How-to

How to Reduce Attrition with Predictive Technology

By Kunal Das

Remember the Apple promotion “there’s an app for that?” Now more than ever, we live in a world where there seems to be different technology solutions for everything you could possibly think of.

At first, it seems fantastic — if you have a problem, there’s a solution out there somewhere. But for businesses, it can become all too easy to fall prey to “shiny object syndrome” and get caught up devoting time and money toward technologies that don’t actually deliver incremental value.

So how can you distinguish between shiny objects and truly valuable solutions? Plain and simple: Follow the money.

The Value of Predictive Technology

Predictive technology, in particular, can very easily seem like a shiny object. After all, the ability to predict the future sounds like it’s all fun and games, right? Well, as much as it might sound that way, AI-driven predictive technology is quickly becoming one of the most valuable solutions in retail marketers’ tool belts.

Specifically, predictive technology delivers significant monetary returns by providing accurate insight into data and increasing efficiency in decision-making and actions. Together, this creates what I like to call “turnkey decisioning.”

Turnkey Decisioning in Action: Reducing Attrition

To fully illustrate the value of turnkey decisioning powered by predictive technology, let’s look at the use case of reducing attrition.

The Problem: Identifying At-Risk Customers

Traditionally, retailers have used standard models for bucketing customers into lifecycle stages of active, at-risk and lost. Under this standard model, someone who hasn’t made any purchases in 90 days is considered at-risk. While this model has been around for quite some time, we all know it’s flawed because everyone buys at a different cadence. Still, it’s been the best measure we’ve had until now.

In addition to not taking into account the unique behaviors of individual customers, the status quo of identifying at-risk customers also presents challenges when it comes to determining who hasn’t purchased anything within the standard 90 days. Currently, mining data to build such a list can prove a time-consuming task.

Taken together, these two challenges make it difficult to effectively prevent customers from churning.

The Solution: Individualized Lifecycle Models Coupled with Real-Time Insight

Predictive technology provides a better solution than the standard lifecycle models by taking into account each customer’s unique buying cadence to provide an individualized model for lifecycle stages. As a result, one customer might be designated as at-risk after not making a purchase in 30 days while another customer might not be designated as at-risk until not making a purchase after 150 days.

Beyond providing more accurate definitions for when customers become at-risk, predictive technology should also make it easier to discover these findings. In fact, it should take discovery out of the equation entirely by serving you details about newly at-risk customers on a silver platter each day.

In other words, the system should automatically identify customers who have become at-risk within the past 24 hours based on their own unique purchase cadence and then alert you once that audience is ready. And it should cycle through this process every day so that you can deploy reactivation campaigns for customers the moment they become at-risk.

The So What: It’s All About the Revenue

So what if predictive technology alerts you the moment customers become at-risk? Does it really make that much of a difference? The answer is a resounding yes.

For example, let’s take a look at how this all shakes out for one home goods retailer. After looking at the data, our predictive models found that for the home goods retailer in question, 86% of revenue over the next two years (based on predicted customer lifetime value) will come from customers currently in the active and at-risk categories. Narrowing in even more, 40% of revenue will come from customers currently labeled as at-risk. That means that should those customers churn, the retailer stands to lose up to 40% of revenue over the next two years.

As a result, if the retailer can properly identify when individual customers become at risk in near real-time and then quickly and easily take action on that insight by launching targeted reactivation campaigns, the team stands to retain the 40% of revenue that it might have otherwise lost. How’s that for a “so what”?

Turning Insights Into Action with Predictive Decisioning

Of course the linchpin to all of this is the ability to take action quickly and easily on the insights that predictive technology delivers. That’s where a decisioning platform that can surface these types of insights and then push that information out to any customer engagement channel  comes into play. From there, it’s up to retail marketing teams to engage with customers and prevent churn — but with the right technology in place, everything aside from what to say should be extremely turnkey.

Kunal Das

Kunal is a Key Account Director who works with Bluecore’s largest clients. He has 12 years of marketing and advertising technology experience and thrives on being able to help his clients achieve business outcomes by pushing the envelope to do what’s never been done before using innovative Bluecore technology.