There is a long thread of opinion being thrown out in Quora to answer a user question - "What Determines influence?" The neat thing about Quora right now is that a lot of digitally influential folks (by my intuitive measure) are answering questions. I am pretty sure this is experimental behavior and it is anyone's guess whether more than the professional pundits (paid to write, essentially) will stick with it. But I digress.
There are lots of thoughts in there from smart people. Many are interesting. Most don't really help today's marketer beyond giving him/her something to wonder.
Why we should care about influence
Everyone is trying to make use of commonly understood phenomena:
- Consumers in many markets around the world trust in friends, family and people's opinions online for purchase decision and opinion on issues(not equally).
- With the explosive growth of Facebook more and more people get information via their news feed - i.e. from the weak and strong ties of the people they have friended
- 51% of searchers are looking for other people's recommendations. (from Global Web Index 2009) That means Google connects us to both brand-sourced content and third party opinion and that people actively seek out the later
- Network and influencer "influence" seems to be challenging the accepted effects of reach and frequency-based advertising
With all this going on, its no wonder marketers want to tap into personal influence.
A model for influence & relevance
Many of the answers in Quora catch many great points:
- Influence is contextual - the nature of the publisher, the topic and the relationship of the reader are not static states. I listen to some people on car advice, others on music. They are rarely the same people or sources.
- The many "scores" out there including Klout are insufficient yet still some do serve a purpose. The discussion gets academic pretty fast and I am not talking Harvard-academic
- In the online space, we are all just learning how to actually apply our understanding of influence to actual marketing and communications programs.
We have a model for gauging the influence of individuals. It varies depending on the person's main platforms (e.g. I blog, tweet, publish via other blogs, and now have published in Quora, etc....). But here's the model for a blog-centric individual:
We have gone through the painstaking process of defining simple but useful formulas for each of these KPIs that we feel really define how influential a source can be. But how relevant? To assess how important that influence might be for a vehicle recommendation vs. a hotel, we also map relevance. the key criteria for that are as follows:
Yet practically-speaking, this precision needs to be simplified for us to use on a daily basis as marketers. The scores for these simple criteria allow us to map individual sources on a simple XY axis. Now we have something to work with. But notice a few things:
We don't specifically call out qualitative assessments about a person's affiliation or credibility. That doesn't mean we don't value someone's affiliation with a top university, for example. That should show up in the data. Still there are always anomalies. Celebrity is one. The question in Quora related to an article in Ad Age on paid celebrity tweets. Many people in the Quora stream have commented on how influential they feel a tweet from Justin Bieber's stream might be. A celebrity tweet can raise awareness about a product or service. I have no doubt that core fans may even be driven to behavior simply because Justin's stream suggested a product. But in general, it is not realistic to think that any one source can be predicted as the "persuader" or cause of a purchase.
The complexity of influence and compound effects
Marketing has long been struggling with attribution. Which communication actually drove the person into the store? Traditional online marketers try and track the 'last click' but even that is deceiving. The influence of individuals and networks is no less complex.
Weak ties vs. strong ties - this is not a new concept. There is a difference between the impact people close to me may have on my opinions or decisions and those that span out far into my social graph.
Network effects can challenge the impact of influencers. From how clusters work to broad networks, researchers like Duncan Watts have demonstrated that a broad network can have greater influence on a behavior than any single, highly influential individual.
People form clusters that can have great influence on behavior. Online and offline we tend to flock with folks that we share some affinity with. Sometimes this is referred to as homophily. The importance of homophily is likely overblown in my humble opinion. (here is a neat opinion in support of its importance) Still marketers need to be mindful that it takes special effort to communicate and engage with groups across clusters to avoid being trapped in an influence-chamber that does not extend beyond that affinity cluster.
Owned, earned and paid media have a complex effect. We are all experiencing a tremendous number of messages and interactions each day related to products and services. many of those are advertisements, may are third party opinions. They also have a compound effect. For instance, some studies point out the increased effectiveness of paid media in proximity to earned media.
A practical approach to influence for marketers
Tons of sources influence a purchase decision. Social media, advertising, brand reputation, offline relationships and much more. If for a moment, we say what can we as marketers do via social media to affect influence towards our ends, we need to understand:
- What is likely to make a source more influential on a subject with a particular audience
- Understand how networks 'work' and have a reasonable understanding of how people make decisions
- Then build a program that both engages influencers and networks such that each generate authentic and useful word of mouth
It will never be as simple as identifying the top 10 "influencers" for a given topic and then convincing them to chatter positively about a brand.