives long enough eventually walks this tightrope. And the trick isn’t to find the perfect balance and lock it in. The trick is to keep adjusting. To live in that tension. To know that innovation is useless unless it ships, and shipping is meaningless unless it evolves.
The real trick isn’t innovating. It’s making sure your innovation survives contact with the real world. That it gets across the gap, becomes a product, and earns its place on your website.
And everybody will quit because those cut for working in skunkworks are not made for maintaining spreadsheets.
Article is paywalled, but can be found in all the known archiving websites out there