How does Stay Ai determine churn risk, and what factors are considered?
Stay Ai’s engine determines if a subscription is a churn risk based on cluster analysis and predictive modeling. Over time, as your data set grows and our live learning model is trained, the strength of churn risk analysis deepens.
Factors evaluated in this algorithm considered include (but are not limited to):
Customer location
Length of subscription in days
Order frequency
Order price
Discount percent/amount
How do I best use the Churn Risk Analysis graph to optimize my subscription program?
The Churn Risk Analysis graph is a powerful tool for subscription programs at scale to determine where in the customer journey subscriptions are most at risk of cancellation. There are 3 ways to view churn risk by type, illuminating where brands can take proactive action to prevent subscription cancellation.
Churn Risk Types
Days: This breakdown allows brands to compare the risk of churn in relation to the total number of days on subscription.
Lifetime Value: This breakdown allows brands to compare the risk of churn in relation to the lifetime value of a subscription.
Total Orders: This breakdown allows brands to compare the risk of churn in relation to the total number of orders (or order cycle number) of a subscription.
From there, data can be further drilled down based on Category and Cohort.
Category: This dropdown selection box enables you to view churn risk based on the following:
Baseline: compares churn data across all brands on Stay Ai. This is essentially the “middle ground” for helping brands to gauge their churn risk against benchmarked data.
Start Month: The month in which the subscription began.
Product: The product(s) included in the subscription.
Variant: The product variant(s) included in the subscription.
Cohort: This dropdown enables you to select the start month of the subscription (Start Month), the product(s) included in the subscription (product), or the variant(s) included in the subscription for deeper analysis.
Pro Tips
Breaking down Churn Risk by Total Orders is helpful for understanding where in the customer journey (number of orders) a subscriber is at risk of cancellation.
If you see a jump in subscription churn between order 2 and order 3 for example, that’s a great place to test run an ExperienceEngine promotion with a free gift or discounted upsell incentive for subscribers to stay loyal.
For brands that offer subscriptions that don’t run on standard 4-week (monthly) cadences, the Churn Risk Analysis graph is helpful for illuminating churn by days on subscription or by subscription LTV.
Think of this as a more advanced and detailed way of reviewing your cohorts rather than using a standardized monthly cohort view.
For your loyalty programs, viewing Churn Risk by LTV can be helpful for determining loyalty-based incentives.
Understanding average LTV before churn highlights where you could implement spend-based incentives to inspire future purchases.
How can I get the most out of the Subscription Timing graph on the Executive Dashboard?
The Subscription Timing graph details the breakdown of your active subscriptions based on the frequency of order cycling.
For example, in this image, you can see that 51.26% of subscriptions recur on a monthly (30 days) basis, 21.16% of subscriptions recur every other month, and 4.7% of subscriptions recur on a quarterly cadence. This breakdown is valuable for optimizing the order frequency offerings you provide to subscribers, as well as the default order frequency presented for new subscribers on signup pages, such as the PDP.
Another example: if you're seeing a high skip rate, and most subscribers are currently on a 4-week subscription cadence, consider switching the recommended (or default subscribe & save) option on the PDP to every 6 or 8 weeks.
When is it most helpful to view retention by subscriber versus retention by subscription?
Viewing retention data by subscriber or instead by subscription is truly based on a brand’s internal reporting structure. While some brands find it valuable to track subscriber retention, others prefer to track retention by actual subscription ID.
A few important things to call out when reviewing this data:
A brand could see 95% subscriber retention, but a large bulk of those subscribers are actually skipping or paused for multiple months. In the case of a brand with high skip or pause rates, it’s often most helpful to look at revenue retention or month-over-month performance data.
When reviewing data by subscriber, it’s important to consider that subscribers may have multiple subscriptions, thus skewing average subscriber numbers, like AOV, for example.
Looking at the New & Churned Graph, how can a brand best action on this report?
We created this report in 2023, as it was a highly-requested feature from our merchant brands.
Here are a few ideas for using this data:
For brands that do not use Klaviyo as their ESP, seeing a daily report of new & churned subscribers enables targeted email sends for onboarding/welcome information, win-back attempts, or feedback requests. Note: For brands that do use Klaviyo as their ESP, once integrated with Stay Ai, this event information is passed into Klaviyo for setting automated flow triggers.
For brands looking to conduct 1v1 Customer Support outreach with churned subscribers, exporting a list of daily churned customers can be helpful for email outreach.
For graphs that display the LTO metric, why does this matter, and how can brands act on LTO data?
LTO calculates how many orders per subscriber have been completed, starting at 1 month since the subscriber’s first subscription order. This metric is helpful for understanding the health of your subscribers. LTO illuminates how frequently, on average, your subscribers have skipped, paused, or canceled their subscription orders based on the cohort you’re reviewing.
Let’s pose an example situation: Let’s say it is currently November 2023, and you’re looking at your September 2023 new subscriber cohort, in which you acquired 72 subscribers. Each subscriber is on a monthly order cadence. If each subscriber actually completed their order over the course of the past 3 months, you would see a total of 216 orders completed - a 100% completion rate. (LTO = Cumulative Orders / Subscribers, so 216 cumulative orders / 72 subscribers = 3 orders over the last 3 months).
If some subscribers have skipped or paused orders, or they churned, your LTO would be less than 3. If only 150 orders were completed by those 72 subscribers, your LTO would equal ~2 (150 cumulative orders / 72 subscribers = ~2 orders over the last 3 months).
As we mentioned before, your LTO metric highlights subscriber health - and if your cohort of newly acquired subscribers are valuable subscribers, who are actually completing their orders. If you’re seeing low LTO numbers, consider what sorts of promotions or ads you were running when acquiring subscribers. Low LTO might indicate that you have high churn rates, or that during that time period, you acquired deal-shopping subscribers who aren’t generating that much revenue for the business.