Ten new ads do not automatically create ten useful learnings. If the audience, offer, message, format, and landing page all change together, you may find a combination that performs differently without understanding why.
Give each test a decision to support. That decision should be narrow enough that the result helps you choose the next action.
1. Write the question before the creative brief.
Replace “We need fresh creatives” with something specific: “Do suitable prospects respond more strongly to the time this service saves or the uncertainty it removes?” This gives the designer a clear brief and gives you a reason to compare the results.
Hypothetical test brief
Question: which message leads to more qualified enquiries for a bookkeeping service?
Version A: focus on time spent doing the books. Version B: focus on clarity about the monthly numbers. Keep the underlying service, audience, format, and destination consistent.
2. Decide what would count as useful evidence.
Choose the business outcome before launch. For the example above, that could be cost per qualified enquiry, using an agreed qualification definition. Click-through rate can help explain how people respond to the message, but it does not establish lead quality.
- Keep a reference: compare against an existing version or a clearly defined alternative.
- Control the comparison: use a properly configured split test when available. Ordinary ad delivery can distribute exposure unevenly.
- Set boundaries: document a budget, review window, and stop conditions before the test runs.
- Match the journey: make sure the page fulfils the promise made in each ad.
There is no universal number of days that makes a result conclusive. Traffic, conversion frequency, and the size of the difference all matter. If the evidence is thin, mark the result as inconclusive instead of forcing a winner.
3. Record what changes next.
Keep a simple log: question, versions, dates, spend, qualified outcomes, caveats, and next action. Include unexpected observations from the sales team, such as prospects arriving with the wrong expectation.
A useful next step might be refining the stronger message, testing a different objection, or repairing the page before producing more ads. Save the learning alongside the creative so the next round starts with evidence.
A good test brief ends with: “If we learn this, here is the decision we will make.”
Read next