CLV Part IV: Choosing a Time Horizon
From earlier: customer lifetime is the average point when a sub will cancel, within a period of time. To calculate customer lifetime, we have to set an arbitrary end point for our model. It can not go on forever, into infinity & beyond. What’s an appropriate period of time for our projection?
The "period of time" is a decisive assumption and, while there's guardrails, it's a bit subjective. It depends a bit on your own plans, and the degree to which you want to be conservative or ambitious. For those following along in the model, we're now on the CLV Model tab. This post will cover the driver assumption circled below.
Longer Horizon = More Risk
In general, the longer the period of time in our CLV model, the riskier. We're trying to predict the future. We're far more likely to accurately predict what will happen tomorrow than 5 years from now. As we get further from the present moment, the "cone of possibilities" expands, and the model takes on more risk of being inaccurate.
Let's say our hypothetical newsletter, Big Dean's, is doing this as an experiment or a side hustle. They're unsure if this will be a long-term commitment. The team (or individual) behind Big Dean's has been known to cycle through one-off efforts. Further, this is not their primary revenue stream. They're fortunate to have an impeccable restaurant right next to a popular beach pier. Revenue from the restaurant and merchandise sales are their primary sources of income. In this scenario, it may be prudent to be conservative, and assume a 12- or 24-month max lifetime. Who knows if Big Dean's newsletter will even exist for that long...
But perhaps after a handful of months, the Big Dean's team realizes they are on to something. Subscriber growth has exploded out of the gate. The newsletter is helping sell more t-shirts and drawling lines of people to its restaurant. The team is having a ton of fun running the newsletter. They decide to make more of an investment in developing and growing the newsletter.
Also, crucially, they now have data on how well they are retaining existing subscribers (more on this in the next post). In this more committed and informed scenario, the team considers adjusting their time horizon to a 24- to 36-month max lifetime.
Waiting for the Cows to Come Home
A CLV model will produce a definitive financial metric. But it's important we don't take it as scripture - we're trying to predict future revenue. As stated above, there's a margin of error in our prediction, which grows as we extend the model's time horizon. Longer horizons not only have a larger margin of error. They require more dedication and patience.
For example, a 5-year horizon is a commitment to continuing the product for that long. Given the nature of subscription models, the product must, at a minimum, maintain its value. If you stop shipping your product for a few months, that may have a material impact on the entire decay curve.
We also need the patience to collect that predicted earnings over a 60-month time frame. Don't fool yourself into thinking the entire CLV amount shows up in your bank account when a new sub signs up.
How to Choose a Horizon
Summarizing the above, we can start to develop a framework to help choose an initial time horizon:
Like many decisions, it boils down to how conservative or aspirational you want to be. For those just starting, it may be prudent to start with a 12-month (or maybe 24-month) projection. As you take in data and find your groove, consider lengthening your time horizon. And if you're extending the model's horizon, do so with awareness of the risks and commitments.
Okay, now the fun begins. Next we dive into decay curves - the backbone of the CLV model and one of the most important subscription metrics.