Here is a summary of questions our users often ask us. If there is anything else you’d like to know, please do not hesitate to contact us.
Yes, you are welcome to sign up for free. You can run small studies with fewer than 20 responses without charge. Larger studies attract fees.
You only need to pay for the experiments where you want to analyse more than 20 responses. You can collect any number of responses without paying first, and only pay when you are ready to see the reports. Generally, experiments with 20 respondents or fewer are useful for testing the tool, but typically cannot provide reliable results. If you want to test with ~50 survey testers, you can use our Quick Feedback functionality.
Yes, you can add any number of additional questions (multiple choice, free form text, number, emails, constant-sum, Likert scale, etc.). There is no cost to adding those.
Yes, we will gladly help you with any customisation required. Please contact us for a quote.
Yes, it is possible to change all text of respondent-facing survey pages to be in your language. You can do so in the ‘Choose template’ tab. Our users have run studies in Chinese, Polish, Arabic, and other languages. If you have any questions while setting up, please let us know.
Yes, but you should keep the number of attributes and levels low. Refer to the automatic sample size calculation when you set up the experiment for guidance on how many respondents you need depending on the settings you specify.
Yes, you can add pictures in a variety of popular formats (JPEG, PNG, etc.).
There is a limit to prevent unreasonably large numbers, but our users have not come close to it. If your study requires additional attributes and levels, please contact support. However, practically, it is often unreasonable to have more than six attributes, and about ten levels per attribute.
As indicated above, the system practically does not limit the number of attributes. However, respondents will find it difficult to process information on more than 6 attributes because of the high cognitive load. If you do have a need to test a dozen or more attributes we suggest one of two options:
“None of the above” option should be included in most marketing studies because it reflects actual market behaviour of not buying a product. Forced choice (i.e., where there is no “None of these” option and participants are forced to choose between alternatives) may be relevant in medical applications of conjoint analysis (e.g., when choosing treatment for a patient).
Yes, but we recommend specifying only a few for statistical reasons.
Yes, you can supply your own design for our DIY conjoint product.
Elasticity curves are not delivered automatically at this stage, but we do perform this analysis as a separate service. Please contact us to discuss this option.
Yes, see brand-specific conjoint.
Yes, it is done automatically. The more complex the design, the higher the number of blocks.
Conjoint.ly uses Markov chain Monte Carlo Hierarchical Bayes (MCMC HB) estimation to calculate individual-level coefficients. These coefficients are used in market share estimation, helping account for heterogeneity in the market. They are aggregated at the market and segment levels to show partworth utilities.
Bayesian hierarchical modelling is a statistical model in conjoint analysis is a type of modelling that estimates parameters (partworth utilities) not for market as a whole, but rather for individuals. The word “hierarchical” refers to the nested structure (individuals are nested in the market). “Bayesian” refers to the statistical paradigm of Bayesian statistics that is based on Bayes' theorem. Amongst other benefits of HB, this approach allows more parameters (attributes and levels) to be estimated with smaller amounts of data collected from each respondent.