Interested in pharmacology? You must be a female with right-winged political views, in your late 20s, who likes vegetarian sausage rolls and believes the world is controlled by a secret elite organisation. This is according to a new feature on survey website YouGov, which claims to allow you to “get the profile of your target audience across multi-channel data sets with greater granularity and accuracy than ever before”. Indeed, the website does offer a lot of detail about each “profile”, including a range of demographics, self-assessed personality traits, preferences for food and hobbies, online behaviour and favoured brands.
The data are gathered from over 200,000 members of YouGov, presumably as an amalgamation of various surveys that these individuals have completed on a variety of subjects. The sample size is impressive, although self-selected (which already produces sample bias when dealing with such data as facets of the personality). The scientific basis of these results is, of course, somewhat lacking. We do not know how many people were questioned for each trait or preference in each profile and I doubt those building the website had pre-determined hypotheses of any kind before analysing the data.
At best, these “profiles” (designed to look like they are representing one individual) provide a basic overview of some traits of some customers. A customer fitting all of the information provided probably doesn’t exist. In healthcare, particularly, it is important not to generalise customers and patients. Knowing that your customers are “likely” to have opposing political views to your own, or make more money than you do, should have no effect on the healthcare you provide. The idea of understanding your client base can be enticing, especially from a sales or marketing perspective, but with this can come prejudice. Even when customers seem predictable – when a website claims to know what they are thinking – the old adage that everyone is unique stands true, and everyone deserves to be treated as an individual. The sweeping generalisations made by the YouGov profiling tool are entertaining but potentially misleading.