Influencer marketing: The appliance of scienceby
As principal scientist of analytics at Lithium Technologies, Michael Wu is one of the leading thinkers on behaviour within online communities and social networks, and has spent years exploring the dynamics of social interaction. MyCustomer.com caught up with Michael at the recent European Customer Experience World to discuss the value of influence, and how to identify - and engage with - influencers.
MyC: How do you define influence?
MW: Influence, at the really high level, is the ability to change somebody’s thought or action. It doesn’t always have to action. A lot of people out there say that influence has to create action but not necessarily. Every time you move people down the purchase funnel, when you change someone’s mind from being, say, not aware of your product to the start of being aware, to being interesting and desirable, then you have changed their mind. So that is some form of influence as well. And then eventually you influence them so much that they took action to either purchase or refer a friend – then that is also influence. So it is changing people’s mind or action.
The thing to point out is that it matters how you do it too. A lot of people say it doesn’t matter how you do it as long as you change people’s mind – that is not true. That is not really influence because you could pay people to do it, and that is not really influence, you just bribe them. You could force people to do it and you could put a gun to their head – and that’s not really influence. So I have four points and I call them no carrot, no stick, no annoyance, no trick.
No carrot meaning no money. It doesn’t mean no reward but just no money. No stick means no force. You can’t force people to do things, that is not really influence. No annoyance means you can’t frustrate them, you can’t spam them or anything, they have to really want to do it themselves. And sometimes people want to do it themselves but they just didn’t get the whole picture, they get tricked or they get deceived into doing things, so the last one is no trick.
So basically influence is the ability to cause a change in some people’s mind or action under the condition or no carrot, no stick, no annoyance, no tricks.
MyC: The saying goes “Target the influencers, and you can move the crowd”. So how can firms identify influencers?
MW: Finding influencers is one of the things that a lot of people also misunderstand out there. Many people and vendors claim that they can find influencers but they are actually finding the noise makers. I believe that finding influencers requires six factors. The model that I developed basically models a two-party dynamic. In order for an influencer to influence a target – a target audience like an influencee – his influence has to be propagated to him. Otherwise there’s no chance. So for the influencer’s influence to propagate to the influencee there are six factors that have to be met.
The first factor is credibility – domain credibility – because there is no such thing as a universal influencer. People are influential in certain domains. You could be a car expert and you have some influence there. You could be a camera expert and you have some influence there. So domain credibility is the first factor.
The second factor is bandwidth – you have to have high bandwidth. Clearly if you have credibility but never say anything you’re not going to be influencing anyone. The bandwidth essentially is what most of the people out there focus on. How loud you are, essentially. So there’s two types of data to determine how much bandwidth you have. One is what we call the participation velocity, how quickly you participate, for example how many times you tweet in a day, the number of blogs you have written per week... it is kind of participation per unit time. And the other kind of data that would be used is what we call social equity data. Those are things that you accumulate over time, for example the total number of followers you have, the total number of friends you have, the total number of blogs you’ve written throughout the lifetime of your blog. And so those all contribute to your bandwidth.
And the third factor is relevance – you have to have relevance. That basically says you have to take into account what the target wants. If I’m a camera expert and you’re trying to buy a laptop, I’m not going to be able to influence you because my expertise is not relevant to you.
Relevance actually has two different components. One, is the content relevant, so this is the content relevant, the third factor, and so the fourth factor is temporal relevance – this is relevance in a time domain. So every consumer has a decision window where they are most susceptible to influence. For different products it’s different. For example, if I’m trying to buy a laptop I may do some research for a couple of months before I make a purchase so that decision window is about two months. If I’m trying to buy a car maybe I’ll research for six months. If I’m buying a house it’s probably a year. But if I’m trying to buy a beer maybe it’s the next five minutes. After that it’s too late – I’ve already bought the beer, I’ve already drunk the beer, it’s too late. So the relevant time is usually now, and the relevant window is usually how long you have to look back in order to determine the influence. Because some influencers maybe are really influential on the topic of your interest but it was too long ago and they may not talk about that topic anymore and that becomes irrelevant to you.
The fifth factor is channel alignment. Channel alignment simply is the alignment of where the influencer is and where the target is. They better be in the same place, basically. So if you are trying to influence people on Twitter it doesn’t do you any good if you find a Facebook influencer. And the same thing if you have a campaign in LA, it doesn’t help you if you find an influencer in New York. So it could be in a physical location or it could be in the digital world in a different online channel. We call this factor online channel alignment.
And then the last factor is essentially what we call trust. So regardless of whether you have any of the first five factors, if the target doesn’t trust me as an influencer I can’t influence them.
And one of the interesting things about this model is that all of these six factors have to be present and accounted for. If you miss any one of the factors essentially you break that chain and so the influencer’s influence cannot propagate to the target and you don’t have any influence.
MyC: Once you have identified an influencer, what is the best way to engage with them and build a relationship?
MW: That is actually the hard part. The easy part is to find them. Once you find them it is actually very difficult to build a working relationship with them, long-term, because you can’t really treat them like employees and work them like an employee because they will hate you and they will turn against you. And you probably shouldn’t pay them either because if you pay them they lose credibility. When most people realise they get paid to propagate a certain message then they are not going to trust the influencer anymore. So they lose credibility that way.
And to influence and influencer again you can’t use the same four things, you can’t use carrot, you can’t use stick, you can’t frustrate them and you can’t trick them either. So working with them is really hard. The only strategy that I know that is repeatable long-term is to co-create value with them. So in some way everybody has to get some benefit out of this of this relationship and then this will be a long-term interaction.
One example I can think of is using a community. So the community provides audience for the influencer, so the influencer actually gains influence by providing help and expertise in the community, and the community members get benefit from his expertise and clearly the sponsoring brand for the community gets benefit of effective word of mouth through the influencer. So this is truly a win-win-win strategy for everyone. So that is one way you could do it.
MyC: How can businesses measure the ROI of their efforts at engaging influencers?
MW: ROI is a tough nut to crack. I have done a little bit of study on that and basically we did a control customer lifetime simulation using real network data from our platform. And we found that if you target random users you will get some benefit. But if you target the influencer, the benefit is greater, and it is actually greater by 50%, so it is pretty significant.
Neil Davey is the managing editor of MyCustomer. An experienced business journalist and editor, Neil has worked on a variety of newspapers, magazines and websites over the past 20 years, including Internet Works, CXO magazine and Business Management. He joined MyCustomer in 2007.