Home/Resources/Comparisons/ Synthetic respondents
Pretotyping and synthetic respondents: when each method fits.
Synthetic respondents and digital twins use AI to simulate answers from target groups. Pretotyping measures the behaviour of real people. Horizon runs pretotyping as a painted-door test.
Compared
Two methods, two strengths.
Both help to decide faster. The difference lies in where the signal comes from: a model or people in a buying situation.
Context
Where the methods differ.
Source
Synthetic respondents derive answers from existing data. Pretotyping observes real people.
Novelty
For new offers, the model lacks a basis. Pretotyping measures behaviour here too.
Use
Synthetic respondents deliver first hypotheses in minutes. Pretotyping delivers, in around four weeks, the robust basis on which an investment can rest.
Combination
Combining both methods.
Synthetic respondents
Gather ideas, objections and hypotheses.
Build variants
Select up to six of the strongest variants.
Pretotyping
Test the variants on the behaviour of real people.
Decision
With AI exploration and behavioural evidence.
More comparisons
Pretotyping compared with other methods.
Method
Pretotyping and conjoint
Weighting features in a survey or testing variants on behaviour.
To the comparisonMethod
Pretotyping and surveys
Stated willingness to buy or measured purchase intent.
To the comparisonMethod
Pretotyping and Van Westendorp and Gabor-Granger
Stated willingness to pay or prices compared on behaviour.
To the comparisonMethod
Pretotyping and concept tests
A rating in a survey or measured purchase intent in the online environment.
To the comparisonMethod
Pretotyping and A/B tests in the shop
Optimising what is live or deciding before investing.
To the comparisonProvider
Horizon compared with providers.
Provider
Horizon and Appinio
Surveys via an app panel and a behavioural test in the online environment.
To the comparisonFAQ
Frequently asked questions.
Will AI soon replace tests with real people?
AI already helps with exploration today. The more expensive and newer the decision, the more important the behaviour of real people becomes.
Does Horizon use AI itself?
Yes, in setting up and running tests. The signal always comes from real people; metrics are not generated by a language model.
Is anything sold with pretotyping?
No. No contract is concluded and no payment is taken. Anyone who chooses an offer is then told transparently that the offer is not yet available.
How long does a test with Horizon take?
Around four weeks from question to data analysis, depending on the target group and the number of variants.
Which method fits your question?
Bring your decision question and we will show you which method fits and how the two work together. Your insights team is welcome to join.
You will speak with Daniel Putsche
Founder & CEO, 30 minutes
Read more