Your Innovation Lab Is Probably Theater. Here's What Actually Works.

Picture this: You're a retailer. Somewhere between mid-sized and massive. The tech is moving faster than you can keep up with. Amazon exists. Your stores are closing. Your margins are getting crushed.

So you build an innovation lab. You get some venture capitalists. You bring in a technology partner. You design a space that looks like the future—all glass walls and standing desks and people with MacBook Pros looking thoughtful.

And then nothing fucking happens.

The lab becomes a showcase for journalists. It becomes proof to shareholders that you're "taking innovation seriously." It becomes theater.

I recently talked to Carl Boutet, who's spent years studying innovation labs across retail and other industries. He's seen labs at Microsoft that actually work. He's seen labs at Intel that function. And he's seen dozens of labs that are pure innovation theater—expensive stages for a show nobody in the organization actually wants to watch.

Here's what Carl learned: Most retail labs fail for one reason. Not because the technology isn't good. Not because the team isn't talented. But because they're built on a fundamental misunderstanding of what a lab is supposed to be.

The 2008 Lesson Nobody Learned

Carl describes the moment retail realized it needed to innovate differently. It was during the financial crisis. E-commerce was growing. Supply chains were fragile. Physical stores were looking increasingly vulnerable.

And the first thing retail organizations did when money got tight?

They killed the innovation departments.

That's the pattern from 2008: Crisis hits, innovation gets cut first. Because innovation is seen as a luxury. A nice-to-have. Something that matters only if everything else is secure.

Fast forward to now, and Carl's noticing something different. The majority of retail organizations aren't cutting innovation. They're investing in it. They're recognizing that the crisis isn't going to pass by waiting for old models to work again. It's going to require fundamentally different approaches.

But here's the problem: They're still building labs the wrong way.

Innovation Theater

Carl is blunt about what he sees: Most innovation labs are not actually labs. They're stages for a performance.

A lab—a real lab—is a place where you test hypotheses. You learn. You fail fast. You iterate toward something that works. It's messy. It's unglamorous. Most of the time, it looks like people sitting around having boring meetings about what works and what doesn't.

What most retailers build is the opposite. They want spectacle. They want something Instagram-worthy. They want a space that makes them feel like they're cutting-edge, even if nobody's actually cutting anything.

And here's why: A real lab requires admitting you don't know something. A real lab requires the humility to say "We're going to try this and we might fail." Real labs make people uncomfortable because they expose what the organization doesn't understand.

Fake labs are comfortable. They're controlled. They let you appear innovative without the risk.

But—and this is critical—customers can tell the difference. Carl says "The best labs are the ones you don't know are labs." The technology is running in the background. You're having a good experience. You don't know your behavior is being tracked by RFID-powered predictive analytics. You just know your experience is smooth and personal.

That's not theater. That's actually working.

The Resilience Playbook

When Carl thinks about what's coming for retail, he uses a word that keeps coming back: Resilience.

How do you build an organization that can absorb shocks? That can adapt when a crisis hits?

His answer: You need to be strong in multiple modes simultaneously.

He calls it the 50-50 blur. Physical and digital are so intertwined that you can't tell which is which. You're not a 90% digital company with some physical stores as an afterthought. You're not a physical retailer with a half-hearted e-commerce store. You're something genuinely integrated.

This matters because when something breaks in one mode, you still function in another. During COVID, physical retail essentially shut down. The retailers who survived were the ones who had invested in digital. Not as an afterthought. As a core capability.

The playbook Carl advocates is: Build toward doing half your business in digital channels and half in physical. You might end up at 60-40. You might hit 65-35. But if you're aiming for 50-50, you're building real balance.

The innovation lab, then, isn't a separate thing. It's where you figure out how to make that integration work. How to make digital seamless with physical. How to let people move between modes without friction.

This is completely different from what most labs are trying to do, which is "innovate in isolation and then try to shove it into the existing business."

The Location Problem

Here's where Carl's expertise gets really specific. Most retailers building labs ask: "Where should we locate it?"

