Chapter 6
Turn the numbers into one experiment
Finish this chapter with a one-experiment brief you can use immediately to turn the numbers into one experiment, while avoiding mistaking activity for revenue progress.
What you’ll get
- A completed one-experiment brief, not a list of abstract ideas.
- One explicit decision rule for turn the numbers into one experiment.
- A 45-minute execution sequence you can repeat yourself or hand to a teammate.
- A quality gate that tells you whether to move forward, run one small test, or revise the input.
Use this chapter as a working session
Do not read this chapter passively. Open the document, spreadsheet, or Tapmy workspace where the decision will live. Replace sample language with the exact offer, buyer, channel, price, and evidence you have today. If a required fact is missing, label it as an assumption and give it a smallest-possible test.
| Time | Work | Concrete output |
|---|---|---|
| 0–10 min | Collect the evidence already available: buyer words, traffic data, sales conversations, support questions, and current page copy. | A short evidence list with source and date. |
| 10–20 min | Make the single decision this chapter is designed to resolve. Do not combine it with a second optimization. | A one-sentence decision and the reason behind it. |
| 20–35 min | Build the one-experiment brief with real names, numbers, links, and owners. | A usable operating asset, not a draft outline. |
| 35–45 min | Review it from the buyer’s point of view and run the quality gate below. | Keep, test, or revise decision with one next action. |
Work through the chapter
Use this experiment card:
- Observation: Product clicks are healthy, but few people start checkout.
- Hypothesis: Buyers cannot quickly see what is included.
- Change: Add a three-item deliverables block and one product preview above the CTA.
- Primary metric: Checkout starts ÷ product clicks.
- Guardrail: Completed purchases and refund/support messages.
- Decision date: After the next comparable campaign window.
Change one primary variable whenever possible. If you change the price, headline, image, CTA, and traffic source together, a win teaches you very little.
Decision gates
| Evidence state | What it means | What to do next |
|---|---|---|
| The input is specific, current, and tied to a real buyer action | The one-experiment brief is ready to use. | Put it into the Tapmy journey and measure the next observable behavior. |
| The direction is plausible but one important input is assumed | The decision is testable, not yet proven. | Run the smallest test that can confirm or reject that one assumption. |
| The asset tries to serve several audiences, offers, or goals at once | The scope is too broad to produce a clean result. | Narrow to one buyer, one job, one offer, and one reporting window. |
| The result depends on a claim you cannot show or deliver | Trust and delivery risk are higher than the upside. | Rewrite the promise around an outcome you can support and verify. |
Quality standard
Before moving on, confirm that:
- the one-experiment brief names one buyer and one job;
- every number has a timeframe, currency, or denominator where needed;
- every link and handoff has an owner;
- the next action is visible without rereading the chapter;
- the decision can be measured with a real customer behavior, not a vanity metric;
- no claim is stronger than the available evidence.
Before you move on
- The one-experiment brief is complete with real inputs.
- One assumption is marked for testing, if needed.
- The owner and next review date are recorded.
- The next customer-visible Tapmy action is clear.