And building programs to best leverage your most influential customers
Most companies internally assign a value to their customers that predicts how much money the customer will spend in the future, otherwise known as 'Total Customer Value' (TCV). Conventionally, the best way to determine a customer’s TCV is to look at how much they have spent in the past: United Airlines gives its best customers frequent upgrades along with a special express line, free drinks, and no charges (and higher priority) for checked bags. This is because these customers have spent a lot of money with United in the past which is a very good predictor that they will spend a lot of money with them in the future.
While calculating TCV from historical spending is simple, doing so does not value the customer based on the number of people they’ve told about your product and the resulting money these people spend. Yet people who convince others to buy your product can be more valuable than big-spending customers.
If Walt Mossberg — the technology columnist for the Wall Street Journal — buys your gadget, he's likely to be your most important customer for years even if he never buys from you again. Mossberg's true TCV should be off the charts! Mossberg is an example of an influencer.
Influencers are a small group of customers that have the potential to act as evangelists of your product. In cases where these influencers have bad experiences, they might actively tell people to avoid you (or actively protest your company).
Companies that understand the power of influencers are now proactively targeting them. FastCompany reported that marketers spend over $1 billion annually on campaigns targeting influencers – a figure growing 36% a year. But few companies have any idea who their influencers are and what to do once they have found them.
Determining who is an influencer isn't easy. Anywhere from 1% to 10% of your consumers are influencers. Below are some of the metrics Rapleaf uses to find influencers for our customers:
Last year, we conducted a study on 31 million people analyzing friendships on social networks. Results of the study suggest that almost 20% of users had over 100 online friends, while a tiny fraction of users (<1%) are “super connectors” with over 1,000 friends. Imagine the reverberating effects of mobilizing these people. These are the customers you’d want to target with exclusive offers, benefits, etc.
2.) Social persuasion. Let's assume John says he loves an obscure band on his MySpace page that none of his friends talk about. Then, two months later, 30 of his friends also list that band as one of their favorites. While it’s possible that John is an early adopter or that he just happened to listen to the band before they made it big, it’s much more likely that he is an influencer and told his friends.
Unfortunately, measuring social persuasion is not easy and can be quite expensive from a data-collection perspective. One way to gauge persuasion is to take snapshots of a person’s information at periodic intervals. By juxtaposing changes in information (such as music interests) alongside a social graph (who he is friends with), one can infer whether the person under consideration is influencer or more likely to be influenced. If a bunch of your friends buy the same product as you in the next few weeks after your purchase (which might be a six sigma event), there is a strong likelihood that you told your friends about it.
3.) Influence context measurement. Just because someone is an influencer in books doesn’t mean they are also an influencer in electronics. Matt Hurst of Data Mining blog looked at the categories of influence in the blogosphere and found that influence is a function of not only reach but also subject matter. Looking at the linking of blogs, Hurst found that Om Malik, for instance, has reach in the tech space, whereas Michelle Malkin has reach in the socio-political arena. On the other hand, Jeff Jarvis is influential on both tech and socio-political issues. Since influence depends on both reach and context and all three of these bloggers differ on both fronts, they are influential in different ways.
Some people will always be influencers – it is part of who they are – while others are influencers due to their current position (careerwise, within an organization, etc.). And that position might not be obvious. One wouldn't think of a personal assistant as being an influencer but Reggie Love talks to the most powerful person in the world twenty times a day -– he is President Obama’s personal assistant.
Influencers = $$$
After identifying influencers, how can you encourage them to promote your product? Here are some suggestions:
You can put your influencers at the top of the customer service queue. Or have an executive personally call and thank them for being a customer. Retailers might consider sending hand-written "thank y
ou" notes or providing priority shipping free of charge. In turn, influencers will tell their friends about you.
2.) Show appreciation. One way of making customers feel appreciated is by asking for feedback. Before finalizing a new product, get input from your influencers. Bare Escentuals does this for new cosmetic products. If you’re not soliciting feedback, at least give them sneak previews to new products. In return for showing appreciation, you can leverage focus panels and get critical input at no cost.
3.) Offer coupons and special discounts. Providing coupons and special discounts is a proactive form of customer service and marketing. Retailers can offer their top influencers unique promotions to keep them content and engaged.
4.) Employ special outreach. Rather than just sending influencers the standard newsletter, some political campaigns have interns reach out to influencers directly. Cultivating influencers give campaigns significant leverage to spread positive news about their candidate (and unflattering profiles of their opponents). Companies should follow suit to nourish their own evangelist network.
Now all we have to do is get companies to wake up and smell the coffee. Anyone willing to influence them?
(special thanks to Vivek Sodera and Michael Hsu for their help putting this together)