Why Brands STILL don't Understand Digital Influence?

[#1]

The topic on influence and influencers is very complex and continues to baffle the industry. In an attempt to clarify the intricacy of this subject, I presented the foundation and defined what influence really means last time. I also showed how the simplistic definition is insufficient. Moreover, we’ve put this definition to use and explained why the follower count metric shouldn’t contribute to someone’s influence. However, there are still much misunderstanding about influence and how it works.

When I started writing about influence and influencer two plus years ago, it was primarily because of two reasons:

1. I was developing an influencer scoring algorithm on our community platform based on Social Network Analysis (SNA)
2. I see a lot of misconceptions about influencers and how influence works in real life as well as on social media

But two years later, despite thousands of articles and dozens of good whitepapers written on this topic, brands still don’t understand digital influence. The interesting question is why?

No One Has Any Data on Real Influence

One of the reasons that brands don’t understand digital influence is because they don’t seem to realize that nobody actually has “data” on influence (i.e. data that says precisely who influenced who, when, where, how, etc.). Just because someone has a number, it doesn’t mean that number is measured. In the case of influence score, it is computed via some algorithms. These influence scoring algorithms are called black box models in statistics. They take inputs, which are typically activity and interactivity data of a user plus some environmental data on various social channels (e.g. posting messages, connecting with others, retweeting, etc.). Then they crank out an output score (via their black box) that supposedly predicts that user’s influence.

But where do influence vendors get those inputs? They certainly don’t own the user interaction data on various social channels. Vendors typically get their input from the respective social media platforms: either through public APIs, special data feeds through partnerships, or they buy the data.

However, social media were never designed to measure influence either, so they are not able to track explicit data on who influenced who. Social media were designed to let users connect, communicate, and interact in novel ways, so they only have data about connection, communication, and interaction between users. Social media platforms don’t have influence data either. Therefore, all influence scores are computed based on some models. As such, these scores are merely a prediction of someone’s influence, not his actual influence. Those who don’t understand this difference between real measured influence vs. predicted potential to influence are probably wasting their valuable dollars.

Real Influence Depends on the Influencees

Another aspect of influence that brands don’t seem to understand is that influence is a concept involving two parties: the influencer and the influencee (i.e. the target audience being influenced). Moreover, the definition of influence inherently depends on the effects produced on the influencees, not the influencers. So influence only occurs when there is a change in the influencees’ thought or behavior. An influencer can only try to cause that change, but ultimately it is the influencees who determine when real influence actually took place.

This is also why it is so hard to measure true influence, because we will need to track when a desired change has taken place among the influencees, then tally up how many of these are caused by specific influencers. As you can see, to measure real influence, we need to measure the influencees, not the influencers. That is why no one actually has this data, because no one can measure it. The influence score that is computed (not measured) from someone’s social media activities is merely his potential to influence.

Alright, back to the influencees. These influencees are actually implicit in every definition of influence I’ve seen. For example, let’s take the definition from “The Rise of Digital Influence” by Brian Solis: “The ability to cause effect, change behavior, and drive measurable outcomes online.” Cause effect on whom? Change who’s behavior? And drive measureable outcome from whom? If I recommended the new iPhone 5, and you (the influencee) did just what I said, that is influence. If I went and bought the iPhone 5 myself, is that still influence? Obviously not!

Remember, it is the influencees’ reaction to the influencer’s action that constitutes real influence. Without these influencees, there is no influence at all! Regardless how influential an influencer might be, without the influencees, the influencer cannot produce any influence in isolation.

Conclusion

Real influence is a difficult concept for brands to grasp, and there are two reasons for that:

1. Brands don’t seem to realize that nobody has influence data (i.e. explicit data that says precisely who influenced who, when, where, under what context, on what subject matter, etc.). Therefore all influence scores are really influence estimation, which quantifies someone’s potential to influence
2. Real influence depends on the influencees. No actual influence took place until the desired change of thought or behavior is carried out by the influencees. Influencers only have the potential to influence. They can’t actually influence anything without the influencees.

OK, now we understand why brands misunderstand influence, next time we will address the missing link in the influence industry. Stay tuned for the exciting revelation... In the meantime, I’m happy to discuss further any misunderstanding you might have about influence.

