Glossary
What is message testing?
Definition
Message testing is the practice of evaluating alternative messages with a real audience to learn which framing, claim, or angle lands before committing spend to it.
The survey toolkit includes monadic tests (each respondent rates one message in isolation), sequential monadic, and MaxDiff (forced trade-offs that rank many messages efficiently). The in-market toolkit is live A/B testing across ads, landing pages, and email. Qualitative work sits underneath both, explaining why a message lands rather than just whether it does.
Discipline matters more than method. Test the claim, not the execution: if two variants differ in message and image and layout, the winner teaches you nothing. Keep exposure realistic, decide the success metric before fielding, and size the test so the answer is signal rather than noise. Test with the audience that will actually buy, too, not the panel that happens to be cheapest to reach.
The structural pitfall is forced exposure. A panel respondent who must read your message and answer questions behaves nothing like someone thumbing past it in a feed, and stated preference reliably diverges from revealed preference. Messages that win in the lab can lose in market, which is why experienced teams treat survey tests as a filter and in-market response as the verdict.
How this shows up in Waldo
Before testing your own messages, it helps to know which angles the category has already voted on. Waldo’s /category/{id}/angle-library returns the messaging angles running across a tracked category, and the audience conversation around them shows how each was received: a cheap way to shortlist candidates before formal testing spends a dollar.
Related terms and reading
Put Waldo behind your agents
Brand, category, and audience intelligence over 200+ API and MCP endpoints. Sign up, mint a key, and run it against the brands you actually track.