If you haven’t yet had the time to read our latest book on leveraging customer analytics to build a scalable churn prevention system, here are the top takeaways.
1. Retail is evolving faster than ever before in the history of the known universe. 30% of U.S. customers say they change brands often just for the sake of variety and novelty, and 49% of customers will gladly switch brands—for a coupon.
2. The economics of saving customers from churning really adds up; repeat buyers are essential to long term growth. Preventing just 1% of your very best customers from churning can lead to an overall revenue increase of 5%. Yowza.
3. Most retailers today rely on broad-stroke rules to guide their churn prevention program, such as “send a discount to anyone who hasn’t made a purchase in 90 days.” These rule-based programs have two big flaws. One, the arbitrary, blanket “90 day” rule is often too late or too early for most customers. Two, though conceptually simple, even rule-based programs can be a real nuisance to manage.
4. We recommend a more productive, personal way to implement a churn prevention program.
Rather than relying on blanket 90-day rules, we’ll use predictive analytics to detect when specific shoppers start veering off their norm.
Most effective churn prevention systems share five core capabilities:
Consolidate customer data: Customer data (including CRM data, transactional data, email data, on-site data, and more) is often siloed across data sources and marketing tools. With data spread across so many places, marketers struggle to build a complete view of their customers.
Predict behavior: the key to saving a relationship is detecting early warning signs of churn. This is where predictive analytics comes into play. Or, as the wind blows with popular terms of the season: Machine Learning, Artificial Intelligence, Fancy-Pants Statistics, Really-Good Math, Magic-Eight-Ball-Action, etc. Whichever term you prefer, they all perform the same function: using algorithms that apply some form of pattern-matching technique to highlight customers who show signs of veering from their expected purchase behavior.
Connect marketing tools: Most marketing teams utilize several different tools for customer communications. One tool to send email. Another for marketing on Facebook. Another for display ads. A different team or tool that manages direct mail. And so on. Churn prevention systems really take off when the segments of “at-risk” customers are automatically synchronized with each of those tools. Once they are, teams can set up automated programs to launch relevant, targeted campaigns on a weekly basis.
Test and measure performance: We need to be willing to try a variety of ideas—from communication strategies to incentive packages—until we find which approaches connect with various customer types. As marketers run campaigns across email, direct mail, Facebook, and display, it’s important to track how many incremental customers are being saved. This requires a system that makes control group testing easy.
Automate campaign delivery: Automation is essential for a successful churn prevention system. As the team runs experiments and observes results on which messages are working with different types of customers, they need a way to set effective tactics on autopilot. That’s when the machine starts rolling. It’s kind of like cruise control.
5. Once these core capabilities are in place, it’s time to put the churn prevention system to work. We advocate a crawl, walk, run approach. Through each stage the program grows in two main ways: the number of sub-segments targeted within customers identified to be at risk of churn, and the number of marketing channels used.
In the crawl stage, we suggest focusing on one channel. Email is the most common starting point and often has an immediate impact on churn with the least amount of effort. We won’t roll out the fancy micro segments just yet. We want to test a handful of communication and promotional tactics for customers within each risk stage, see what works, then automate those effective ideas.
With a basic system in place and running smoothly, it’s time to kick it up a notch.
The goal of the “walk” stage is to get even more granular with our segmentation. Now that we’re already segmenting by churn stage, the next step is further segmenting by value tier.
We’ll focus on the churning customers with the highest probability of return (the “cooling off ” shoppers) and sub-segment that group by three tiers of value: high-spending VIPs, medium value shoppers, and lower-value, discount-oriented shoppers. We’ll also expand into an additional marketing channel (Facebook) to reach and win back even more customers.
Within a few weeks, there should be an evident impact on churn. Excellent.
It’s time to layer in some more nuanced behavioral segmentation, “personas,” and add a new channel, direct mail. These aren’t your ad agency’s personas. Remember, your churn machine employs advanced statistical capabilities, so let’s see what else we can discover to further inform our churn prevention communications. One technique known as “cluster analysis” segments customer populations into groups based into distinct purchasing patterns and demographics.
We should continue to refresh email creative periodically (e.g., quarterly) and look for additional optimization opportunities, but at this point, it’s time to transition the program into an always-on, steady-state mode and look to the next set of incremental opportunities.