4 reasons why LTV/CAC is not a great metric for early startups, and what to use instead
Introduction
Lifetime customer value over customer acquisition cost is one of the primary metrics that companies use to measure ad spend efficiency. Intuitively, it makes sense as a metric. For every dollar you spend, you want to make back more than a dollar. It's a useful metric for large companies that spend millions because a small improvement in LTV/CAC results in significantly more profit. However, there are many reasons why for startups, it's not as useful of a metric. In this post, I'll talk about a few of these reasons, and give an alternative metric to use instead.
1. Startups don't live forever
When calculating LTV, usually companies assume a user will retain between 3-10 years. However, a lot of startups don't have 3-10 years of runway to recuperate their ad spend loss. So, the lifetime of the user is capped by the lifetime of the startup and the "effective" LTV is lower than what's projected.
2. CAC for first 100 users is different than next 1000 users.
Every marketer is familiar with this. For example, let’s say you get one ad set + audience combination to have a 3X ROAS (return on ad spend). For every $1 you spend, you're making $3 back (minus cost to produce the service). Why not raise the ad spend sliders up to $1MM to make an easy ~$2MM profit? It doesn't work that way. As you scale your ads, your ROAS generally goes down to a point where it won’t be profitable past a certain point. Various factors are in play here, for example, saturation of audience and how ad platforms usually find the lowest cost users first. So, it doesn't make sense to run a small test, calculate a LTV/CAC, and assume that will apply for a larger spend.
3. Startups likely don't have enough data to accurately calculate LTV/CAC
There are a lot of variables involved to get accurate LTV and CAC numbers. For example, long term retention data. It's hard to accurately predict long term user retention when a startup doesn't have a significant amount of users over one year old. In addition, products change a lot over a year for startups so cohorts now may not behave like cohorts a year ago. So, most startups don’t have the data needed to accurately model LTV.
4. It's not worth a startup's time to calculate LTV/CAC correctly
To get an accurate LTV/CAC number, it requires significant work and usually an analyst or data scientist. It also requires segmenting your users by different cuts because different types of users have different LTV and CAC values. Then, you have to create an accurate model of how long an average user of that segment retains and spends. Finally, you have to predict the CAC for just that segment which is not that simple either. The large amount of time spent calculating an accurate LTV/CAC is probably better spent growing the company in other ways.
Use payback period instead
So now that you are hopefully convinced that LTV/CAC is not great for early startups, what should you use instead? Early startups don't have the data to predict the lifetime user revenue multiplier they get on ad spend, so the next best thing is to make the ad spend break even quickly. It's not great to spend $1 to get $1 back 5 years later, so the faster, the better. I've found that payback period is a good metric to use. Payback period can be defined as the time it takes to recuperate the cost of an ad spend through revenue. After the payback period, gross margin is almost pure profit. If your payback period is low (< 6 months), you know you can probably make a profit off of your ads before you run out of runway.
How do you calculate payback period? To calculate it accurately is also slightly complex because it involves predicting revenue over time, and may need to be done in user segments. Instead of spending significant time trying to get the exact payback period, my advice to early stage startups is that it should be very obvious if your ads are profitable, or they aren't worth running and optimizing at early stages of a startup. This means your payback period is 0 or close to 0.
For example, if you spend $1000 on ads which result in subscriptions netting $1000 in the first month, that's obviously a great payback period and you should keep spending. If you spent $1000 on ads which result in $500 in the first month, $300 in the second month, $150 in the third month, and that's all the data you have, then it's unclear how long the payback period is. It may or may not be profitable, but we can be pretty confident it's not *super* profitable. So, the opportunity cost of shutting down these ads is not high, and I would not recommend focusing resources on optimizing these ads in this case.
Summary
If you're an early stage startup, don't just use LTV/CAC because that's what the big companies use. It's an important metric for them, but not great for early startups. Instead, try to roughly estimate payback period. It should be obvious that your ads are paying for themselves, or they probably aren't worth the time and effort running. Best of luck running profitable ads!
Thanks Vincent Tian for giving feedback on earlier drafts of this post.
Get daily growth advice by following me on Twitter: @JeffChang30