Social networks need phone companies to create good advertising models

Social networks are a minority of communication, and great behavioral advertising strategies will need to think beyond them, out into the world of phones and face-to-face interactions.

Social networks plan to target advertising at people by first understanding where groups of friends exist, and then figuring out who is most influential within that group. By advertising to the most influential member of the group, they hope to reach the rest of the group.

Social networks plan to do this by measuring how often people communicate with each other, and how many times they “like” and “comment” on each others content. The problem however, is that social networks have a very limited view of the world. They only see a tiny fraction of a group of friends’ interactions. They miss all the face-to-face communication, all the phone calls and text messages, and all the email conversations.

Research we’ve conducted on how frequently people use different communication tools showed that social networks make up a minority of communication. Even with people who use social networks every day or couple of days, they communicated with others four times more often face-to-face, four times more often via phone calls, and twice as often via text messages. In other words, the vast majority of the influencing is happening outside the world of social networks. And most of it is happening off the web. Even the best advertising models that only use social network data will have a very skewed view of the world.

Businesses need more than the content and connections that exist in social networks. This is about relationships and conversations, not connections. If we think about…

1. Who is talking to who?
2. What is their relationship?
3. What are they talking about?

…we see that Facebook (for example) knows [1] and [3], but doesn’t know a lot about [2]. Yet [2] is the most important piece. [2] determines levels of trust. It determines who is likely to influence who. Crawling the graph and content on social networks won’t tell us a lot about [2]. The data required to really understand [2] lies in offline interactions and on people’s phones.

The best behavioral advertising models will be the ones that can tap into and understand off-web interactions. Who is calling and texting who? Who is giving who advice about what? Who trusts who?

Tweet This Post