How to Recover a Stalled Experimentation Programme

Answer these questions about your programme over the last six months: How many experiments were completed? How many had a decision recorded? How many results were communicated to stakeholders? When was the last brainstorming session, backlog grooming, review board meeting? Is there an executive sponsor who actively references the programme? If the numbers have declined, the rituals have lapsed, and the sponsor is absent, your programme is stalling.

Diagnosing the Stall

Common causes: loss of a key person, leadership attention shifted, no visible impact, quality erosion, process fatigue, resource squeeze, isolation.

Checkpoint: Name the person whose departure would most damage your programme right now. What would be lost that is not documented anywhere else?

Checkpoint: When was the last time your executive sponsor mentioned the experimentation programme unprompted?

Checkpoint: Pull your last five experiments. Do they all have a clear hypothesis, a completed measurement plan, and a decision protocol written before launch?

Checkpoint: Can you produce six months of programme data without significant effort? If pulling that data requires opening multiple tools, asking colleagues, and piecing the picture together, that effort level is itself diagnostic.

Recovery

Step 1: Secure sponsorship. A programme without an executive sponsor will stall again.

Step 2: Start small and visible. Two or three well-designed experiments in one strategically important area. Execute flawlessly. Document thoroughly.

Step 3: Rebuild the foundations. Address the structural cause.

Step 4: Re-establish the cadence. Reintroduce the rituals.

Checkpoint: Which rituals were happening before the stall and are not happening now? That list tells you what infrastructure the programme was relying on informally.

Step 5: Prevent recurrence. What will prevent this from happening again?

What Good Looks Like

A recovered programme is not just running again. It is running on a stronger foundation. The question is not how to restart. It is what you are restarting on top of. Restarting the activity without fixing the structure is recovery theatre.