Client contact: +44 330 818 3103 [email protected]

February 19, 2019

Customer knowledge - when user tests are no longer enough: do you really know your customers?

Romain Secco, Product Designer at Linxo, tells us how, starting from a user test, he carries out a quantitative study / survey and then qualitative user interviews, all to improve his knowledge of his customers through 3 very specific user profiles.

Essential lessons of this conference

In order to get to the main part of this intervention, please find here the summary of our witness Linxo in video, followed by the key lessons below:

The following lessons are specific to the Linxo case:
  • Segmentation and knowledge of users is essential to meet their expectations.
  • In the first Linxo user test, we targeted people who were new to the application. This excluded existing users. Following the application's redesign, these existing users had opposite reactions: some loved it, some hated it. So, what was going on? After a quick remedial phase, the Linxo team decided to scan its existing customers to identify the various profiles.
  • The challenge of this approach is to obtain this information in order to customize the application in response to the various usages.
  • Linxo then triggers a quantitative study phase to formulate hypotheses and user insights. Then, a qualitative study validates (and invalidates!) the various hypotheses.
  • For the quantitative study of its customer database, Linxo defines a short collection period - 6 months - in order to have fresh data and keep the data volume manageable.
  • The quantitative data related to socio-professional categories could be quite similar when analyzing Linxo customers, and it was, for example, the analysis of their differing connection frequency behaviors that allowed various profiles to be identified.
  • Qualitative studies make it possible to map these various profiles, to bring to light people who did not initially seem obvious, and, on the other hand, to remove repetitious profiles.
  • In the qualitative approach, beyond 20/30 people, verbatims become useless and limit the advantage of increasing the sample size. On the targeting side, the more similar people are, the more similar their experience and therefore their verbatims will be. The sample size can then decrease.
  • If you don't have hypotheses before your tests, focus groups or individual interviews will submerge the main problems and then the hypotheses to be tested.
  • Following these different study stages conducted both with Ferpection and internally, Linxo has identified 3 profiles among existing users and is now working to provide customized responses to these profiles within its application.
  • Finally, it's important to point out that Linxo's testimony is all the more striking because it is not "standard" in its approach without it being questioned. Generally speaking, qualitative methodologies are used to explore an issue while quantitative methods are used to confirm and measure the importance of the various subjects identified. It is therefore crucial to build your methodologies according to your objectives and with the advice of an expert because each case is different.


Would you prefer to see the Linxo procedure slides yourself? They are available here on slideshare:

Do you wish to learn more about the available user research methods to better understand your users and test your products and services? Why not start with our individual user interview page or our customer knowledge page?

All articles from the category: | RSS

Thibault Geenen

Thibault Geenen

Thibault is fascinated by the power of UX, especially user research and nowadays the UX for Good principles. As an entrepreneur, he's a huge fan of liberated company principles, where teammates give the best through creativity without constraints. A science-fiction lover, he remains the only human being believing that Andy Weir's 'The Martian' is a how-to guide for entrepreneurs.

Starting a user research project? Contact us