They haunt us. The one(s) that got away.
Shuffling around the street with other companies’ apps and games on their phones, other service brands handling their banking, insurance, travel, and cable.
They’ve become zombie customers.
If you’re an enterprise marketing executive with these eerily silent customers who haunt you with fear that they’ve churned, it might be time to rethink your strategy.
After all, you might quickly end up with a full-on zombie outbreak.
As we continue to barrel toward a subscription-based, managed services economy, predicting and preventing customer churn is an often-overlooked piece of acquisition-heavy marketing organizations.
Whether you’re in a recurring billing-based subscription model (telecommunications, cable, media, software), traditional service vertical like banking, insurance or travel, or rely on in-app purchases or advertising engagement such as gaming or social platforms, retention rate is one of your most important KPIs. And yet, this is probably the one that paralyzes you the most because it seems too scary complicated to address.
In the Harvard Business Review’s 2016 article, “Winning Back Lost Customers,” Georgia State University researcher V. Kumar studied 53,000 customers who left a telecom company over a seven-year period.
He cites three reasons companies should focus energy on lapsed customers:
- The customer has a demonstrated need for the service.
- The customer has familiarity with the company equates to less resources spent on brand awareness or basic education
- The company should have significant product usage data to draw upon to identify most profitable defectors and craft relevant win-back offers
This last point is particularly crucial.
While drawing upon existing product usage data is important in crafting win-back strategies, it’s even better if you’re able to predict churn risk behaviors in advance.
So why aren’t more companies predicting and capitalizing on customers at risk for churn?
Simple. It seems too hard with today’s tools. Manual process just doesn’t cut it when you’re dealing with a potential zombie outbreak and trying to zero in on individual customer’s reasons for leaving. It’s really hard to keep up with the empowered customer using traditional data modeling or even rules-based marketing systems that treat all customers the same.
It takes machine learning-fueled technologies to test and identify the most effective combinations of the thousands of customer behavioral data points and messaging permutations to address customers as individuals as opposed to rule sets.
In our experience fighting churn zombies for Fortune 500 companies, here are three things to keep in mind for your 2017 marketing strategies.
Identify zombie risks—while you still can
You definitely don’t want to be sitting in your easy chair totally oblivious that the zombie apocalypse has started. That is, until you hear the chorus of “braaaaains” outside your front door.
Most companies don’t have the longitudinal view on their primary customers at risk for churn. Their models can predict churn, but only in the days leading up to a cancellation or abandonment. Traditional churn models are like smoke detectors: they don’t predict (and help you prevent) churn, they just alert you when it’s too late. Like they say, “Where there’s smoke, there’s fire,” and old-school churn model’s only solution is to douse the flame that’s already started.
Within each service or app, there will be telltale signs for a consumer who’s at risk for churn (low engagement, payment lapse, etc.,). The trick is in making sure this information is surfaced and acted upon with engagement amplifiers in a timely manner to anticipate and prevent customer disengagement (or the fire) from ever happening.
If you’re reaching out to customers after they’ve been infected by a zombie virus, you’re considerably less likely to save them because customers become zombies long before your traditional models can warn you.
Whether someone has already lapsed or at risk for churn, it’s equally important to ask the question, “Is this customer worth saving?” You can only answer that by knowing two things: the root cause of why each individual customer is at risk, and the value you place on keeping them.
Many companies simply blast every lapsed customer with a generic turn-around offer, but having answers to the following questions will help guide which discounts, upgrades, and engagement offers to dynamically present to each customer segment:
- Why is the customer at risk of leaving?
- Do you have a remedy for it?
- How likely is the customer to come back?
- How much will they spend if they come back?
- How likely are they to stay and for how long?
The antidote can be spread as quickly as the virus
Just like bad news travels fast, a good experience can send ripples through an at-risk customer’s community of relationships or social graph. Using the built-in social graph function of the Amplero too, we ran a test across 6,000 users of one of our larger clients.
Our study quickly found that social connections of customers increase their consumption and are less likely to churn due to a campaign that was neither targeted toward them nor offered any direct incentives.
Likewise, it’s possible to stimulate increased usage throughout an entire customer social ecosystem through intentional sharing campaigns.
For example, imagine a dormant customer receives a notification from a friend and responds within five minutes. You then classify them with a high propensity to become a more active customer. Within 12 hours, the customer who sent the original notification could be given an offer to win two movie tickets if they answer a movie quiz. As part of the rules, they can text four people that they’ve texted in the past 24 hours for help, and they’ll also win two tickets each. You’ve now spurred increased usage from the at-risk consume, and made them more open to an upgrade or renewal offer.
In a zombie hunt, it’s the equivalent of injecting the antidote into a person who has been bitten, and all the surrounding zombies also get well.
Not bad, right?
Gearing up for your own zombie hunt
Any good zombie adventure requires some serious gear. For marketers, the tools of choice rest on the ability to act upon consumer behavioral data to continuously and dynamically optimize the customer experience.
Traditional marketing technology tools and analytics processes aren’t up to the task. In Forrester’s recently released 2016 Digital Intelligence Playbook, they point at that the “backward-looking analysis of aggregated online customer interactions is not enough for firms to maximize the opportunity of digital customer interactions and to compete in the age of the customer.”
Zombies or not, exploring machine learning-powered digital intelligence platforms to rapidly ingest and act upon your consumer data is a mandate for enterprise marketers in 2017. Otherwise, you could end up dealing with a full-on zombie outbreak.
Fighting customers that silently become zombies is not for the faint-of-heart-lean-back-on-your-heels-fear-of-the-unknown marketer. It takes bold marketers with vision, who lean forward at the prospect of tackling one of the most complex problems facing marketers.
It takes the types of marketers who are thinking about customers as relationships over transactions, subscription of SKU sales, and retention over acquisition.
For more information on leveraging machine-learning to predict and prevent, check out our case studies or contact us.
Glenn Pingul is VP of Scientific Marketing Strategies for Amplero, a machine-learning-powered digital intelligence platform that helps enterprises better connect with their customers. Prior to joining Amplero, he was a co-founder of the online video advertising ad tech company, Mixpo, where he was VP of marketing. He has an extensive background developing digital marketing, analytics, and loyalty and retention strategies while serving in executive level positions at T-Mobile USA, Nordstrom.com/Inc., AirTouch Wireless (Verizon Wireless), Starbucks.com and The American Express Company.