Split your repeat purchase flow based on predicted CLV
The more data insights you have at your fingertips, the more creative you can get with your retention efforts. By knowing a customer’s predicted CLV, you can strategically influence their purchase behavior by incentivizing them into spending more or placing more orders. This is where customer engagement programs come into play: loyalty or rewards programs, referral programs, even brand ambassador programs.
In this video, we will walk through how to export predictive analytics data for your key segments, and analyze it to find the average CLV of your audience. Then, we will create a repeat purchase flow triggered by Expected Date of Next Order that is branched by whether someone’s predicted CLV is higher or lower than the audience average.
Step-by-step guide
We’ve documented the same material covered in the tutorial video into step-by-step instructions. Follow along to build this flow in your own account, and modify the flow as needed based on your specific business, and any other retention flows you may have running in your account:
1. Find the average predicted CLV of your customer base.
- The goal here is to find the average Predicted CLV of all of your customers (anyone who has placed an order). To do this, click into your Has placed 1 or more orders segment. If you don’t have this segment, build it with Segments AI.
- Click Manage segment > Export segment to CSV
- Select the data points you want to include in the export: at minimum, include historic CLV, predicted CLV, and total CLV.
- Open the .csv file in a spreadsheet and calculate the average of each column. This will show you the average measurement for your customer base as a whole.
- Take note of the average predicted CLV. This is the number we will use to split our flow.
2. Build a flow triggered by Expected Date of Next Order.
- Build a date-based flow that is triggered by Expected Date of Next Order.
- Set a time of day to trigger the flow.
- Set the flow to “never” repeat. Choosing monthly or yearly would cause this flow to continue sending during fixed future time intervals, which is not what we want. We only want this flow to be sent when an Expected Date of Next Order is detected on a profile.
- Add a profile filter to remove folks from the flow if they place an order at any time since starting the flow. To do this, use the conditions What someone has done (or not done) > Placed Order > zero times > since starting this flow.
- If you have separate subscription reminder flows or replenishment flows running in your account, add a profile filter to remove folks from this flow if they have received one of those other flows in the last 365 days.
3. Add a conditional split based on someone’s predicted CLV.
- Drag a conditional split into your flow.
- Configure the split to say Predictive analytics about someone >is at least >INSERT YOUR AVERAGE PREDICTED CLV (the number we found from our .csv export).
Drag messages into each path and customize.
- YES path: this path is for customers whose predicted CLV is above average, meaning you don’t need to do much incentivizing to get them to purchase more. This is a great opportunity to collect testimonials, feedback, or encourage customers to join your rewards program if they aren’t already a member.
- NO path: this path is for customers whose predicted CLV is below average. This means you have an opportunity to incentivize them to spend more with the intent of raising their total CLV above what the prediction is. You may consider offering volume-based discounts, bundles, special free items with purchase, or some access to an “exclusive” perk that will only be given if they purchase within a certain time window.