This was my first approach to this environment. I was studying in order to capture a layer of humorous messages that we simply could not access from other open platforms. But when I entered WhatsApp, it was 2018 in Brazil. I previously started to research the Brazilian 2018 elections that elected Bolsonaro here. I observed that these far right groups already had a very structured ecosystem of different groups. Each group performed different functions. So you have a group that organized public choreographies supporting Bolsonaro, you have groups that organized motorcades, you have groups that sell T shirts, you have groups that discuss politics. I tried to engage within these groups using a covert research strategy, because once I identified myself as a researcher, I would be immediately banned from these groups. Also, once I identified myself as a researcher, I was immediately subjected to security risks. So I used covert research, which is a strategy of not deliberately not identifying myself as a researcher to observe these environments. I tried to understand how content circulated, which were the main users, the main senders of these contents, and a bit from an ethnographic point of view, trying to understand these activities. But I also noticed that most of the content circulated, most of the users that sent the majority of the messages were not as organic as they said. They clearly assumed some behavior that one could identify as inauthentic behavior, because some of these users, for instance, sent the very same message for the very same groups during the morning, during the afternoon, during the evening, and also during the night, in order to capture the attention of different users. They spread these messages systematically in a way that no spontaneous user, no organic user could do. They were clearly a professional agent intervening in this environment. So I started to look at this to understand these patterns. And one concept that was very useful for me was the astroturfing concept. The astroturfing concept already existed before, of course, but most of the time it was discussed in only a theoretical way. You do not have much empirical literature to discuss these nuances of a practical environment of a political agent performing a spontaneous supporter of a politician. I tried to understand these patterns in an empirical way, using the concept of astroturfing and dealing with these concepts in a multi-method approach — using most of the time an ethnographic approach together with a social network analysis, and also some descriptive statistics, to understand the circulation of these contents in these environments, and how we could identify these astroturfing patterns within these WhatsApp groups I was observing.