All writing

A launch plan needs a kill question

Teams define what success looks like before launch. They should also decide which evidence would make them stop, narrow, or change the bet.

4 min readBy Kritika Rastogi

Launch planning is built around positive motion.

The team defines the audience, promise, channels, assets, milestones, and success metrics. The underlying assumption is that once the product enters the market, the work is to make it travel farther.

Some launches should not travel farther. They should be narrowed, repositioned, delayed, or stopped.

That decision is difficult to make after launch because effort creates attachment. A team that has spent months building the product and campaign can explain away weak evidence for a long time. The useful moment to define a stopping condition is before the results arrive.

Write the kill question

A kill question names the evidence that would make the team reconsider the launch thesis.

Examples:

  • If the target audience understands the promise but does not begin the core workflow, what assumption about urgency or usability was wrong?
  • If activation is strong but repeated use is weak, does the product solve a recurring problem?
  • If users adopt but the intended buyer will not pay, is the audience, package, or economic story incorrect?
  • If deals progress until security review and then stop, is the launch premature for the segment?

The word “kill” does not mean every miss ends the product. It forces the team to name what evidence would be serious enough to change direction.

Pair each success metric with a decision

A metric is more useful when the team knows what it will do at different values.

For a new workflow, the plan might say:

  • Strong activation and repeat use: expand distribution.
  • Strong activation and weak repeat use: investigate the value moment and frequency.
  • Weak activation after qualified interest: revisit onboarding and the promise-to-product gap.
  • Weak qualified interest: reconsider the audience or problem framing.

This turns the dashboard into a decision system. It also reduces the temptation to substitute reach metrics when product behavior disappoints.

Protect the learning window

Stopping conditions need a fair test. Define the audience, time period, minimum sample, and product readiness required before judging the result. Otherwise, a team can kill a good idea after a noisy week or protect a weak one indefinitely.

The launch brief should state both sides of the bet: what the team expects to happen and what would cause it to think again.

Optimism gets a product into the market. Decision discipline determines what the company learns once it arrives.