Online research methods to explore your target audiences

and simulate their real-world choices

Exploratory surveys correspond to the classic questionnaires you're used to. These questionnaires are perfect for discovering the underlying structure and thought patterns of a population you're exploring. In this type of survey, what matters most is minimizing biases to avoid steering respondents' answers.

And conducting statistics (PCA, CA, SEMs) to avoid a superficial reading of results: indeed, raw results and percentages can often be misleading if not cross-analyzed. But this type of questionnaire only investigates the declarative (what respondents believe or what they're willing to show) and you must therefore be careful to stick to very concrete usage questions rather than pushing respondents toward projection.

In this type of questionnaire, we don't just ask users what they believe and think. They're not always aware of their deep motivations. We force them in questionnaires to take a position, to choose between different options. And we then introduce these trade-offs into a mathematical model that calculates participants' real profiles and their true preferences, which allows us to get much closer to the reality of their needs.

Realistic simulation surveys are Randomized Controlled Trials (RCT), meaning genuine scientific experiments where two (or more) distinct subject groups are presented with different questionnaires to evaluate the real impact of choice situations. A bit like an A/B test.

But we use games and behavioral economics models here that allow us, like a digital twin, to realistically simulate real-life choice situations and capture individuals' true intentions and true preferences in order to build a democratic choice space and informed consent.

Our Psychologists, Sociologists, Behavioral Economists & Statisticians master these three levels and help you choose the method suited to your strategic challenges.

• Creating a reliable questionnaire

Since the advent of online questionnaire tools such as Google Sheet or Survey Monkey, most people think that conducting an online survey is a quick and easy process: you brainstorm, write down a few questions that you enter into an online questionnaire tool, launch the survey on the web like a bottle in the sea and then take a look at the pie charts automatically produced by the publishing site to get reliable answers to your questions.

What we've been teaching our students in human sciences, marketing and UX for the last 12 years is that these different stages are all pitfalls that will skew your results, make you make mistakes and lead you to make the wrong strategic decisions.

The online survey is the most delicate exercise in data recovery: at every stage, you need the support of a specialist in human sciences and statistics to ensure that the answers are correct: ask unbiased questions, make sure the test sample is valid, check the results with statistical tests rather than flat sorting, etc.

• Distribute your questions widely

The Internet makes it possible to survey a larger sample in a shorter time. This improves the reliability and accuracy of the survey results. An online survey can also overcome geographical constraints by surveying an audience in several countries. A considerable advantage when conducting market research for international expansion.

• Create continuous questionnaires to improve response rates

The permanent accessibility of online questionnaires makes it possible to reach people who cannot easily be reached by traditional means or who are not available at normal times. It is therefore often useful for obtaining a high response rate with a view to penetrating new markets. The response rate is also improved by the possibility offered by online surveys of automating the skipping of question categories, or of inserting multimedia documents to make the questions clearer and more attractive.

• Statistical tests to ensure the reliability of responses to your questions

Software used to generate online questionnaires and surveys is generally equipped with datavisualisation functions and therefore provides structured results reports in the form of graphs and other summary tables. But beware: these graphs are not statistics: they are flat sortings, simple aggregations of results that are generally misleading as to the true direction taken by users. Only statistical tests carried out by our statisticians at ENSAE and Modalisa will enable you to discern the true from the false and the sometimes complex behaviours that hide behind appearances.

This will guarantee you real strategic advantages in terms of understanding users and defining the most relevant value propositions.

• Improve data security and confidentiality 

Solutions for creating online surveys ensure the security and confidentiality of company data, as well as compliance with the RGPD, which is built into the software.