When you value your customers, your basic LTV model can be summarized with the following sub-metrics:
- Average order value
- 1st purchase conversion rate
- Subsequent purchase conversion rates
The speed at which your cohort converts for the first time, and all subsequent conversions is summarized in your cohort decay rate.
Your cohort decay rate is perhaps the single most important assumption that can skew your LTV analysis.
A decay rate is like a waterfall
Many business models will exhibit an accelerated first conversion that tapers off over time. This is particularly true in a direct response type business, or for direct response type marketing campaigns that are meant to convert prospects TODAY.
A single campaign that generates revenue over time is represented in the following curve:
Again, the curve represents a cohort with high conversions early in its lifecycle, slowing down over time. While the table's conversion numbers are entirely made up, they broadly exhibit a similar cohort decay.
In the photos above you can see that the waterfall, or cohort decay rate, is constrained to a certain amount of distance, or time. This amount of time is defined by your LTV threshold, or the window in which your business is comfortable assessing customer LTV revenue.
In this example, each of the four cohorts on the right exhibits a different decay rate. This behavior is likewise typical in the real world, where different groupings of campaigns are likely to convert in similar ways. This behavior pattern makes it very logical for businesses to perform cohort LTV analysis at the "acquisition source" level.
Ways to change your cohort decay rate
The long term consequences of altering your cohort decay rate can translate into revenue upside.
Your business is a pie. If you're able to improve the speed at which cohorts convert and re-convert, you are growing your pie. Firms that realize this early into their acquisition efforts improve their chances of pie-growing before competition has time to eat your pie.
Don't forget about customer returns (eCommerce & Digital Goods), unsubscribes (list businesses), and uninstalls (Apps).
Here are a few ways you can improve your cohort decay rate:
- Improve post-acquisition customer CRM engagement to portray value for your customers, improving conversion to sale (and repeat).
- Measure where warm leads drop off from your conversion funnel and proactively build a case for proving value with them before they become cold leads.
- Reduce refunds & returns by analyzing customer pain-points.
Refunds, returns, & things you don't want...
From the perspective of your cohort decay, there are two schools of thought on how to deal with "the returning of money to customers:"
- Bake refunds, returns, unsubscribes, uninstalls into your cohort decay to account for this.
- Do not bake these numbers into the decay, and instead create a refund-specific decay rate to modify your model's output based on refunds.
Depending on the size and scope of your business, there are arguments for using either method. My recommendation is to pursue number 2, as it separates revenue from returns creating two streams of pure metrics. By doing so you're more likely to apply the appropriate assumptions and catch any mistakes.
Inverse cohort decay rates
There are certain business models that will exhibit an inverse decay rate, where your customers spend little initially and grow spend over time. However, many of these cases are unsustainable, and all will eventually reach an upper ceiling. Here are a few examples:
- Some forms of SaaS businesses, where a firm receives discounted integration and setup costs before monthly recurring fees. Once integrated, the firm may often be asked to pay premium prices due to the costs associated with finding alternative integrations.
- Free trial businesses, where a customer receives value and teaser rates before they begin paying.
- Businesses dependent on feature upsells, where the basic product comes at a low cost but further upgrades come at additional fees.
- The drug market, where an individual gets hooked at discounted prices before they become addicted and pay full price.
Have more ideas on how to improve cohort decay rates? Mention them in the comments below.