Does Influence Matter in Customer Service?

No.  My opinion of that was solidified last week thanks to a band of wikid smat (translated: “really smart” for you non-Bostonians) people led by Wim Rampen.  Wim tossed some tweet chum out there last week with this: 

This whole search for influencers and influence doesn’t sit right with me…what are your thoughts?

I took the bait.  And, this started a fascinating twitter chat about this complex, multi-dimensional issue.  My response was: 

I’m in customer service. I don’t care about influence. Customer has a problem, gotta fix it.

Do we ask callers who they know before we provide service?  Do we respond to emails with “I’d be happy to help you if you could first tell me how many twitter followers you have and how many people subscribe to your RSS feed”.  Of course not.  Being a bottom line kind of person, this is where I jumped to.  But, as others like Prem Kumar, Brian Vellmure and Mitch Lieberman pointed out, it’s complicated.

While I think we are all in agreement, each of these folks came at the topic from slightly different angles.  And there are many.  Case in point:  based on a recent article about Delta Airline’s use of Twitter for customer service, Mitch saw it this way.  Wim pondered the validity of influencers in general in a post that stimulated a great discussion.  Eric Jacques also weighed in.

So you can clearly see, it’s an issue with many moving parts.

My point is this.  Customers need service.  That part is simple.  So, I stand by my assertion that I don’t care about influence.  That is not to say, however, that every customer gets the same white glove level of service treatment every time, all the time.  Economic realities and scarcity of resources dictate that, in order to deliver a superior service experience, many organizations have a need to segment their customers for treatment.  But, what possible value is created by making those segmentation decisions based on influence?  According to Wim, he’s already witnessing this practice.  It’s even more insane in my mind to use this criteria over or in place of other measures like CLV, profitability or loyalty.

So, as I commented to Wim over on CustomerThink, picture this scenario:

Average Joe Consultant, who’s a Delta million miler and flies them exclusively, gets bumped from a flight, is disconnected from hold after 15 minutes or is charged cancellation fees or made to swipe his credit card to use the head because special treatment is being given instead to someone who tweets his request instead of calls, has 5 million followers but has never flown Delta before.

What if that was you, Joe?


  1. Another excellent post, Barry! Thanks for sharing.

    Customers and organizations can both benefit from customer segmentation and corresponding levels of customer service. Nothing wrong with that, as long as customers know what service to expect, and as long as companies live up to these expectations.

    Adding influence to the mix may obscure things for companies and customers. As Wim points out, influence is still elusive, and as such it would be just an extra and above all fuzzy variable in a multi-level equation for the most optimal customer segmentation.

  2. Great post Barry, good to read after a couple of days away…

    I agree to your point that Customer = Need Service and as such influence is not a critical factor.
    However I think the definition of 'Influencer' is changing rapidly these days as social media are greatly expanding the influential reach of our Customers.
    While I fully agree that the influential level of a Customer should not determine the level of Service he gets (although it would be fun on an automated phone menu:'Press 1 for more than 1000 followers…') we need to consider the influential reach of our Customer base as a whole as an important parameter. Employees in Customer Service should be aware of the power of influence and understand that in this rapid evolving world of social media it's even more important than before to get it right the first time.

    Take care


  3. thanks for the insights Christophe and Bart. I think the one issue common in both your thoughts is the, as Christophe calls it, fuzzy nature of influence in a multi-variant equation that determines segmented treatment treatment or 'service level'. How do you quantify it, if it exists. What value do you put on it.

    To your point, Bart, the definition is changing. No doubt. And you could certainly argue that, in the future (way in the future in my mind), it might be a variable in calculating lifetime value. In the sense that, if my network reach is X and I've demonstrated that I can "influence" Y percentage of my network to some call to action, and each new customer has a CLV of Z. Then maybe the equation looks something like:

    A(XY)*(ZA) ; where A is the percentage of my network that I can influence that is statistically likely to make a purchase of my product or a similar product from a competitor.

    If we can quantify that, then I think influence can potentially serve a place in service segmentation. But even then, its probablistic vs actual dollar spend of a frequent flier for example so shouldn't trump those measures.

    Thanks again.

  4. mistake. Here's the edit


  5. Barry, I admire your brave attempt to capture such a complex matter in such a simple equation! I love simplicity, but I am not sure if we can get away with such an simple equation (considering the complexity of the algorithms used by Klout and other services).

    At this moment, I think it is still difficult to measure influence. As a result, it will be even harder to measure true conversions as a result of influence. And… it will be even one bit harder to _estimate_ the impact of a customer's influence on other customers' actions, and to provide a particular level of service accordingly.



  6. Barry,

    In my opinion the concept of the formula makes sense indeed, however as Christophe mentions I wonder if we can measure such a complex matter with such a simple formula.
    Would your formula allow for service segmentation or would the distribution be suboptimal (i.e. too narrow or too wide)? I also wonder whether the linear relationships in the formula would make sense at both ends of the influential spectrum.

    It's clear that influence in service is a hot topic these days and I think it definitely needs more thought. Your formula described here could be the basis for attempting to quantify influence. I think by gradually adding parameters to this base model we could come a long way.



  7. Customer expectations about service are usually pretty realistic. As long as if you have a good customer service, you’re going to have happier and more loyal customers, which will help produce happier and more loyal employees.

Speak Your Mind