Find the message people actually understand
People can prefer a headline without understanding it. Message testing shows each participant one randomly assigned message and asks them to explain the offer in their own words, so you measure comprehension, not taste.
Why this is hard today
Guesswork
A/B tests need traffic you do not have yet.
Wrong signal
Preference polls reward clever copy.
Slow answers
Teams argue about wording without evidence.
How Scrutinizr does it
- 01
Add variants
Two to four headlines, value propositions or ads.
- 02
Randomise
Each participant sees one variant only.
- 03
Ask them to explain it
Unaided comprehension, relevance and believability questions.
- 04
Pick the winner
Compare variants with sample sizes shown.
The result
- Comprehension rate per variant
- Common misunderstandings
- Relevance and credibility scores
- The quotes behind each result
Example: “delivery” confusion
Message A was read as a grocery delivery service by a third of people. Message B, which led with planning, was understood by 68%.
Good to know.
How is this different from an A/B test?
It works before launch with 50 to 100 people and tells you why a message works, not only which converts.
Can I test ads and landing pages?
Yes. Upload screenshots or paste copy for any variant.
Can experts review the copy too?
Yes. Add a marketing expert review to the same test.
Make your next move with evidence.
Start with a question and bring your websites, documents, data and customer voices into one evidence trail.