8 questions to help explain influence

brianreid The Cancer Letter, huh? I'll have to check that out. Sounds like the guy who puts it out has influence over the small group of people who regularly read his newsletter. I'd deal with this guy's influence just like I would any otehr influencer I was researching. The question is how could his influence benefit one of your clients? Or could it? Because not all influence is relative.

JGoldsborough Yup: lots of influence over a very small number. But that small number includes folks at NIH (who control the money spigot for research) and FDA (who determine what drugs make it to maket).
The point about all influence *not* being relative is well-taken, and I guess that guides where I end up on question #4: ranking becomes really, really tough when you are considering multiple types of influence.

brianreid JGoldsborough I'm glad to see Brian jump into this discussion because he and I have had some excellent back-and-forths about online influence. He's one of the smartest guys I know. The trouble is that Cancer Letter is not an isolated publication. I'm sure Brian (or any one of us) could highlight several similar publications. However, I would say that as the media world becomes even more digitized that these kinds of publications will lose influence if its online presence doesn't keep up. The other thing to keep in mind is that the way we discuss influence online is far from ubiquitous... There are most certainly instances of online and offline influence. How they come together in the same analysis? They probably don't...at least not initially. What you're ultimately raising here is the necessity for multiple kinds of research within marketing. It'll never be enough to say we're listening to online conversations only. Or monitor traditional media. It's a blended approach. Has been and will always be.

commammo shonali ValerieSimon You raise solid points as usual. For me, the value of the human analysis and understanding of the relationship and brand cannot be overstated. In fact, IMO, it's the best gauge of influence by far. So what I'm wondering is if finding a way to weight that POV, per Chemann 's weighted index idea, might show the importance people play in the influencer forumla. What do you think?

pierreloic People ask why I blog. How about smart comments like the ones I'm getting on this post. Your Hamlet POV is a different way of thinking about it that I hadn't considered before. I like it.

When you create a scorecard, do you ever consider qualitative, knowledge of the relationship metrics? You obviously look at human analysis, as you said. But how do you report that part with other metrics? Do you?

If you listeneed to and engaged with a niche set of influencers long enough, do you think you could eventually rank them based on ability to get target consumers to do what the brand wants them to if you defined what that was and even if what the company wanted them to do was different at different times? Make sense?

Thanks for stopping by. Cheers!

pierreloic I'm pretty much in agreement with everything you say. Thinking out aloud, though:

"Influence is not popularity." I agree. However, is there an element of popularity in determining influence? I have a couple of examples in mind but wanted to throw this out there for discussion first.

Shonali Hmmm. Is there an element of popularity in influence? If we accept degrees of popularity, from little to great, would we see a correlation to influence? I'm not so sure - the influence differs according to many variables, and the influencer you reach who gives you a 3rd party endorsement among a small group may be ideal - the influencer may not be widely popular, but popular among the right people for your objectives. Certainly someone unknown, friendless, follower-less, might be influential only to her or his family, so perhaps there's an element of scale -- a threshold of diminishing returns? Zero popularity equals zero influence, but Justin Beiber's 100 Klout score won't help me reach more potential clients -- enormous popularity, zero influence among my intended market.
Hmmm.

Shonali I agree that you can't completely strip out popularity from influence: to commammo's point, if noone listens, one's influence is inexistent. BUT we see time and again with our clients that the correlation between context/relevance and influence (as an action-driven concept) is much stronger than the one between popularity/reach and influence. Agreed that neither reach nor relevance can be ignored but they are not on equal footing.

pierreloic commammo That's what I was thinking too. The reason I mentioned it was that a lot of people will say "X is just a popularity contest" and so on... and while that might be true, it's the relevance of that popular person/brand to the end-user (in a contest, a voter) as well as how strongly they relate to them (and sometimes they just LIKE them, esp. when it comes to people) that will determine who pulls ahead... assuming X tries to exert some influence over said "audience."

Thoughts? Am I making sense?

Shonali Agreed. Popularity (or reach) is one of the key ingredients alongside relevance, resonance (engagement) relationship (social graph), and our new addition (available soon): reputation.

pierreloic I'm looking forward to seeing it! And I have on my list to ping you to chat soon...

