How are mobile or web data analysis carried out?
Studies try to segment users by groups, and from there, predict behavior so that they do something. Human beings are predictable and permeable, so it is a question of seeing the probability that the users of such a group will do such an action, and in what way it can be achieved. If you include me in a population group: male, male, of such age range or status, with children or without children… My behaviors begin to be very predictable. They know that I am going to consume certain products, that I have certain interests, that I make trips very similar to those of my segment.
But when they analyze the users, for example, of a mobile application, they do not have much of that personal data …
There it is played with the crossing of many data. There are many techniques. The simplest are the forms that users fill out, but there are also usage patterns that allow making inferences, hypotheses and assumptions. If they detect that on my phone I use certain applications at certain times, they will put me in groups. There are studies that even infer my gender, with a relatively high degree of success. At the individual level there will be cases that fail, but what matters to you is the bulk. There are mobiles that different people use, but they are used by different applications, at different times … You have to adjust the data.
Does your company work with data offered by the operators?
We do not. There are companies that do it, especially in terms of travel or positioning, but our clients are companies that operate on the Internet or mobile operations. The data is usually from customers, users who use their web pages or applications, and we help them get more data. It seeks to improve something such as the return or sales ratio, or open a market in a certain geographical area …
Let’s say I use an application to listen to music, like Spotify. Do I give you a lot of information about myself as I choose songs?
When I register with Spotify, I already give you my email, date of birth, country, and payment information, and certain things can be inferred from the name of the email. But the consumption of music is what gives the most information: the hours of consumption, the types of songs … Music is very significant, because it is closely linked to our emotions and state of mind and vital moment. The music we listen to, the moment we listen to it or the frequency gives us a lot of information. Then we combine it with analysis by teams of psychologists and sociologists, who tell us: people who listen to this type of music at this time have this profile. If I connect during work hours, I listen to certain music for work; the person profiles that do this are X, they will like this other music. If I connect at night, I give it a certain affinity. They classify me.