Marketers who are responsible for reducing passive attrition are in for a big challenge today. Predicting when a customer is about to leave is certainly not an easy job. But with the ability of marketers to identify “churn signals” before it’s too late, they can proactively communicate with customers, send them relevant offers, and entice them in various ways to remain active.
Today, we
have to be able to predict churn in the best possible way as it helps us:
Take action before it’s too late
Entice certain customers rather than
target our full customer pool
Now, there
are many different methodologies to calculate and predict churn but most of
them are based on purchasing behavior, or lack thereof.
While
purchasing is a critical factor in predicting churn, it is not always a
frequent activity and there are other elements that can be used to supplement
the churn risk prediction.
Let’s talk
about these elements in more detail.
Email
Engagement
Most
marketers use email engagement data as part of a static journey or to limit
audience sizes and reach better deliverability.
We
implemented an analysis to make it simpler for you. In this analysis, we took
samples of long tenured Active customers across different brands. These
customers were not new ones and did not make a purchase in the last few months.
Emails Opened and Churn Rates:
We can
conclude that despite these customers who are not purchasing, opening emails is
as predictive for remaining active as actual purchasing. Customers who cross
the 30-day mark are 35% more likely to churn than customers who opened an email
in the last month.
Therefore,
it is recommended that criteria for email engagement should be layered into
your churn prevention campaigns.
For
customers who are not “Email Engagers,” try taking the following actions:
Target via other channels – On site
pop ups, Facebook, Google, Instagram retargeting, Push, SMS
If you do want to stick to email – Use
subject line testing and optimization to increase likelihood of opens
Site
Visits
“Abandon series”
campaigns, such as customer who abandon carts and abandon browse, are all
common use cases. But have you ever used browsing activity in a more holistic
manner rather than as a trigger?
By analyzing
site and app visits of multiple purchasing customers (Active customers who
haven’t purchased in a few months) since their last purchase, we can see
significant differences between visiting customers and non-visiting customers.
Site
visitors are 75% more likely to purchase again (even though they
haven’t purchased in the prior months.)
Site Visits and Churn Rates:
One way to
address these types of customers and “give them a nudge” is to address their
site visit with a “we noticed you noticing us” campaign. You can also try
focusing the messaging of the campaign on encouragement to visit the site
rather than on purchasing.
The content doesn’t have to be promotional since they are already expressing interest and showing intent. They are probably more interested in your current offering and may need a personal touch to encourage them to take an action.
Here are some great examples that you might want to mimic:
Customer Service Tickets
The information you get from customer
service tickets is typically used for a variety of use cases, such as assessing
satisfaction of service and products, when your customers are seeking support,
and seeing how effectively support agents are solving customer inquiries. But
rarely (if at all) are customer service tickets used for predicting propensity
to churn.
Surprisingly, customers who
contact customer service are 36% less likely to churn!
Now, you may say that the
customers who contact customer service often are customers who are also
purchasing often, otherwise why would they contact customer service?
But again, this is a sample of
actively purchasing customers with multiple purchases, looking at their
lifetime customer service communications.
The hypothesis is that customers
who contact customer service, or at least long-tenured active customers who
contact customer service during their lifetime, are more “invested” in the
brand and are taking an active step to get a resolution from your brand rather
than just returning the product and “cutting ties.”
Customer Service Tickets and Churn Rates:
Final Takeaways
We spoke about 3 examples of
additional nuggets of data that can be used on top of standard purchasing based
churn prevention criteria – email engagement, site visits, and customer service
tickets.
There are still more crumbs of
activities and data points that can be utilized, such as reviews, and survey responses
to reduce churn that we will talk about in the future.
Yoni Barzilay is Optimove’s Director of Data Science, North America. He has a knack for finding creative solutions for extreme data challenges, and has led some of Optimove's biggest e-commerce onboarding projects. Yoni holds a BSc in Industrial Engineering and Management, specializing in Information Systems.
