Propagation of influence on social speech

Social networks are a fertile new ground for massive patterns of influence. We have seen examples of successful influencer (social creator) campaigns leveraged by brands daring to play in the new world of earned media. Some brands and agencies see this new world of earned media as the rightful evolution past the now more and more challenging and limited field of paid media (although the vast majority of advertising dollars are still firmly in this bucket). An entire crop of individual micro brands, the social creators (influencers), has risen and some of them making a comfortable living doing it full time. Even through the days of just launching a YouTube channel of your own and becoming an overnight success are well behind us, it is still a growing and interesting field. Indeed, the promise is strong and inspiring: Resonating a message with consumers by injecting your brand naturally into their conversations sounds a lot better than demographically targeted and relentlessly retargeted attention bombardment. Would we not all welcome this evolutionary step of bringing marketing into the next generation?

But influence propagated across social networks is not all about having trusted friends helping us with purchase decisions for area we believe they know well. We have seen the darker impacts of this model during the 2016 presidential elections in the US and through the surprising Brexit decision in the UK. We are just now piecing together how it was done. This manipulation was all based on extremely effective use of the mediums available (and not yet understood) for mis-information, dis-information and propaganda at levels we have not seen so blatantly in action since the fall of the Soviet Union or Nazi Germany. It is incumbent upon us to do our best to apply our latest technology and science to get a full grasp of what is going on and — at the very least — have a level of transparency. Only then will we be able to earn the consumers trust to reap the benefits of the new medium’s capabilities. This is not a natural approach for advertisers and marketers, but it could very well decide our future existence.

Studying influence propagation across social speech (traveling across social networks) is an amazingly multidisciplinary field. It brings together techniques and sciences that have not often been brought into ensemble with each other. This mix of social sciences like psychology and neuroscience as well as established disciplines like epidemiology. We have long used analog wording to describe what happens on social networks, like ‘things going viral.’ However, we are just now starting to see propagation truly as measurable with the same tools we have learned to utilize to understand infectious disease outbreaks. Epidemiology is a field utilizing advanced statistics over imperfect and small data-sets. In contrast, social speech propagation deals with enormous data sets. So standard statistical methods have to be augmented by ensembles of machine learning (including the magic black box of deep learning).

On the side of social science, it has become clear that topics and posts and mechanistic understanding of channels interacting with both the followed and followers cannot fully describe the seen phenomenon. We need to have a much deeper understanding of the psychology at work. The personality makeup of the expressed channels and posts and its quantitative understanding is a critical missing piece in our understanding. So descriptions of personality as in the HEXACO model of psychometrics become additional tools in the work. Only by bringing all of these elements together can we start to explore the observed patterns.

Definitions of influence, even though they might seem an esoteric concept are part of everyday live. Google launched a vastly superior approach to search because it was based on a relevancy ranking of influence. The usage of PageRank, which is at the heart a measure of how influential a page is to other pages broke new ground in such a dramatic way, that we soon forgot every other internet search system. When a tool becomes a verb, it has clearly won. So the next time you use Google, you are seeing an influence ranking at work. The leap of using PageRank to help define influence across social speech is a small one. But it turns out to just be a starting point. Time series propagation is much more of a modeling effort than a simple statistical measurement.

At Orbitwerks, we are living and breathing the orchestration of these approaches to get a measure of this disruptive paradigm change in the way we communicate as a civilization.