The Art and Science of Figuring out Your CAC Payback Time
I’ve talked a lot about CAC/LTV, how to measure it, and why it’s such an important metric, especially for later-stage SaaS companies. CAC/LTV is also a central theme of my “5 Ways to $100 Million” post, since the key take-away from that post is that SaaS companies need to align their GTM motion with their pricing in a way that they can achieve a favorable CAC/LTV ratio at scale.
However, for earlier-stage companies with less data on longer-term retention, and thus less clarity on LTV, another metric is often more practical: CAC payback time. As you probably know (and if not, might have guessed), CAC payback period measures the time it takes to recoup your customer acquisition costs.
CAC payback time is a simple and intuitive concept — which makes it so useful — but when you use it in practice, you may run into questions like these:
- What costs should be included in my CACs? The simple answer is “all sales and marketing costs related to customer acquisition”. But how about, for example, one-off expenses related to creating your marketing website or building your sales machine? How about the costs of AEs who didn’t work out and were terminated a few months after joining? Or, if you have a freemium model, the costs of supporting free users?
- How do I factor in account expansions and churn? CAC payback formulas usually have ARPA (Average Revenue per Account) in the denominator. Should this be the initial ASP (Average Selling Price) of a cohort, the ARPA of a cohort, or something else? How to account for logo churn?
- How do long sales cycles change the equation? How should I account for sales cycle length and other factors, like AE ramp-up time, that increase the time until I’ve recouped my CACs?
The bad news is that the answer to most of these questions is: It depends. It depends on what you want to know. The good news is that if you know what you want to know — if you’re clear on why you’re interested in your CAC payback time — everything follows. To use a somewhat overused term, you can work it all out from first principles.
So what is it that we want to know if we talk about CAC payback times? In my experience, the question you’d like to answer usually falls into one of these buckets:
1) “Did we break even on our sales and marketing spend?”
You may want to know if your sales and marketing spend from a certain period of time has already been recouped, or whether a specific cohort of customers has already reached the break-even point. In any case, you’re looking in the rearview mirror to understand the profitability of your customer acquisition in the past.
Here’s an example with dummy data:
The table shows the cumulative gross profit for a number of customer cohorts over the cohorts’ lifetimes. You can see, for example, that the January 2023 cohort has turned positive in month 8. Here’s a nice way to visualize the data:
Creating these charts is of course much easier than getting the data needed for them, so let’s talk about that. The revenue side of the equation is pretty straightforward: You just take the actual revenue generated by a certain cohort of customers and multiply it by your gross margin percentage (here’s where things become less precise because you probably don’t know what exactly you’ve spent on that cohort in terms of CoGS). That gives you a reasonably accurate number for the Gross Profit generated by that cohort over time.
The costs side is more complicated, mainly because of tracking and attribution issues. The general principle should be that you include all costs associated with acquiring the customer cohort you’re looking at.
Depending on your GTM motion, this may include, for example:
- Advertising spending to generate demand and acquire leads that you’ve spent weeks or months before the leads became customers (depending on the quality of your tracking and attribution system, you may have to work with some rough numbers based on your average marketing spending in the months before the conversion month)
- The costs of your sales team, including supporting roles like SDRs and sales management, and costs associated with hiring and training of your sales team.
Note that in the example above, several cohorts have already broken even, which makes it easy to (approximately) determine the payback time. If you’re earlier in your journey or your payback time is longer, you may not have any cohorts that have reached the break-even point. In that case you’ll have to estimate how your cohorts will behave (in terms of churn, expansion, etc) based on the data that you have. I’ve described how to do this in this post and this Loom.
2) “Should we double-down on customer acquisition?”
While the rearview mirror look is interesting, what you’ll really want to know, and where things become highly actionable, is what your CAC payback is likely going to be in future. There are two versions of this question:
a) What is our CAC payback time if we keep spending what we’ve spent in the past?
b) What is our CAC payback time if we spend more, potentially much more?
For (a), the starting point should be your historic numbers, what I’ve called the rear mirror look. You’ll mainly have to decide which cohorts you want to use for your estimate. Older cohorts have the advantage of being more mature, so you have more longitudinal data on them. The downside is that the data might be outdated if your CACs, ASP, or churn, have changed in the meantime. On the flip side, younger cohorts are incomplete but more closely mirror your current performance.
The right choice depends on your assessment of the data and various trends that you may see in it. As an early-stage startup you usually don’t have many quarters of relatively stable data, so you may have to use your average CACs of the last 3–6 months while you use the MRR and NDR numbers from older cohorts. If you were wondering why this post is titled “art and science of…”, now you know. ;-) This adds uncertainty, but if your CACs have changed significantly in the last 1–2 years, or if you’ve only recently started to spend money on customer acquisition, this may be your best bet. It all comes down to you taking an honest look at the data and trying to understand what it means for the future.
If (a) comes with uncertainty, things become even more certain with (b). Trying to predict how your CACs will change if you try to 3x, 5x, or 10x your spend is very difficult. I wrote a post about this question some years ago, so I won’t go into detail here. As a side note, while the change in CACs is the most critical factor, ARPA and NDR (Net Dollar Retention) may also change if your customer acquisition mix changes (e.g. if you’re planning to do more outbound or more paid marketing as you scale).
It helps to be conservative here. Unless you have very strong reasons to believe that your CACs will go down, I’d assume that they will increase as you scale. That’s because there are “low hanging fruit” effects everywhere, and positive scale effects (e.g. from having built a brand) usually don’t kick in until much later.
