How is discovery different from a survey?

A survey answers the questions you already had. Discovery reads the market before the plan exists: which segments are in it, how large each is, and what opens and closes a sale, so the segment you did not brief for shows up with its size.

Can it find segments I did not define?

Yes. Because the audience is built from a population rather than from your quotas, a read can be broken down by any field the audience carries, and a group you did not plan for appears at its real size. In the market-to-scale tile above, that group was the 19% who had never once compared brands. A study that only recruits people who say they are considering a change never sees them.

How many people should a discovery read include?

A few hundred to a few thousand, depending on how many segments you want to size. The read above used 900 so that switchers, habitual buyers and non-buyers each had enough people to break down separately.

Which task types does a discovery study use?

Usually a survey for the sizing, interviews or a focus group for the reasons, and follow-ups to the segments that surprised you. All run on the same audience, so the reasons come from the same people as the numbers.

Discover · Synthetic market research

Understand any market in an afternoon.

Who buys, why they buy, and what would change their mind. Ask thousands of Synthetic Individuals about any market and read the answers in hours, not weeks.

Map your market

Plant-based drink buyers · Germany

SurveyFind out which brands they buy today, how often, and what would actually make them change.

900 respondents

Benchmarked in public. Trusted by teams at global brands.

See details

Match with real surveys

89%

against a 93% human test-retest ceiling

Answer consistency

95%

across 12 rephrasings of a question

Studies run

10K+

surveys, interviews, focus groups

Sources · iMario Synthetic Audience Benchmark, public method and raw files · Pew Research Center, survey methodology · Updated 2026-09-05, reviewed by iMario research

In short

Most research arrives after the decisions are made. iMario asks the market first, so the plan starts from what buyers want, not what the category assumes.

The problem

You are deciding for people you have never met.

[1]

1–2 quarters

What a global segmentation typically takes end to end. The strategy call that needed it happened in month two.

[2]

1 wave

Answers the questions you wrote in March. The one your CMO asks in May is not in it.

[3]

6 weeks

To re-field one question you did not think to ask. You can only cross-break by the variables that were in the wave.

How it works

Ask a market a question, and get the market back.

Pick a task, point it at a population you describe in one line, and read the answers as segments with real sizes. No screener to write, no quotas to agree.

Northwind/New task12,480 credits

Hello, I'm iMario.

What do you want to find out?

Plant-based drink buyers · Germany

SurveyFind out which brands they buy today, how often, and what would actually make them change.

900 respondents

Northwind/Report12,480 credits

Result

What would make you change brand?

It tastes better41%
Costs less28%
Better for me21%
Nothing would10%

Crosstab

Would change brand · by buyer type

	Switchers	Habitual	New
Yes	34%	11%	6%
Maybe	39%	28%	17%
No	27%	61%	77%

What you get

A market you can question, not a deck.

34%

47%

19%

Would change brand
Buy on habit
Never compared

The market to scale

Every segment with its real size and its recoverable share, not a persona board.

“I switched for taste, not the health thing. The health thing is only why I looked.”

Changed brand last year · Hamburg · 34

The reasons in their words

Open answers coded into themes, with the verbatims kept and attributable to people.

34%

An audience that stays

The audience stays in your workspace, so next quarter’s follow-up costs one afternoon: the same market, asked the next question, with no new panel to recruit.

Compared

A segmentation study and a discovery read, side by side.

The same market question, answered the usual way and on iMario. The left column is a typical agency schedule for a global segmentation; the right is the run shown on this page.

Time to a read: Traditional segmentationOne to two quarters end to end: brief, screener, quotas, fieldwork, analysis.; Discovery on iMarioThe same afternoon. The first answers land in minutes and the full read in hours.
Who answers: Traditional segmentationA recruited sample, capped by quotas set before anyone knew the market.; Discovery on iMarioA population you describe in one line, built from official statistics and your own data. No screener, no quotas.
The question you forgot: Traditional segmentationSix weeks and a new wave to re-field one question.; Discovery on iMarioAsk it now, of the same audience, and read it beside the first wave.
Breaking it down: Traditional segmentationOnly by the variables that were in the wave.; Discovery on iMarioBy any field the audience carries, so a segment nobody briefed for appears at its real size.
Cost: Traditional segmentationAgency fieldwork, recruiting and incentives, for every wave.; Discovery on iMarioCredits, from one per respondent per question. The follow-up wave costs an afternoon, not a budget line.

A segmentation study and a discovery read, side by side.
Decision	Traditional segmentation	Discovery on iMario
Time to a read	One to two quarters end to end: brief, screener, quotas, fieldwork, analysis.	The same afternoon. The first answers land in minutes and the full read in hours.
Who answers	A recruited sample, capped by quotas set before anyone knew the market.	A population you describe in one line, built from official statistics and your own data. No screener, no quotas.
The question you forgot	Six weeks and a new wave to re-field one question.	Ask it now, of the same audience, and read it beside the first wave.
Breaking it down	Only by the variables that were in the wave.	By any field the audience carries, so a segment nobody briefed for appears at its real size.
Cost	Agency fieldwork, recruiting and incentives, for every wave.	Credits, from one per respondent per question. The follow-up wave costs an afternoon, not a budget line.

Case study · IFOP

Reaching the people fieldwork struggles to reach.

Hypotheses, stimulus and sampling priorities were sharpened before a single person was recruited, so the human study started from a stronger question.

Synthetic consumers run upstream of the study, mapping attitudes, objections and segment differences before anyone is recruited.

See more details

Where they were

IFOP is a market research agency. As AI arrived in the industry it looked past faster moderation to a harder constraint: finding and engaging the right people when an audience is small, specialised, sensitive or geographically dispersed.

Before
fieldwork

AI moved upstream instead of automating the interview itself.

Hard-to-reach
audiences

Explored early, where recruiting is slowest and most expensive.

Where the synthetic read sits

Synthetic readattitudes and objections mapped
Hypotheses sharpenedstimulus and sampling priorities set
Fieldworkstarts from a narrower question

FAQ

Questions,answered.

Straight answers before you start. Anything else, ask us live.

Book a demo

Market research run on Synthetic Individuals, simulated people built from real customer data and a modelled population of the market. You ask them what you would ask a panel, and read the answers as distributions and as individual reasons, in hours rather than weeks.