What Is Monadic Testing?
Monadic testing is a survey design in which each respondent sees and rates only one concept, ad or product. To compare several options, you show each one to a separate, matched group of people and compare the groups' scores. Because nobody sees more than one option, no option is judged against another, only against what people already know.
Ask people to rate two ideas one after the other and the second answer is never independent. A strong first idea makes the second look worse; a weak one makes it look better. Monadic testing removes that problem by giving each person one idea only.
This guide covers how monadic testing compares with the other designs, when it is worth the larger sample, how many respondents each concept needs, the questions to ask, and how to tell a real difference from noise. It goes deeper into one part of our concept testing guide, the pillar for everything on testing ideas, ads and products before launch.
The Four Concept Test Designs
Design | What each respondent sees | Main strength | Main weakness |
|---|---|---|---|
Monadic | One concept only | Clean, unbiased scores | Needs a separate sample for each concept |
Sequential monadic | Several concepts, one after another, in rotated order | Smaller total sample | Each rating is influenced by the concepts seen before |
Comparison (paired) | Two or more concepts side by side | Shows a clear preference | Says nothing about whether either concept is good |
Proto-monadic | Several concepts one after another, then a comparison question | Ratings and a preference in one study | Has the bias of sequential monadic |
Monadic vs Sequential Monadic
The choice between them is a trade between accuracy and cost.
Monadic | Sequential monadic | |
|---|---|---|
Respondents for 3 concepts at 200 ratings each | 600 | 200 |
Order effects | None | Present, reduced by rotating the order |
Fatigue | Low: one concept, a short survey | Grows with each concept |
How close it is to real life | Close: people meet a product alone | Further: nobody shops by rating products in a row |
Best for | Final decisions between a few options | Screening many early ideas |
Why order matters. Once someone has rated one concept, it becomes their reference point. A sequential test can show concept B ahead simply because it followed a weak concept A. Rotating the order spreads that effect evenly, but it does not remove it; it adds it to every score.
A useful compromise. In a sequential monadic test with rotated order, every concept is shown first to about a third of the people (with three concepts). Those first-position ratings were given with no other concept in mind, so they work as a small monadic test inside the larger one. With 300 respondents and three concepts, that is about 100 clean ratings per concept: enough to check whether the full results tell the same story.
When to Use Monadic Testing
Use a monadic design when:
The decision is final or expensive: which product to launch, which campaign to fund, which name to register.
The concepts are similar. Small differences are exactly what order effects can create or hide.
You need scores you can compare with later tests. A monadic score can be compared with your past monadic tests; a sequential score depends on what it was shown with.
The survey is long. Rating one concept in depth is realistic; rating four in depth is not.
Use sequential monadic when you have many ideas (five or more), an early stage and a limited budget. Keep it to three or four concepts per person, drop the clearly weak ones, then test the finalists monadically.
How Many Respondents a Monadic Test Needs
Plan each concept as its own study, or cell:
150 respondents per concept is the minimum for a reliable top-two-box score.
200 respondents per concept is the practical standard.
300 or more per concept if you want to read subgroups, such as two age groups, on their own.
The total is the cell size times the number of concepts: two concepts at 200 each is 400 respondents, three is 600.
The cells must be comparable, or the difference you see may be a difference in people, not concepts. Use the same audience definition, the same screener and quotas, the same sample source and the same fieldwork dates for every cell. When the results come in, check that the cells match on age, gender and how often people buy in the category before you compare the scores.
The Questions to Ask in Each Cell
Every cell gets exactly the same survey, with only the concept changed. These are the questions in the SegmentOS Concept Testing template, in order. Replace [brand] with the name of the concept.
