Permission to Pull the Plug: A 90-Day Framework for Immersive Experiments That Refuse to Die

Somewhere in the United States right now, a brand team is six months into a virtual showroom pilot that nobody has opened since the executive demo. The agency is still billing. The roadmap still exists in a slide deck. And no one — not the brand, not the vendor, not the internal champion — has said the obvious thing out loud: this isn't working, and it should have been shut down weeks ago.

Immersive technology has a failure problem, but not the kind most people assume. The problem isn't that pilots fail. It's that they don't fail fast enough. In our work with brands across retail, automotive, and entertainment, we've observed a consistent pattern: immersive initiatives are treated as prestige projects — showcases for innovation theater — rather than as experiments subject to the same ruthless discipline as any other marketing spend.

The remedy is counterintuitive but simple. Give every immersive pilot a hard expiration date, and mean it.

Why Immersive Pilots Refuse to Die

Traditional digital campaigns have natural terminus points. A paid media flight ends. A landing page stops converting and gets sunset. Immersive experiences lack these off-ramps. A WebAR filter or a VR training module can run indefinitely at minimal incremental cost, which makes it painless to keep them on life support — and painful to admit they never worked.

There's also a psychological component. Immersive pilots typically require visible, cross-functional buy-in. The CMO saw the demo. The innovation team championed it. Killing the project feels like killing the coalition that built it. So teams quietly extend timelines, redefine success metrics, and wait for a quarterly review where the conversation will be less awkward.

Meanwhile, the budget that could fund the next, better experiment stays locked inside a zombie project.

The 90-Day Commitment

We advocate a simple operating rule: no immersive pilot should run longer than 90 days without a formal go, pivot, or kill decision. Not a status update. A decision, made by people with the authority to reallocate budget.

The framework has three components.

Pre-committed kill criteria. Before launch, the team writes down — in plain language, signed off by the budget owner — the conditions under which the project dies. Examples: fewer than 15% of store visitors engage with the AR experience; average session duration under 90 seconds; cost per qualified lead exceeds the paid search benchmark by more than 2x. These thresholds are negotiated before anyone is emotionally invested in the outcome, which is precisely what makes them credible.

A fixed learning budget, not a fixed build budget. The 90-day pilot should be scoped as the cheapest artifact capable of testing the core hypothesis. If the hypothesis is "shoppers will use virtual try-on to reduce returns," you don't need a photorealistic 3D catalog of 4,000 SKUs. You need 40 of your best-selling items and a real analytics pipeline. Spend the savings on iteration speed, not asset volume.

A mandatory reallocation plan. This is the piece most teams skip. Before day one, decide where the money goes if the pilot fails. A kill decision without a redirection plan isn't discipline — it's just surrender, and everyone in the room knows it. Naming the successor experiment in advance transforms a failure into a portfolio move.

What Killing Early Actually Looks Like

Consider a large US specialty retailer we advised last year. The team launched a 90-day virtual showroom pilot targeting wholesale buyers with a hard criterion: at least 12% of invited buyers must place an order through the experience. By day 60, engagement was strong but conversion sat at 4%. The kill criteria triggered. Instead of extending the pilot another quarter, the team held the pre-agreed decision meeting in a single afternoon.

The instructive part is what happened next. The reallocation plan redirected the remaining budget toward a modest AR sizing tool embedded in the existing e-commerce flow — a far less glamorous concept that had lost the internal beauty contest months earlier. Within two quarters, that tool was measurably reducing fit-related returns. The showroom wasn't a wasted investment; it was tuition. The 90-day structure is what converted that tuition into a return.

The pattern repeats across categories. A quick-service restaurant chain killed a gamified AR loyalty concept in week nine when early telemetry showed open rates collapsing after the novelty window. The retained spend funded spatial audio work in a different channel that outperformed the original concept within a single quarter. In both cases, the teams credited one thing above all: the kill criteria were written before anyone fell in love.

The Cultural Obstacle

The barrier here is rarely analytical. American corporate culture rewards the narrative of persistence — the founder who refused to quit, the team that shipped against all odds. A 90-day kill decision reads, in that mythology, as failure of nerve.

It isn't. It's the opposite. Teams that operate on fixed experiment windows make more bets per year, not fewer, because capital recycles instead of pooling in half-dead projects. Over a 24-month horizon, a brand running four disciplined 90-day pilots will have more real evidence — and more shipped wins — than a brand running one multi-year "flagship" immersive initiative that never had the courage to define what success meant.

Agencies bear responsibility here, too. The commercial incentives of our industry favor extending engagements, not ending them. Which is why we now write the 90-day rule into our own statements of work. If your immersive partner resists pre-agreed kill criteria, that resistance is data.

Fail on Purpose

The most valuable sentence in any immersive program is a pre-written one: "We will stop this project on this date if these numbers don't materialize, and the budget will move here." Everything else — the demos, the decks, the executive enthusiasm — is downstream of whether that sentence exists and whether anyone intends to honor it.

Immersive technology is finally good enough that bad ideas are the primary constraint on progress, not bad tools. The organizations that will lead the next decade of spatial computing are not the ones with the largest innovation budgets. They're the ones with the shortest feedback loops and the least sentimental relationship with their own prototypes.

Ninety days. Hard criteria. A reallocation plan signed in advance. Let your worst ideas die quickly so your best ones can inherit the budget.