What is Message Testing?

Message testing is the evaluation of communication with members of its target audience to measure comprehension (do they understand what it says?), relevance (does it speak to a need they have?), credibility (do they believe it?), differentiation (does it sound different from the alternatives?), and response (does it make them want to learn more, try, or buy?). The message can be as small as a button label or as large as a company's positioning, and the methods scale accordingly, from a five-second test on a headline to a monadic study in which hundreds of respondents each evaluate one of several positioning statements. It is the communication counterpart of concept testing (which evaluates the idea) and of value-proposition testing (which evaluates the promise), and it borrows from both.

What to Measure

Comprehension. Ask respondents to explain the message in their own words, and code whether they got it. The most important measure and the one most often skipped in favor of "which do you like?".

Relevance and importance. Does the message address something they care about? "How relevant is this to you?" on a labeled scale, plus which part mattered.

Credibility. Claims that sound too good produce skepticism; a believability rating and an open "what makes you doubt this?" catch it.

Differentiation. Could this be any competitor's message? Shown alongside anonymized competitor claims, does it stand out, and on what?

Response and intent. Interest, likelihood to click, try, or buy (purchase intent), read comparatively across messages rather than as forecasts.

Language. The words respondents use to restate the message are the words the audience thinks in; they routinely improve the copy more than any score does.

The Methods

Monadic testing. Each respondent sees one message and rates it; messages are compared across matched groups. Clean, no order effects, sample-hungry (a few hundred per message); the standard for high-stakes positioning.

Sequential monadic. Each respondent sees several messages in rotated order and rates each; more efficient, with order effects managed by randomization.

Comparative and preference formats. Messages shown together, respondents pick the strongest and explain why; fast and useful for narrowing many drafts to a few, with the usual caution that preference is not performance.

MaxDiff on claims. A MaxDiff across a list of candidate claims or benefits produces a ranked importance that a five-point scale can't.

Qualitative reaction. Recorded first reactions and restatements from a handful of target users (a Ballpark study shows the message and captures the response on video within seconds of exposure) reveal the misreadings and the language that surveys score but can't hear.

Live experiments. Once a message is in production, an A/B test on the landing page or the email measures the behavior the survey estimated.

Running It Well

1. Test with the audience the message is for, screened behaviorally; messages land differently on people who have the problem.
2. Show the message in context (on the page mock, in the email, on the ad) rather than as a sentence in a box; context changes comprehension.
3. Ask comprehension before evaluation, and open questions before scales, so that reactions aren't primed by the rating task.
4. Test messages that genuinely differ; five variations of one idea produce five near-identical scores.
5. Keep a fixed set of measures across tests so that messages can be compared over time, and read comprehension and credibility as gates: a message that isn't understood or believed shouldn't advance on preference.

The Bottom Line

Message testing checks whether communication is understood, relevant, believed, distinct, and motivating, with the people it's aimed at, before it ships. Measure comprehension first, show messages in context, test real alternatives, use monadic designs when the stakes justify the sample, capture the audience's own words, and confirm the winner with a live experiment. The writer knows what the message means; the test finds out what it says.