I'm tired of hearing about LTV:CAC being a useful metric for SaaS/subscription
I think it should be EXTINGUISHED from our vocabulary.
Why?
1. LTV is a shoddy way to measure value capture
2. You need piles of VC $$$
3. There's a superior metric you should use instead
β
LTV:CAC is fundamentally trying to answer two questions:
1. How quickly can I recoup the cost to acquire a customer?
2. How much can I reasonably afford to pay to acquire a customer?
We'll get back to these at the end.
Here's why LTV does a terrible job at answering those questions...
(and this is coming from someone who used to work at Baremetrics... I've quite literally seen thousands of companies' metrics)
Baremetrics' definition:
"LTV is an educated guess at how much you can expect to make from the average customer before they churn."
This definition essentially assumes that most/all customers eventually churn.
It also assumes that revenue per customer stays relatively static.
This is far from the truth.
In fact, in a healthy SaaS business, it's the exact opposite.
(1) Customers stick around for quite a long time and (2) revenue per customer increases over time.
Colin shows Customer.io's cohort retention. The bottom orange part shows steady revenue all the way back from the 2013 cohort.
In fact, Customer.io has exceptional expansion revenue with most cohorts GROWING in revenue over time.
twitter.com/alphacolin/status/1349121874488279040?s=20
LTV just isn't meant for SaaS.
When one-time sales were the norm (including physical products, software, and services) LTV was super easy to calculate:
The total $ amount customers have spent with you is divided by the number of customers.
LTV:CAC makes a lot of sense there.
That standard calculation of LTV works well when there's a very small standard deviation from the mean.
e.g. LTV per customer looks like: $100, $90, $110, $95, $120.
LTV = (100+90+110+95+120) / 5 = $103
If CAC is $20, you've got a 5:1 LTV:CAC and "profitable" on each customer.
SaaS usually has large standard deviations from the mean because (1) subscriptions take time to accrue revenue and (2) price points can vary largely.
E.g. LTV per customer looks like $20, $100, $500, $5000, $30.
(20+100+500+5000+30) / 5 = $1,130
If CAC is $200, you've got ~5:1 LTV:CAC... right?
Not exactly. You're not profitable on each customer β only 2/5 customers.
What if the vast majority of the customers you acquire are on the lower plan? What if your first 3 months retention is only 50%?
You're in deep trouble.
If you use that formula, you're making way too many assumptions and not accounting for the dynamics of the subscription business model.
That's why you don't see people using that formula of LTV for SaaS anymore.
So we made a NEW formula that attempts to adjust for the subscription model:
Average Monthly Recurring Revenue Per Customer Γ· User Churn Rate = LTV
Simplified: ARPC/Churn
But there are still fundamental flaws. π€¦ββοΈ
It's a classic case of the danger of working with averages.
Averages don't tell the whole story. They're skewed by the range and outliers.
And if you're not careful, it can bankrupt your business.
Let me illustrate with some hypothetical numbers...
First: ARPC still fails to account for the different pricing tiers you may offer.
For example, maybe you offer $50, $100, and $500/mo plans, but your ARPC is $60.
$60 ARPC isn't an accurate representation of the revenue you're generating from each customer.
Second: Churn can be deceiving.
A seemingly innocent 5% monthly churn rate results inβ¦ a 46% annual churn rate! π€―
That means you have to replace half of your customers after a year.
Using the ARPC/Churn = LTV formula, if:
β’ ARPC = $60
β’ Monthly churn = 5%
Then LTV = $1,200
If CAC = $200, it might seem like you're in the clear with 6:1 LTV:CAC.
However, this means that you technically won't be profitable until month 4 on average.
And by month 12, half of them will likely be gone.
How much profit is left after that? π
You'll be completely out of cash before you can find out.
The good news is that you can get away with this if you have piles of VC money to (1) afford to wait until month 4 and (2) pay for everything else on top of the cost to acquire customers at scale.
But LTV:CAC is just a proxy for another, far superior metric you can track.
Payback Period is what you REALLY want to know and should completely replace LTV:CAC for SaaS companies.
LTV:CAC hates that you have to wait to accrue revenue.
Payback Period is totally fine with it. π
To calculate Payback Period, simply divide the average cost to acquire a customer by the average first month's revenue:
CAC/ARPC = Payback Period
To avoid the same delusions of working with averages, you can break it out by dividing CAC by each monthly price point you offer.
To use some of the same numbers from earlier, that might look like:
Plan A Payback Period = $200/$50 = 4 months
Plan B Payback Period = $200/$100 = 2 months
Plan C Payback Period = $200/$500 = ~2 weeks
> 1. How quickly can I recoup the cost to acquire a customer?
Payback Period answers the first fundamental question directly so you know how much cash you need to have to float acquisition costs.
We can check that box β
> 2. How much can I reasonably afford to pay to acquire a customer?
Payback Period helps you answer the second fundamental question by giving you a basis for understanding *when* a customer becomes profitable.
We can also check that box β
Personally, I don't believe you should spend any more than the first year's revenue to acquire a customer.
For example:
Plan A = $50/mo = $600 max CAC
Plan B = $100/mo = $1200 max CAC
Plan C = $500/mo = $6000 max CAC
If you know that larger ARPC customers are going to require different and more expensive marketing strategies, you can justify a larger CAC for them.
At a baseline though, you use your lowest plan to establish a conservative maximum CAC that you can use across the board.
Spending more than the average first year's revenue to acquire a customer is risky business.
Again, you can make it work with piles of VC money.
But unless you have amazing retention (<=1% monthly churn), you're doomed to fail.
Especially if your goal is to be profitable, you want to reduce your payback period as much as possible.
Ideally, your payback period would be 1 month or less.
That way, you're not running on a deficit and scale up and down depending on your needs and goals.
Plus, the question should never be "What's the maximum I can pay to acquire a customer?"
You should be asking "How can I pay the bare minimum to acquire a customer?"
To sum up:
LTV:CAC makes it easy to justify insane acquisition costs.
Payback Period is a much better representation of reality. π