And the answer from consultants is usually: "Somewhere cool. Somewhere that signals innovation. Maybe near a tech hub or a university."

Carl's answer: "In an actual store."

Not in a special facility. Not in a lab building. In a real store. Where real customers come. Where real transactions happen. Where your people have to manage real constraints.

Why? Because if you innovate in a vacuum, the results won't transfer to reality. You get excited about something in your lab. You build it. It works perfectly. Then you try to deploy it in an actual store and it falls apart.

The store manager tells you: "Yeah, that worked in your lab. But in my reality, kids are spilling juice on the equipment. My Wi-Fi is spotty. I have limited staff. My equipment budget doesn't exist."

So the innovation dies. Not because it was bad. Because it was lab-born, not reality-born.

But there's another element: The person factor.

Carl talked to a friend who runs innovation for a major retailer. This company has several thousand stores. And they've chosen 20 of them for innovation testing. When Carl asked how they chose the 20, he expected a scientific answer. Demographics. Geographic representation. Proper sampling.

The answer was simpler: "We picked the stores where we have a manager who's open to this and excited about it."

That's the secret. Everything else is secondary. If the manager doesn't want the innovation, it won't work. If the manager is excited about it? It has a shot.

The Business Model Problem

Here's where most innovation labs stumble: They're money pits.

A real lab—a working lab—costs money. You're running experiments. You're testing things that don't work. You're burning resources on learning.

But you can't burn indefinitely. At some point, the organization asks: "When does this start generating revenue?"

And most labs don't have an answer.

Carl's shift in thinking is that the lab shouldn't be a pure sand box with no connection to reality. It shouldn't be a revenue generator either. But it should be at least cost-neutral. It should be running actual transactions. It should be learning from real economic activity.

This is a hard balance. Carl initially thought labs needed to be completely separate from reality, completely untethered from commercial pressure, to do real innovation. But he's evolved. He now thinks the best labs are the ones doing real work in real contexts while still maintaining the freedom to experiment.

It's not a pure black box. It's a controlled environment that still behaves like the real world.

Why This Matters

You're probably not a major retailer. But the principles Carl discovered apply anywhere you're trying to innovate:

First: Understand that innovation theater is not innovation. It feels good and looks good, but it doesn't change anything. If your lab isn't producing learnings that change your business, you're just running an expensive PR campaign.

Second: Build your innovation in contexts that resemble reality. Controlled experiments are fine for lab science. But for business innovation, you need real customers, real constraints, real conditions. Otherwise, what you learn won't transfer.

Third: Location matters, but not in the way you think. It's not about being cool or being near a tech hub. It's about having people who give a shit. A manager who's excited. A team that believes in what you're doing.

Fourth: Resilience comes from balance, not from betting everything on the future. Figure out how your organization functions in multiple modes and get good at all of them.

And finally: Money is a teacher. Real costs and real revenues teach you things that pure research never will. Don't make your lab entirely dependent on them. But don't isolate them entirely either.

The Uncomfortable Truth

Here's what Carl won't say directly but is true nonetheless: Most organizations aren't actually interested in innovation. They're interested in the appearance of innovation.

Real innovation requires changing how you operate. It requires admitting you don't know something. It requires accepting that some experiments will fail. It requires cultural changes that touch compensation, decision-making, who gets promoted.

That's uncomfortable. So most organizations choose theater instead. They build the lab, they get the press coverage, and then they leave everything else exactly as it was.

And then they're shocked when the lab produces nothing.

---

Want the Full Story?

Carl Boutet has studied innovation labs across industries and has now taken on the role of advisor for a retail innovation lab at McGill University in Canada. He's seen what works, what doesn't, and most importantly, why most of what's being called "innovation" in retail is actually something else entirely.

If you want the full conversation—including specific examples of labs that work, the economics of running a viable lab, and how to choose locations based on people rather than real estate—listen to Yaniv Corem's interview with Carl on The School of Innovation podcast.

Because here's the thing about building an innovation lab: Everyone can do it. Building one that actually produces change? That's the hard part. And it requires a hell of a lot more than glass walls and standing desks.