I believe second hand research on influencers would be scarce and far too removed to unveil material insights. I would begin my primary research by: 1. Defining a scope of influencers/influencing space. 2. Generating a list of influencers that qualify under the category, or problem statement. 3. Go directly on the relevant influencer channels (i.e. YouTube, Instagram, Tiktok) to compare high performing individuals or groups and look for existing or emerging trends in those space. 4. In terms of data collection for comparison, I would look at performance metrics in terms of engagement to identify effective or ineffective influencers. I would also take note of the qualitative factors that might be common characteristics of good-performing influencers.