Collect observations that can be checked
Record a source link, advertiser, visible offer, format, opening message and capture date. Note what is unknown rather than filling gaps with confident guesses about targeting, spend or performance. Retain only the material and access the team is permitted to use.
Build the evidence table
Scroll horizontally to read all columns.
| Field | What belongs here | What does not |
|---|---|---|
| Observation | A visible hook, format or claim with a source | An unverified return on ad spend |
| Interpretation | A clearly labelled possible pattern | A claim that every similar ad performs well |
| Original hypothesis | Your offer and a specific change to investigate | A competitor’s material copied into a new ad |
| Later result | Your recorded test conditions and observed outcomes | Library visibility treated as your campaign result |
Write a brief someone can approve
- Define the offer, intended audience hypothesis and one creative question.
- Choose a small varied reference set and preserve source dates.
- Write one reason the pattern may matter and one reason it may fail.
- Create an original message and authorized asset list.
- Specify a later test measure, budget and stopping rule without inventing results.
A worked planning example
A hypothetical product team notices that several references demonstrate a use case before showing a feature list. It proposes a short original demonstration of its own product. The hypothesis is that the explanation may improve understanding for the intended audience. Whether it improves any campaign outcome remains unknown until the team runs and records its own evaluation.
Use Atria for the research requirement
Review the Atria profile for the saved-reference and creative-research workflow. Inspect workspace access, exports, search conditions and any required AI allowance directly. DataHawk addresses a different question involving authorized marketplace data. Neither profile establishes future campaign performance.
Working files
Blank templates for your own planning. These contain no customer records and are not submitted to Sales Fit Lab.
A blank planning file. Record your own observations; this is not a completed product test.