---

Michael Wu, Ph.D. is Lithium's Chief Scientist. His research includes: deriving insights from big data, understanding the behavioral economics of gamification, engaging + finding true social media influencers, developing predictive + actionable social analytics algorithms, social CRM, and using cyber anthropology + social network analysis to unravel the collective dynamics of communities + social networks.

Michael was voted a 2010 Influential Leader by CRM Magazine for his work on predictive social analytics + its application to Social CRM. He's a blogger on Lithosphere, and you can follow him @mich8elwu or Google+.

[#2]

Michael, I couldn't agree more. This is an important piece and your work on this front continues to push marketers and strategists to rethink what they think they know and what they should know about digital influence. 

Just to clarify, my definition referenced here is incomplete without the context of the report. It actually aligns the points you make here. The relationship between cause and effect is at the heart of what you're asking..."Cause effect on whom? Change who’s behavior? And drive measureable outcome from whom?"

One of the problems with influence is that brands aren't thinking through these questions, but it is where everything begins. What is the effect you wish to cause or the behavior you wish to change. Who are the individuals that can help you acheive these goals and are the people they reach the likely candidates for your efforts? 

That way, brands can start to think through how to trigger desired scenarios as you described here, "If I recommended the new iPhone 5, and you (the influencee) did just what I said, that is influence." 

What are your thoughts on brands confusing influence and advocacy and influence with WOM?

Cheers Michael!

[#3]

Great post Michael!

Influence data does exist though (here the BlogFrog we ran many influencer programs for brands over the last two years); it’s just not the type of lower purchase -funnel data you’d expect to find on a PAID campaign and can’t usually link back an individual influencee to the individual influencer that influenced them. Since influencer marketing, just like social media by the way, is a much better fit for upper funnel conversion, that’s awareness and consideration, influence data is usually linked to brand lift, intent to recommend/ intent to purchase type of data.

[#4]

Hello Brian,

Thank you for taking the time to comment on my blog.

I actually read your report and I thought it was probably one of the best I've seen. I understand your definition, and I wasn't challenging it or saying that it's incomplete. Just that there are implicit influencees in your definition (as well as others) that the readers often overlooked. For people who bother to define what influence is, they all seem to understand it, but somehow it was not made explicit. The focus is often placed on the influencers rather than the influencees. I just want to make this point very explicit so the readers won't have a chance of missing the important roles of the influencees. Regardless, I'm glad that we are align fundamentally.

But I agree that brands are not thinking through the fundamental questions deep enough to reap the real benefit of their influencers.

Concerning influence, advocacy, and influence with WOM, the difference between these are quite subtle. With respect to a brand, advocates are just people who support the brand and defend the brand. They can be anyone. They may not be influencers yet, because they may not have a high enough capacity to influence others. However, they can certainly influence other through WOM. WOM is just one method to propagate influence. It's a method that everyone can use. Everyone has some influence. Influencers simply have significantly greater capacity to influence.

If brands cultivate their advocates, they can one day become an influencer. In short, an influencer is an advocate, but not all advocates are influencers. Although influencers are more powerful, there are much fewer of them than advocates. Sometimes, the collective voice of all the advocates may be greater than those from a few influencers. Since advocates are already prime to support a brand, they tend to be more positive toward the brand, and they are usually the influencees who carry out the “last mile” of the actions recommended by the influencers as well as other advocates. Because influencers need to maintain their credibility in order to sustain their influence, they can be critical at times. Regardless, they are often constructive even if they have genuine criticisms.

Advocates and influencers have different roles, depending on what brands want to achieve. Influencers have greater reach. Being so broad they are often better for driving the top of the funnel (i.e. awareness and interest). Since advocates have a smaller audience, they can reach people at a more personal level. Even though advocates may not reach as many people as influencers, they are more effective at driving the bottom of the funnel (i.e. conversion and actions).

Alright, this is just my quick brain dump on the distinction between influence and advocacy.

Thanks again for stopping by. I hope to see you again next time. 

[#5]

Hello Raviv,

Thank you for the discussion.

If you can't link the influencees back to the influencer who influence them, then you didn't actually measure influence data. You estimated it. That is not data. That is an estimation, a prediction, it's the result of statistical computation, not measurement. Like I said, just because you have a number it doesn't mean you measure it. Just because you have a number it doesn't mean that number is a datum.

