Baketivity YoY Growth Soars after Analyzing Customer LTV's

Overview

Baketivity takes the hard work out of baking together with your children. They sell baking subscription packages for kids. Customers will receive monthly boxes bursting with new recipes, activity booklets, and their favorite baking kits, delivered right to their doors. They chose AdVenture Media to spearhead their advertising strategy on Google, Facebook and Instagram.

Goals

The client sells exclusively through their eCommerce website and the Account Management team actively manages both Google Ads and Facebook. The client wants to grow their business YoY.

Challenges

After launching the campaigns, the Account Management team soon found, depending on which subscription package, each package’s CPA is significantly different. 

Each subscription package has a different price point. If a customer commits to an annual subscription, the upfront cost is higher, but the average monthly fee is cheaper compared to a 3-month subscription package.

Though, using either CPA or ROAS obscures the true value of each subscription acquisition. Both Facebook and Google’s attribution models take only into consideration the conversion value at the point of purchase. What if a customer renews? What if the renewal rate for each subscription package is different? If both are true, we can be more aggressive with our CPA knowing that the lifetime value of each subscription is higher than its sticker price.

Data Analysis

We know our CPAs for each subscription package on both Google Ads and Facebook. We now want to find out the LTVs for each package. The client has a few subscription packages. 

• One Time Purchase
• Monthly Subscription
• 3-Month Subscription
• 6-Month Subscription
• Annual Subscription

We first exported multiple data sets by package. One list contains all the orders for one-time purchasers, another list for monthly subscription purchases, and so forth.

There are a few methods to calculate LTV. We adopted:

LTV = Sticker Price / Churn Rate

To identify churn rate, we analyzed each data set by cohorts. If 100 users subscribed in January and only 90 renewed in February, the churn rate is 10%. Since each data set has orders tagged by unique customer ID, we can iterate, observe the proportion of customers return after a period and calculate churn over time. For a 3-month package, we calculated churn by using a 3-month interval (customers can only cancel after the term ends).

Once we know the churn rate for each package, we were able to calculate LTVs for each.

Results

The analysis had two actionable insights. First, ROAS for each new acquisition is higher if we factor in LTV, regardless of the length of the package. The Account Management team can be more aggressive in acquiring customers. Second, the analysis identified that single-order purchasers have the lowest LTV due to high churn rate. The Analytics & Market Research team recommended focusing on acquiring subscription customers only. If any customer somehow only desires to make a single order purchase, we recommend nurturing these customers post-purchase and guiding them towards a monthly subscription package.