This article is part of our blog series on panel management. In this blog series, we’ll present the work of our panel managers and share their daily challenges and considerations with you. Please have a look here to get an overview over all topics covered by this series.
In this article, we are still answering the question who all these people are, that join our panel. As we mentioned in previous articles, we try to get to know them as good as possible, the sooner the better.
When they register to the panel, we collect some initial information about the person, like the email address, the name, the age and the gender. This has mainly to do with legal requirements, especially to make sure that they are of legal age to share personal information with us. However, even afterwards, it is important to getting them to know a bit better and this is what profiling is all about.
In theory, the purpose of profile information is to invite panel members only to studies they are eligible for and, as a consequence, to improve their membership experience. Unfortunately, this does not always work out in practice. We frequently have to screen out participants because they do not belong to the required target group.
This is a serious problem for every online panel. Just think about the signals you are constantly sending to a poorly profiled panellist. Email invitations repeatedly ask you for the favour to respond the survey in a timely manner. As a panellist, you somehow overcome your reluctance to take the time and share personal information. However, even though you show good will, the survey tells you that something is wrong with your profile and that you have to wait until the next time (without guarantee that you’ll be able to complete then). It is obvious that these frustrating moments have a negative effect on the response rates and the average duration members are willing to remain active in the panel.
Knowing to which target groups a panellist belongs can help to improve the membership experience. If you want to interview only owners of a certain car, you should avoid inviting panellists who, according to their profile data, own a different car.
This explains the vast amount of the profile variables we have. They simply reflect the wide range of topics we are regularly covering with our research. In total, we have eight different categories in our profiles, a basic profile that covers personal information, and seven special profiles: household and family, mobility, media usage, money and insurances, jobs and career, health and hobbies. Each of these categories contain a wide range of questions and provide us with about 500 data points for every panel member. It is worth mentioning, that every panel country has some additional local variables, like favourite TV channels or newspaper subscriptions.
However, it’s not only the amount of variables that make a profiled panel valuable, it’s also the quality of the profile information, especially whether all information is still valid and up to date. Some things change rather seldom (e.g. chronical diseases), some things change more often (e.g. preferred TV station) and some other things are changing continuously by definition (e.g. amount of business trips during the last year). This is why we prompt all panellists to reassess their profiles on a regular basis.
We will discuss at a later point in this blog series, how we ensure that all profile variables are valid, complete and updated. We will also have a look at how profile variables help us to assess the quality and representativeness of our panel. However, it should have become clear already, why profile variables are so important for the quality of a panel. They help to maintain the panellists motivated by tailoring our sampling to their profiles.