Mark Schaefer commented on my recent episode about influencers:
“You are rolling old tapes here Chris. You’re viewing the world from a very narrow (B2B tech) perspective. Here’s the fact, influencers sell stuff. A lot of stuff. For many companies this is the most cost-effective sales and marketing efforts they have ever seen. The case studies are powerful and compelling. The techniques are becoming increasingly sophisticated, as is measurement. Influencers can move markets. Are you accurately seeing what is happening out there?”
Can’t see anything? Watch it on YouTube here.
Listen to the audio here:
- Got a question for You Ask, I’ll Answer? Submit it here!
- Subscribe to my weekly newsletter for more useful marketing tips.
- Find older episodes of You Ask, I Answer on my YouTube channel.
Some of the influencers who sell stuff – and a lot of it – are folks like Gary Vaynerchuk. He’s got his own successful sneaker line, as an example. Gary’s a great example of the sophistication Mark mentions, creating content on literally every channel possible by syndicating content across channels and branching out with new programs all the time. For that top tier method of content, Gary currently spends around $10,000 per day on advertising and his content team is 17 people.
I’d disagree that measurement – at least on the brand side – is improving. As recently as two months ago when I was working at an agency, measurement of influencers was still disconnected from business results. Even in some of the influencer programs I’m part of now, measurement is neither standardized nor connected to business results, though there are some glimmers of hope. One company recently asked for access to my Facebook Pixel and for a collaborative tagging effort with Google Tag Manager, which was incredibly smart. That’s definitely the right way to go.
Mark’s comments got me thinking about the fourth influencer archetype. When we examine the ways to measure influencers, our three archetypes – lots of connections coming in as the expert, bilateral connections as the connector, and lots of connections going out as the broadcaster – these focus on the network as a whole. For brands seeking to identify influencers, using these measures of centrality makes a lot of sense, because the brand’s goals are reach, engagement, and meaningful action. To Mark’s point, sell stuff.
- We understand authority using Eigenvector Centrality, the importance of an individual influencer based on who talks about them.
- We understand connectivity with Betweenness Centrality, how many times an influencer bridges connections to others in a network.
- We understand broadcast capability with Closeness Centrality, how few degrees separate an influencer from their audience.
Mark’s question – do these algorithms accurately see what’s happening (and by extension, people like me who use them) – is important. The blind spot in all these network algorithms that we cannot currently solve – but that our companies desperately want us all to solve – is which influencers create meaningful impact.
We can’t solve for that right now because we have no access – from the outside – as to who’s driving results. We could, as mentioned earlier, tackle this with tag management and analytics on an individual company basis. From the outside, though, before we sign a deal with an influencer, we want some way of inferring their ability to generate results.
Perhaps the answer lies in a combination of the different centrality algorithms. Each one judges influencer in a different way, but what would an influencer be capable of if they did all three? What if they were the respected leader of their tribe, that everyone talked about, but also the bridge to connect tribe members, and also so responsive that they were incredibly close to a large part of their network? That fourth influencer archetype, the true super-influencer, would logically be able to compel business results as well, perhaps better than influencers who only ranked well in one algorithm or another.
That’s worth experimenting with. I know I’ll be working on it, to determine a scalable way to compute that. Thanks, Mark!
You might also enjoy:
- Transforming People, Process, and Technology
- How to Think About High Bounce Rates in Google Analytics
- Marketers, Stop Panicking About Apple Mail Privacy Protection
- You Ask, I Answer: Best Language for Marketing Data Science, R or Python?
- What Is The Difference Between Analysis and Insight?
Want to read more like this from Christopher Penn? Get updates here:
Get your copy of AI For Marketers