Your Innovation Lab Is Theatre. Here's What Actually Works.

I've toured a lot of innovation labs.

The good ones have exposed brick and industrial lighting. There are beanbags, or maybe those standing desks that nobody actually uses. Whiteboards covered in sticky notes. A 3D printer in the corner, largely decorative. Inspirational quotes about failure on the walls.

The people inside are smart, enthusiastic, and genuinely trying to do good work. They run workshops. They generate ideas. They build prototypes. They create pitch decks.

And almost nothing they do ever reaches a customer.

Innovation labs are theatre. They exist to signal that an organisation takes innovation seriously. They provide cover for leadership to say "we're investing in the future." They absorb the restless energy of employees who want to build new things.

What they rarely do is actually innovate.

The Lab Failure Pattern

I've watched the same pattern play out across dozens of organisations. Universities. Corporates. Government agencies. The details vary. The trajectory doesn't.

Year One: Excitement.

The lab launches with fanfare. Leadership announces a commitment to innovation. A cool space is built out. Smart people are hired or seconded. There's energy, optimism, permission to think differently.

The team generates ideas. Lots of ideas. They run design sprints. They interview customers. They build prototypes. They present to executives. Everyone agrees the work is exciting and important.

Year Two: Friction.

The lab tries to move ideas into the core business. This is where it falls apart.

The core business has priorities. Budgets. Roadmaps. Governance processes. The lab's ideas don't fit neatly into any of these. They require resources the business hasn't allocated. They need technical capabilities that don't exist. They threaten existing revenue streams or power structures.

The lab's sponsors are supportive in principle but can't force adoption. The lab's critics (and there are always critics) point to the lack of shipped products. The team spends more time navigating politics than building.

Year Three: Decline.

The founding team gets frustrated and leaves. Budget gets questioned. The lab pivots to "supporting innovation across the organisation" which means running workshops instead of building products. The space becomes a venue for corporate events.

By year four, the lab is either shut down quietly or persists as a zombie, technically existing but producing nothing of value.

Year Five: Repeat.

New leadership arrives. They notice the organisation isn't innovative enough. They announce a new innovation initiative. The cycle begins again.

Why Labs Fail

The failure isn't about the people. Most innovation lab teams are talented and motivated. The failure is structural.

Labs are separated from the business.

The whole point of a lab is to create space away from day-to-day operations. Freedom to think differently. Protection from business-as-usual pressures.

But separation creates its own problem: no path to implementation. The lab can invent, but it can't ship. It can prototype, but it can't scale. It can generate ideas, but it can't make them real.

Xerox PARC invented the graphical user interface, the mouse, Ethernet, and the laser printer. Apple and Microsoft commercialised them. The lab was brilliant at invention, disconnected from execution. Xerox could have owned the personal computer industry. Instead, they watched others profit from their innovations.

Labs lack commercial accountability.

Most labs are measured on activity, not outcomes. Workshops held. Prototypes built. Ideas generated. Executives engaged.

None of these metrics require anything to actually work. None of them require customers to pay money. None of them require the business to change.

Without commercial accountability, labs optimise for what they can control: looking innovative. The theatre becomes the product.

Labs can't compel adoption.

Even when a lab builds something valuable, they can't force the core business to adopt it. They can present. They can pitch. They can demonstrate. They can't make anyone change.

The core business has every incentive to resist. Lab innovations threaten existing products, existing teams, existing budgets. The people who would need to implement lab innovations weren't involved in creating them and don't feel ownership.

The lab becomes a suggestion box. The suggestions pile up. Nothing changes.

Labs attract the wrong problems.

Labs tend to work on what's interesting rather than what's important. Blue-sky thinking. Emerging technology. Future scenarios.

Meanwhile, the business has actual problems. Customers churning. Costs rising. Competitors gaining. The lab isn't working on these problems because they're not exciting enough, or because they're too politically sensitive, or because they require deep engagement with the messy reality of current operations.

The lab innovates in a vacuum. The business struggles with real challenges. They never connect.

What Actually Works

The organisations that successfully innovate don't do it through labs. They do it through structures that connect invention to implementation.

What works: Embedded experimentation.

Steven Bartlett doesn't have an innovation lab at Flight Story. He has a Head of Failure who works across every team. Experimentation isn't separated from the business. It's embedded in the business.

Every team runs experiments. Every team reports on what they tested and learned. The experimentation function provides method and rigour. The teams provide domain knowledge and implementation capability.

This eliminates the handoff problem. The people who discover what works are the same people who implement it. There's no gap between lab and business because there's no lab.

What works: Small autonomous teams.

Amazon doesn't have innovation labs. They have "two-pizza teams." Small groups with end-to-end ownership of a product or capability. They can build, ship, and iterate without waiting for other teams.

These teams have commercial accountability. They own metrics. They're measured on customer outcomes, not activity. If something doesn't work, they feel it directly.

The structure enables speed. A two-pizza team can decide and act in days. A lab trying to influence a large organisation takes months just to get a meeting.

