Since the 1:1 interviews started with the broader question, we began collecting insights, and these started to populate a list that was clustered into several topics. For every 3 guests that mentioned the same insight, we created a shortlist which at a certain point had 3 main insights. As the conversations with guests proceeded, so did the literature review, and soon the insights were 5, and then 8. In the last round of conversations, after the interviewees had some time to respond to the open question, we would present the list of the insights that were clustered and organized, and asked them to comment and rank these according to their understanding of their impact in the thesis question. At the end of the 25th conversation, and out of a list containing more than 16 clusters, we settled on the 10 insights presented in detail in the research document. The order of this list is based on the sequence they appeared in the discourse, as they became stronger and relevant, enough for us to add them to the list. A couple of the interviewees thought these insights were somehow biased against design.