First, the concept itself is shown, with a short description and an image if you have one. Then:
1. What is your overall reaction to [brand]? (I love it / I like it / It's okay / I don't like it / I don't fully understand it)
2. How unique is [brand] compared to products currently available? (1 Not at all unique to 5 Completely unique)
3. How relevant is [brand] to your life or needs? (1 Not at all relevant to 5 Extremely relevant)
4. How likely are you to purchase [brand] if it were available today? (1 Very unlikely to 5 Very likely)
5. What do you like most about [brand]? Select up to 3. (Solves a real problem / Innovative approach / Easy to understand / Good value proposition / Fits my lifestyle / Design or aesthetics / I don't particularly like anything)
6. What concerns you most about [brand]? Select up to 2. (Too expensive / I don't see the need for it / Seems difficult to use / Doubt the quality / I prefer existing solutions / No concerns)
7. How much would you expect to pay for [brand]? (Under $10 / $10 to $25 / $25 to $50 / $50 to $100 / $100 to $250 / Over $250)
8. What would make [brand] more appealing to you? (Open text)
The template also includes an attention check between questions 4 and 5: "Attention check: To confirm you are reading this survey carefully, please select 'I am reading carefully' from the options below." Remove the people who fail it from every cell before you compare.
How to Read Monadic Test Results
Compare top-two-box scores. For each rating question, report the share who answer 4 or 5. For the reaction question, report the share who answer "I love it" or "I like it."
Check whether the gap is real. Take an example with illustrative numbers. Concept A gets a 26% top-two-box purchase intent from 200 people; concept B gets 34% from another 200. The gap of 8 points looks decisive, but it is not statistically significant at 95% confidence (a two-proportion test gives z = 1.75, below the 1.96 needed). With 400 people in each cell, the same scores would be significant (z = 2.47).
To detect an 8-point gap reliably, at 95% confidence and 80% power, you need about 500 people per cell. A 10-point gap needs about 330 to 360, depending on where the scores sit. If the concepts are close, either test with more people or accept that they perform about the same and choose on cost, fit or strategy. Check your own numbers with the sample size calculator and the margin of error calculator.
Read the reasons, not only the scores. Questions 5, 6 and 8 tell you why a concept won. A concept that wins on purchase intent but whose top concern is "Too expensive" may be winning on a price you cannot offer.
Apply a reality check to purchase intent. People say they will buy more often than they do. Use purchase intent to compare concepts with each other, not to forecast sales.
How to Run a Monadic Test in SegmentOS
The SegmentOS Concept Testing template tests one concept per study, which makes it monadic by design. To compare two concepts:
1. Create one study from the template for each concept, changing only the concept description and image.
2. Give every study the same audience and the same number of respondents.
3. Launch them on the same day, so the cells field at the same time.
4. Compare the top-two-box scores and the reasons between the studies.
The Concept Testing template is not one of the Premium templates, so you can publish it to your own audience on the free plan. To reach people in your category who are not your customers, the Pro plan adds a verified panel of more than 30 million respondents in 127 countries, with the cost of each study shown before you launch.
For a sequential monadic test, build one study with each concept and its questions in a section of its own, and turn on Randomize Sections in the study's Randomization settings. Each respondent then sees the concepts in a different order. Keep the screener outside the sections so it always comes first.
Common Mistakes in Monadic Testing
Cells that do not match. Different audiences, dates or sources make the comparison meaningless.
Changing more than the concept. Same questions, same order, same length in every cell.
Calling a winner on a small gap. At 200 per cell, gaps under about 9 points are usually within the noise.
Testing too many concepts monadically. Five concepts at 200 each is 1,000 respondents. Screen first.
Reading purchase intent as a sales forecast. It ranks concepts; it does not predict demand.
Key Takeaways
What it is: each respondent rates one concept only, and separate matched groups rate the others.
Why it matters: no concept is judged against another, so the scores are free of order effects.
The cost: a separate sample per concept, 150 at the minimum and 200 as standard.
The alternative: sequential monadic is cheaper and fits early screening, but its scores are biased by order.
Check significance: at 200 per cell, an 8-point gap is not yet a real difference.



