Write the stop rule before launch

A campaign threshold protects the decision from the optimism, fatigue, and sunk cost that appear after the money is spent.

Write stop, refine, and continue rules before launching a test. Define the fixed budget and window, the costly action that matters, the spread between angles that would count as meaningful, and the result that would invalidate the current problem or audience hypothesis. Prewritten rules prevent sunk cost and post-hoc storytelling from moving the threshold.

Key takeaways

  • Every possible result should lead to an action.
  • Thresholds protect judgment after sunk cost appears.
  • Technical failure is different from market failure.
  • Changing the rule creates a new test.

A test is easiest to interpret before it begins. Nothing has been spent, no creative has become a favorite, and no one has explained the launch to a team. Afterward, every ambiguous number arrives with a story about why one more week might make it clear.

Define three exits

Stop when no angle produces meaningful action and there is no qualitative recognition. Refine when attention appears but the next step fails, suggesting an offer or handoff problem. Continue when one angle clearly wins on the action closest to intent. The categories matter more than universal benchmark numbers.

Separate broken delivery from weak demand

A rejected ad, dead form, incorrect pixel, or audience too small to deliver invalidates the experiment. It does not invalidate the idea. Write technical minimums beside market thresholds so a failed setup is repaired rather than interpreted and a functioning campaign is not excused as a setup problem.

Record exceptions instead of hiding them

Sometimes new information justifies another run. Preserve the original result and write why the next test differs. This keeps the evidence trail honest. Quietly extending the window or changing the audience turns one experiment into several overlapping experiments with no clean conclusion.

The threshold is a promise made by the version of you who has not spent the money yet.

Frequently asked questions

What decision rules should I set before an ad test?
Set a fixed budget and window, technical delivery minimums, the primary costly action, and explicit outcomes for stopping, refining, or continuing. Record any later rule change as a separate test.

Keep reading

  • GTM for everyone

    Go-to-market should not begin with a department, a deck, or a giant budget. It begins when one person describes a problem clearly enough to test it with strangers.

  • Your notes are already market research

    Voice memos, call notes, and half-finished documents already contain the customer language your first campaign needs. The work is finding the repeated detail.

  • The first brief should sound spoken

    A useful first brief sounds like the way you explain the problem to another person—not the way a company describes itself in a document.