JGoldsborough I love this conversation. I too have figured out why you blog :)
Let me try to answer your questions:

1- How do we account for / report on qualitative metrics? First let me say that we don't try to replace humans going over the data we generate to analyze it, instead we try to supplement that analysis by doing the quant work that can't be handled by a person and expose as much data and insights as we can to make it easier for people to conduct an analysis on top of the data. The quant work itself takes into account qualitative data - for example we use all endorsment metrics across the web that are highly qualitative data (someone telling someone else they liked a piece they did - like me with you on this post) - but we don't try to infer too much meaning from it. Then we let our users react to and engage with the data itself: they have the ability to add more qual data to an influencer profile, tag, add notes, like/dislike a profile we found, re-rank our list based on their campaign goals, etc. In other words, add human intelligence once we've reduced the scope of study to something manageable. The work our clients do with this is really fascinating and helps us build future versions of our app when we find repeatable patterns in their work on the data. Sorry if I ramble on this but I get very excited about this aspect of what we do :)

2- We could already (though we don't do it yet) take into account in an influencer's score/rank their successful engagement with brands. We have built in our app what we call an influencer monitor that measures the success by influencer of an engagement campaign. The thing we haven't figured out yet and the reason why this isn't yet reflected in our score is that I'm not sure it's the right thing to do... When a brand engages with influencers, they may have a very specific idea of what they want out of that relationship but our best case studies come from brands that were willing to be surprised and let the relationship evolve in unanticipated ways (hire an influencer, bring a harsh critique of their product to a product development workshop, etc.). These instances yield very high value for brands and we don't want to penalize the influencers who spearhead them by boosting those who fit the mental model of a brand.

Thanks again for the post and the conversation.

PL

Shonali traackr courtv chuckhemann Ok, I love the shooting fish in a barrel analogy. Had to read it a couple of times to get it -- call me slow on a Friday. But I think that is a great comparison anyone can understand.

We have been doing influencer outreach for years with the client I had in mind when asking about ranking. I really think we could do it and have the "knowledge of relationship" based on past experienes and outreach to have it make sense.

Chemann The more I think about it, the more I agree. Do you think ID process has to contain metrics and qualitative (or what I would call knowledge of the realtionship)? How might you position the two together?

So for instance, perhaps you measure exposure metrics (visits, impressions), engagement metrics (comments, tone), influence metrics (rank, authority, Klout, Traackr, knowledge of the relationship, qualitative stories) and actions (shares, story submissions, clicks, purchases, video views) or several othe rmetrics I failed to mention. Then you rank them all based on importance and lay them over each tactic? Am I even close to your weighted index idea? Were you thinking of another approach?

Sorry for all the questions. Just really trying to wrap my head around how people think about this topic.

Chemann Hi Chuck, I'm sorry to say we can't argue (at least much) here :) While I maintain that there is no universal influencer, I do believe standardization of the criteria used to determine influence is not only possible, but extremely important. But any universally accepted process needs to be meaningful. As Shonali points out, many of the current means for evaluating influence really offer only a very small piece of the picture. Really eager to continue to hearing your thoughts... and cannot wait for the book to come out!

JGoldsborough You're there my friend. The trick is blending qual with quant. Too much quant and you (generally speaking) have something most companies will distrust. Have too much qual and companies will think there isn't enough rigor behind the analysis. What's the right percentage of each? Only you and your client know that in conjunction. I'd be copping out if I said 50/50, but if you put a gun to my head its something like 60 /40 quant to qual.

ValerieSimon Bummer! Kidding... Shonali is right of course about a lack of means to evaluate the results. The unfortunate reality is that these tools will never give you 100% clarity into the black box, which is why its so critical we (and by we I mean your employer and you) work toward creating a process that's accepted by internal and external (clients) decision makers. Thanks for the note on the book. Want to write it for me? KIDDING! :)

ValerieSimon I agree, Val. No such think as a universal influencer, no matter how much chuckhemann tries to convince me :). Apologies to Oprah, the Pope and Lady Gaga.

We can't be reminded enough about tying our work back to goals and objectives set for the brand. When we start braindumping on metrics and creative measurement explanations, it's very easy to forget the basics. So I appreciate that reminder. Cheers!