# Plan an acquisition experiment

Prepare an experiment brief to test an acquisition assumption. This procedure produces a plan; it does not send messages, publish an offer, or authorize spend.

## Prerequisites

Choose an owner and a question from an existing offer, lead, or fulfillment proposal. Gather the evidence already available. Check whether the described product behavior is implemented before testing it as an available offer.

1. Write one hypothesis. For example, investigate whether a specific audience understands an operational replay better than a generic product tour. Record why existing evidence leaves the question unresolved.
2. Choose the smallest useful method. The [measurement reference](/stormtrust/acquisition/leads/measurement-and-experiments/) distinguishes early interviews and response evidence from traffic-based comparisons. Do not assume sufficient traffic or instrumentation exists.
3. Define one primary measure, its source, and a decision rule before collecting results. Add quality checks such as qualified responses, unsupported claims, incomplete data, and downstream usefulness. Use the [scoreboard](/stormtrust/acquisition/money-models/scoreboard/) for definitions.
4. Document the audience, variant identifiers, required assets, owner, duration, capacity, and any approval needed for external activity. Keep customer records out of this public repository. Proposed analytics event names are not proof that an event is instrumented.
5. Review the plan, then conduct only the approved activity. Record observations, missing data, and outcomes, including a negative or inconclusive result. Keep the original hypothesis beside the result.

## Verify the result

The record should explain what was tested, what was observed, what remains uncertain, and whether to retain, revise, or stop the proposal. Do not convert an interview preference into a revenue or conversion claim.