wiki / Marketing / marketing-message-testing-matrix
advancedMarketingmessagingpaid-adspositioningab-testingstrategy

Build a message testing matrix across awareness stages

Maps your value propositions against buyer awareness stages, picks the four highest-information test cells, and drafts paired ad executions for each.

The prompt — fill the [BRACKETS]0 copies
Build a message testing matrix for [PRODUCT OR SERVICE] so I can learn what to say before scaling spend on [PRIMARY PAID CHANNEL].

Inputs:
- Value propositions we believe are true: [LIST 3-5 VALUE PROPS]
- Audience segments: [1-3 SEGMENTS, ONE LINE EACH]
- What buyers currently do instead: [STATUS QUO OR COMPETITOR BEHAVIOR]
- Constraints: [BUDGET LEVEL, BRAND RULES, CLAIMS WE CANNOT LEGALLY OR HONESTLY MAKE]

Step 1 — Matrix. Build a table: rows are the value propositions; columns are three awareness stages — problem-aware (feels the pain, does not know solutions exist), solution-aware (comparing options), and product-aware (knows us, has not bought). In each cell, write the one-sentence core message that value proposition becomes at that stage. If a proposition does not work at a stage, mark the cell weak and say why in a few words — an honest sparse matrix is more useful than a filled one.

Step 2 — Prioritization. Pick the four cells to test first. Justify each pick on two axes: how different the candidate messages are from one another (tests that cannot produce a clear winner waste spend), and how large the audience at that awareness stage likely is, reasoning from the status-quo input rather than invented market data.

Step 3 — Creative. For each of the four cells, write two ad executions for [PRIMARY PAID CHANNEL] in that platform's format: headline, body, CTA. The two executions share the cell's core message but differ in execution style — for example direct claim versus scenario — so message effects and execution effects stay separable.

Step 4 — Decision rules. State what to compare after the test: cell versus cell first, then execution versus execution within winning cells, and what each outcome implies doing next. No fabricated benchmark numbers — express thresholds as relative comparisons only.

Every claim in every execution must respect the constraints input.
Open in Promptly Suggest an edit
Updated 2026-08-18 1 revision
Crafted with precision by MnexWeb