That being said, there are some costs that you may want to exclude if you’re trying to figure out your CACs going forward, like one-off expenses (e.g. setting up Salesforce) or costs related to mistakes that you’ve made and are not going to repeat (e.g. hires or marketing experiments that didn’t work out). Don’t slide down that slippery slope too far, though. Even if you don’t repeat the same mistakes, you’ll make different ones. :-)
3) “And what about cash?”
So far, we’ve talked about CAC payback times based on (gross margin adjusted) revenue. What we haven’t talked about yet is how long it actually takes until your CAC has been recouped in cash. Unless you’re sitting on a gigantic war chest that you’ve raised in the 2021 heydays, cash payback time may be what matters most so let’s take a look at that.
Depending on various factors, your cash payback time may be shorter or longer than your revenue-based payback time. Let’s take a simple example for each of these cases.
Short cash payback time: Imagine a prosumer or consumer SaaS product. Self-service, no-touch conversion, no sales team. Let’s assume it takes the company $100 (mostly ad spend) to acquire a paying user. Let’s say 60% of paying customers sign up for a $25 per month plan and 40% sign up for a $200 per year plan. Unless most of the monthly plan customers cancel in month 2, the company will recoup cash spent on customer acquisition as early as month 2. (Note that for simplicity, I’ve neglected CoGS in this example. Don’t do this in real life.)
Long cash payback time: Consider a B2B SaaS company with an ASP of $500 per month, an 80% gross margin, a sales cycle of around 2.5 months, and CACs of $5000. Let’s further assume that almost all of the company’s customers are on a monthly plan and (to keep things simple) that the company has an NDR of 100%. The company’s CAC payback period is 12.5 months, but given the 2.5 months sales cycle, it takes around 15 months to recoup the full CACs in cash.
This assumes you have a fully ramped sales team. If your question is “If we spend an extra $2 million on sales and marketing, how long will it take to break-even on that spend?”, you’ll have to keep in mind the time (and the money) it takes to hire and train additional salespeople, which can easily add a couple of months to your cash-based payback period.
Before I close, let me try to address a few more FAQs on the topic:
Q: How do I account for churn and expansion when calculating CAC payback time?
A: If you use the “average costs to acquire a customer” divided by “monthly gross profit per customer” formula, you’ll run into various questions: Should gross profit be based on the initial ASP or on the ARPA of a cohort or something else? What happens to CACs for customers that have churned after a few months?
The trick is to look at a cohort of customers rather than an individual or average customer. Instead of trying to figure out the payback period per customer, you calculate how long it takes you to recoup the sales and marketing spend associated with a cohort of customers. That way, your calculation is based on NDR (Net Dollar Retention) and takes into account churn, expansions, contractions, etc.
If you’d like to look at customer-level metrics you can, of course, take the cohort-based revenue data and divide it by the initial cohort size to calculate, for example, your first-month ARPA (also known as ASP) and your average monthly NDR, but I find the cohort-level view to be the best starting point.
Q: What’s a good CAC payback period?
A: The big caveat with benchmarks is that a CAC payback period that is good for one business may be bad for another business. What’s acceptable for you depends on a multitude of factors, such as your customer lifetime, potential for future price increases, how much cash and runway you have, and the stage of the market you’re in. I’ve written about this here. The Goldilocks Zone of SaaS Metrics.
That said, it can of course be very useful to look at benchmarks (ideally from companies with similar characteristics).
Here are a few great starting points:
- David Skok’s rule of thumb is “less than 12 months”.
- According to Benchmarkit, the median CAC payback period is:
- 11 for companies below $1M ARR
- 16 for companies in the $1–5M ARR range
- 17 at $5–20M ARR
- 18 at $20–50M ARR
- 22 at $50–100M ARR
- 25 at more than $100M ARR - The Meritech SaaS Index of publicly traded SaaS companies shows that there is a huge range of CAC payback periods (from 4 to 87 months), with most companies being in the (still very large) 10–50 months range. Keep in mind that public companies usually don’t report CACs on specific segments, cohorts or acquisition sources, so you’re looking at a blended value across the entire business, dividing the previous quarter sales and marketing expense by the current quarter net new MRR (multiplied by the gross margin %).
- According to this benchmarking report of Blossom Street Ventures and Lighter Capital, the average CAC payback period of SaaS companies at IPO was 1.2 years, and the median was 1 year.
- In Bessemer’s SaaS portfolio benchmarking report from 2021, the average CAC payback period is 15 months for companies below at $1–10M ARR, 21–24 months at $10–100M ARR, and 30 months above $100M ARR. In the 2023 State of the Cloud report, Bessemer says 12–18 months is good, 6–12 months is better, and 0–6 months is best.
Last but not least, OpenView has published benchmarks based on a large dataset.
Keep in mind that OpenView uses the formula used by Meritech for public SaaS companies, which doesn’t bake in the effect of NDR in the same way as the approach I’ve described above. As Kyle Poyar points out, to know what “good” looks like is therefore highly dependent on your NDR, amongst other factors. According to Kyle, companies with an NDR of below 100% should look for a CAC payback of <12 months and companies with 100–120% NDR can look for 12–18 months, while companies with 150% NDR might be comfortable with a longer CAC payback depending on cash reserves and product stickiness. Couldn’t agree more!