What works: Executive-sponsored ventures.

Google X (now just X) works because it has extraordinary executive sponsorship. Astro Teller reports directly to Alphabet leadership. Projects have real resources. Successful ventures spin out as independent companies.

Critically, X has kill authority. They celebrate shutting down projects that aren't working. "Kill early, kill often" is an explicit value. Only about 2% of ideas survive.

This isn't a lab in the traditional sense. It's a venture studio with commercial expectations and the resources to meet them.

What works: Acquisition and integration.

Some organisations innovate by buying innovation. They acquire startups that have already proven a model, then integrate them into the core business.

This outsources the hard work of zero-to-one innovation to startups who are better suited for it. The organisation's job becomes integration, not invention.

This works when the organisation has genuine integration capability. It fails when acquisitions are left to wither as separate entities, never connecting to the core business. (Sound familiar? It's the lab problem in different clothes.)

The Higher Education Version

Universities have their own version of innovation theatre. Sometimes it's an explicit innovation lab. More often it's distributed across:

The digital transformation office that produces strategy documents but doesn't ship products.

The learning innovation unit that runs pilots that never scale.

The IT skunkworks that builds prototypes the business never adopts.

The research commercialisation arm that files patents but rarely creates companies.

The industry partnerships team that signs MOUs but doesn't deliver joint outcomes.

Each of these exists to signal seriousness about innovation. Each struggles with the same structural problems: separation from core operations, lack of commercial accountability, inability to compel adoption.

The result is theatre. Activity without outcome. Motion without progress.

Building What Actually Works in Education

If you're in higher education or edtech and want to actually innovate (not just perform innovation), here's what the evidence suggests:

Embed experimentation in teaching teams.

Don't create an innovation lab to reimagine pedagogy. Give teaching teams the permission, resources, and support to run experiments in their actual courses. Let them test new approaches with real students. Measure what works. Scale what succeeds.

The Head of Failure model works here. Someone whose job is to increase experiment velocity across teaching teams. Not a separate lab, but an embedded function.

Create small teams with end-to-end ownership.

If you're building edtech products, organise around two-pizza teams that own entire capabilities. Let them build, ship, measure, and iterate without dependencies on other teams.

If you're a university, consider giving small teams ownership of entire student segments or learning pathways. Let them innovate for their cohort without waiting for institution-wide approval.

Demand commercial accountability.

Innovation efforts should be measured on outcomes, not activity. Not "workshops held" but "student outcomes improved." Not "prototypes built" but "revenue generated" or "costs reduced."

This is uncomfortable in higher education, where commercial language feels foreign. But accountability doesn't require profit motive. It requires clarity about what success looks like and honest measurement of whether you're achieving it.

Kill what isn't working.

Most education innovations persist as zombies. Pilots that never scaled but were never killed. Initiatives that lost momentum but still consume resources. Projects that everyone knows aren't working but no one has authority to end.

Build kill decisions into your process. If an experiment doesn't show promise within defined timeframes, end it. Celebrate the ending. Extract the learning. Move resources to something more promising.

Connect invention to implementation.

Whatever structure you create, ensure there's no gap between discovering what works and implementing it. The people who run experiments should be the people (or closely connected to the people) who can scale successes.

If you have a separate innovation function, staff it with people who have credibility and relationships in core operations. Give them authority to influence resource allocation. Make their success dependent on adoption, not just invention.

The Honest Question

Here's the question to ask about any innovation effort:

What has shipped in the last six months that customers are using and paying for (or, in education, that students are experiencing and benefiting from)?

If the answer involves prototypes, pilots, proofs-of-concept, or presentations, you have theatre.

If the answer involves real products, real courses, real services being used by real people, you might have innovation.

The theatre is comforting. It provides jobs. It signals seriousness. It absorbs energy that might otherwise cause trouble.

But it doesn't produce change. And in an era where AI is reshaping education monthly, theatre isn't enough.

The Tear-Down

If you currently have an innovation lab, you have three options:

Option 1: Embed it.

Dissolve the lab as a separate entity. Redistribute the people into operational teams. Transform their role from "inventors" to "experiment enablers." Measure them on the experiment velocity and outcomes of the teams they support.

Option 2: Commercialise it.

Give the lab real commercial accountability. They need to generate revenue, reduce costs, or demonstrably improve measurable outcomes. If they can't, they get cut. No more activity metrics. No more theatre.

Option 3: Spin it out.

If the lab has genuine venture potential, spin it out as a separate entity with its own funding and accountability. Let it succeed or fail on commercial terms. Stop pretending it's part of the core organisation.

What you shouldn't do is continue the theatre. Another year of workshops, prototypes, and presentations that never ship. Another year of talented people generating ideas that never reach customers. Another year of pretending that activity equals innovation.

The beanbags aren't the problem. The whiteboards aren't the problem. The separation from reality is the problem.

Close the lab. Build the capability. Ship the product.

That's what actually works.