Today in marketing, we focus on achieving a few KPIs, including acquisition targets. One way you might measure acquisition targets is Customer Lifetime Value (LTV). LTV, the amount of value a customer contributes to your business over their lifetime – from first purchase or contract and ends with churn — not only tells you if those new customers are worth the cost, but also how to identify and prioritize the highest value ones.
Understanding LTV helps you answer: is a customer’s value worth more (LTV – Lifetime Value) than their cost (CAC – Cost to Acquire a Customer)? When calculated correctly LTV is a powerful tool, when the number is erroneous critical decisions are made with bad information.
What are the basics of calculating LTV?
The simplest approach to calculating LTV is to rely on historical purchases and either Average Profit Per Use (APPU) or cohort revenue. To calculate APPU based LTV, calculate APPU per month.
Then add them up, and then multiply by 12 or 24 to get a one- or two-year LTV.
Cohort analysis takes the APPU approach one step further. Instead of calculating an overall APPU, cohort analysis calculates ARPU per month per cohort (a cohort is a group of customers who share an attribute or set of attributes; i.e. those who joined or made their first purchase in a particular month could be defined as a cohort.).
Whichever historical method you choose you never want to compute historical LTV as “total revenue ÷ total customers”. The issue is this ignores how long customers have been with you, which is slightly a big deal! Because Historical LTV is backwards looking it can produce misleading results if your company, the market or your customer’s behavior have changed, or will change over time. You can guarantee that one of these three things will change.
If that is the case, your only option is Predictive LTV. Predictive LTV combines data points (transaction size, frequency, margin, behavioral, etc) to accurately predict LTV even in cases where data points (such as consumer behavior) seem inherently unpredictable. This is where you’ll want to get your data analysis gurus involved as well. Both of these models involve require sophisticated statistics and modeling skills.
How can you start calculating LTV?
If you intend to calculate long term customer value, we wrote a whole series on this in the Data Driven Daily. First, you’ll need to gather the following data:
- average value per transaction
- transaction frequency
- customer lifespan
- profit margin
You should also consider how you intend to segment your customers for long-term analysis. One method of segmentation for customers is cohort based. Cohort is based on the year in which these customers became customers. You would have one cohort per year of brand existence. If your business is more seasonal or younger than a year, you might consider segmenting customers by day or month. Another way to segment might be based on the changes in your product GUI, so people who used your first product GUI versus second product GUI.
After you segment these customers, you’ll want to track metrics like engagement and revenue per cohort. We’ve seen this a lot with customers who track annual customer cohorts based on engagement KPIs and overall revenue contribution.