This post is part of an ongoing series where I practically walk through important calculations, metrics and unit economics for consumer internet businesses. Today, we compare several different types of lifetime value (LTV) curves.
In my prior posts, I explained how to properly calculate revenue and margins, and then using those principles, I discussed how to properly calculate LTV. As you look at LTV curves, it becomes apparent that there are 3 main types - 1) Exponential, 2) Linear, and 3) Decaying. It is important to understand each type, see below:
Again, LTV here is the revenue generated by an average user in a particular month. The math is based on the retention of your users, or cohort retention. Add each month up over time and you get a cumulative LTV curve. Variables that affect LTV for consumer services are typically things like user retention, average order value and order frequency. Let's quickly examine each type of curve:
Exponential LTV - This is the best possible curve and is a mark of a sticky consumer service. The slope of the curve steepens because your user cohorts are generating more incremental revenue over time. Perhaps retention is getting better, or order frequency is rising, or maybe average order value is going up. To illustrate, if a user is ordering a single product for $60 each month, your LTV curve would become exponential if that user started ordering TWO $60 boxes each month, and eventually even THREE. Or, if the price of the good WENT UP to $100.
Uber is a great example of exponential LTV. When the service first started, users used the service infrequently, say once a week. Today, Uber has better liquidity and better service, so the average user is probably using Uber several times per week. Having a business with exponential LTV means that if you finish a year with $100M of revenue, you can basically do zero marketing the next year and your business will still grow at a healthy clip.
Linear LTV - Linear LTV, though clearly not as exciting as exponential LTV, is generally a healthy thing. It implies a consistent user behavior and a sticky service. If you have an ecommerce business and users are buying $300 worth of goods every single month without fail, then that is linear LTV (and very impressive). If you finish the year with $100M of revenue, you can do zero marketing in the next year and still generate $100M of revenue the following year without doing a whole lot.
Decaying LTV - Decaying LTV means that you are getting a decreasing amount of revenue per user each month, e.g. if a user was spending $300 per month initially and eventually decreased to just $150 per month. If a business finished the year with $100M of revenue and spent nothing on marketing the following year, then it may only retain a small portion of that revenue. I sometimes call this the "treadmill effect", in that the larger you get as a business, the faster you have to run in order to continue growing the business while making up for lost revenue.
Decaying LTV is not always that bad, particularly if you are getting really good "payback", that is, generating enough revenue per user in the early days in order to cover your customer acquisition costs quickly. If you're selling a $1,000 product and acquiring users for next to nothing, then it probably doesn't matter if that user never returns.
Unit economics can be tricky. In a future post, I'll discuss the right questions to ask yourself when thinking through LTV and payback.