It doesn't require the lower purchased funnel to be actually influence data. If you survey people and ask them who change your mind about a particular product / service, who made you interested, and start considering, and even desire a certain product / service. That is sufficient to measure real influence. They are not low level transaction / purchase data. So people can measure influence and everyone has some influence data. They are just not measured on the social web. For example, I know my wife change my mind last weekend on where to go for dinner. That is real influence data. She influenced me on a particular restaurant. People have some influence data, everyone has some in their brain, but not at the social web scale and not on social media.

I hope this clarifies the distinction between estimation vs. real data.

[#6]

I'm no expert, but the popularity of social media says it all...

Especialy w/ women who are generally the decision makers when it comes to purchases for household. Digital influence of womens opionions are  vividly being displayed-in full color! Women are greatly influenced by other women-of their own demographic especially. Using an intenet social site, you can "zero in" on someones profile. Simiar profile's to one's self-are offered up on "digital platter" consistently.

I can go online before I buy a product. In few click's find a lady who could be my twin, based on their of buying habits, websites visited, "likes" on social media profile. I can feel pretty secure based upon these "habits" they will share my likes/dislikes. It's easier to go online for a digital buddies opinion if all I want is "the goods" on product information. A real life conversation can end up taking more time and touching on other topics perhaps not even interested in. Often it's just simpler proposition to get digital influence vs person to person. Not to mention time saver and aninimity advantage too(depending on what product your researching).

[#7]

I'm influenced by all kinds of things.  An example is the Livescribe pen.  I saw our own Michael Puhala using it and asked him about it.  He was so excited about his pen that he gave me an immediate live demo and talked about all the ways it helped him be more productive.  

After he left my desk, here's what happened:

• I did a little research online
• I thought about it for a few days
• On a trip home to Chicago, I mentioned it to my family (my step-dad said, definitively, "GET IT!")
• On a whim, we ran into a Best Buy on the way to the airport and I bought it with pads of the special paper you need
• I ripped it open, tried it, liked it
• While I was thinking about it, I went onto Amazon to buy more supplies

Here's the weird thing - another huge "swing and miss" for brands.  Now, online, I saw display ad after display ad about the Livescribe pen.  Now that I'd bought the thing, these ads were annoying, particularity when they advertised that the pen was now on sale.  I could click on the ad to turn it off (there's some setting feature) but I could not click anywhere to report that I'd actually bought the pen.  Forgive me for drinking the Lithium kool-aid, but even more powerful would be an option for me to say that I bought the pen AND invited me to learn how others LIKE ME had been productive and creative using it.  Influence was clearly woven in at a few touch points but the brand STILL missed an opportunity to engage with me in a meaningful way.  

[#8]

One sign of at least one kind of valuable influence is to be able to attract an intelligent conversation as you have managed to do here Michael

[#9]

Hello Randi and Erin,

Thank you for the sharing your stories and view points.

There is no doubt that influence at work everywhere, on social media and in real life. And I totally agree with both of you that people with similar profile to us, or have some relationship with us (e.g. colleagues, relatives, etc.) are our strongest influencers. That has also been shown in many research studies.

However, that is not what this post is talking about. I’m not saying that influence doesn’t exist, because it certainly does. I am saying that no platform or vendor actually tracks this data. Who has the data that says Erin has been influenced by Michael Puhala and her step-dad? I don’t think any influence vendor has this data, nor does Best Buy has this data either?

And likewise for Randi, who do you think has data on who actually influenced your purchased? Even if someone has that data, they won’t have that data for everyone, and for every instance where influence occurred. That is why influence vendors are not actually measuring anything, they only computed your influence score, which is only an estimate of your potential influence, because they don’t have data on whether influence actually happened.

I totally agree with Erin that brands are still not very good at knowing when to serve what ads. They lack the data and analytics that let them know when to serve which ad. This is where company like Google does really well. But look at how much data they analyze in order to do this. With social media, the hope is that through meaningful and genuine engagement that people will create enough social data for sophisticated algorithm to learn and predict the right ads to serve for you, (i.e. personalized). Wouldn’t that be cool?

Alright, thanks again for commenting, and I hope to see both of you next time.

[#10]

Hello Kare,

Thank you for the nice comment. Comments of any kind are welcome here. Whether it’s a deeper discussions or stories, challenges or criticism, or simply suggestions... I like them all.

So thx and hope all is well with you, and see you again next time.