This comprehensive guide tackles the many nuances of calculating SaaS metrics with usage-based and hybrid revenue streams
I have spent the last two years working with the SaaS Metrics Standards Board to develop a set of metrics all SaaS companies can adopt. As part of that journey, I was tagged as the “usage-based pricing guy” and was charged with understanding the nuances of UBP on each SaaS operating metric. No big deal, right? Well, not exactly.
Most SaaS Metrics are built upon the assumption that a customer, or cohort of customers, has a known annual recurring revenue (ARR) level. In UBP, however, revenue is unknown at launch and changes over time. The foundational building block of SaaS metrics, ARR, is quicksand for UBP companies. And it gets more complicated when you have a hybrid model and mixed revenue streams.
All is not lost, however. In many cases, contractual usage minimums can establish ARR levels; in other cases, ARR can be estimated or calculated retrospectively. Special care must also be given to businesses with seasonal revenue or customers that ramp up usage significantly during onboarding.
In the following sections, we will discuss how to handle the issues listed above and also identify which SaaS metrics are most useful for UBP companies.
Kyle Poyer from OpenView addressed this recently, noting that all recurring and reoccurring revenue should be included in ARR. For UBP companies, I would define that as contractually obligated minimums and any other revenue that occurs year over year without a new sales process. Because different products can have different types of revenue, I recommend that SaaS companies report their revenue as granularly as possible. By getting more granular with the different types of revenue, each can be judged on its own merits.
Most seasoned SaaS practitioners and investors isolate different revenue streams based on:
1.Contracted SaaS Subscription Revenue
2. Contracted Services Revenue
3. Usage-based Minimum Obligations
4. Usage-based Reoccurring (but not contracted) Revenue
5. Usage-based Overages
6. Reoccurring But Not Contracted Services Revenue
(b) Ongoing Consulting
7. One-time Services Revenue
(b) One-time Consulting
8. Hardware Sales
No SaaS business will have all these types of revenue, and some bespoke categories may be required, but this is a good list from which to choose.
The predictability of the revenue might be based on a contract, but not necessarily. I have worked with and funded dozens of SaaS businesses with highly recurring revenue demonstrated by years of retention data but with customers on month-to-month contracts. Investors want to know which revenue is contractually obligated, but the good ones will focus more on the retention history than the contracts.
So what’s included in ARR? All revenue that recurs from year to year on a predictable basis.
For example, I’m working with a SaaS company that provides managed service on a usage basis. An individual customer’s usage is sporadic throughout the year, but it’s consistent year over year with an annual retention rate of 95%. This is clearly ARR.
We also had a borrower at SaaS Capital who provided technology to running races like 5k’s and marathons. They were paid on a per-registrant basis at the time of the race. The size and timing of races varied each year, so retention was difficult to calculate, and revenue was lumpy. Nonetheless, there was little customer attrition, so over the entire year, revenue from the installed base recurred at close to 100%. Their revenue was very different from subscription revenue, but clearly ARR.
This is a critical topic when raising money or selling a business. Annualizing revenue in a subscription business is easy. Take the most recent MRR and multiply by 12. But what if your consumption pricing business has significant variability month to month or has seasonal revenue?
Standard month-to-month variability can typically be handled using a three or six-month average. Try out different periods and see what smooths the revenue line the best without sacrificing recent growth trends. For slower-growing or no-growth SaaS businesses, consider using a trailing twelve-month number instead of estimating a current run rate. It will be less scrutinized and won’t differ much from your run rate.
Seasonality is a bit more complicated. What is the run rate of a SaaS eCommerce platform that gets paid a percentage of revenues and recognizes much of its revenue in the last two months of the year? (See example MRR graph.) You could not pick three or six months that would accurately reflect its run-rate revenue. And using the trailing twelve months of revenue would understate growth.
ARR is attempting to capture what a full year of revenue would look like from the current base of customers.
The most accurate way to calculate ARR in a seasonal business is to determine what percent of annual revenue has been recognized each month. The following chart is based on the previous graph.
With this as your historical revenue spread, you can take the most recent one, two, or three months of revenue and divide it by the sum of the historical percentages for those months to calculate your ARR.
For example: if revenue for this company were $725,000 in the first quarter of Year 4, its ARR would be: ((725,000/(.044+.048+.042)) or $5.4 million. By contrast, its trailing twelve-month revenue is $5.1 million. The $5.4 million ARR number captures a full year of revenue from recently acquired customers.
Some SaaS companies will work to get their contracted minimums to be as close to expected usage as possible and minimize overages. In other companies, contracted minimums are established at lower levels, where customers don’t risk paying for usage they do not use, and the SaaS business can generate higher revenue through higher overage rates.
So one company's overage is another's contracted minimum, yet all this revenue behaves essentially the same.
With enough historical data and when working across a base of at least fifty customers, overages are relatively easy to establish as ARR. My experience has been that overages happen year after year at predictable levels, which can typically be demonstrated with historical data. Ultimately, the contracted minimum revenue may be seen as slightly more predictable and valuable, but it's not an order of magnitude difference in most businesses.
This can be hard. In cases with no contracted minimums, estimating a new booking number is the only way to generate it.
But let’s start with cases where there are contracted minimums. The standard definition of new bookings is the first year’s ARR for customers signed in the month/quarter. If a customer agrees to purchase $50,000 worth of units for use over the next 12 months, the new booking is $50,000. Implementation revenue is not included, and it does not matter at what point in the year the customer is signed or when they will be implemented (within reason). If your company books a deal at $50,000 annually for three years, new bookings are still $50,000, but the Total Contract Value would be $150,000.
In cases with no contracted minimums, the first year’s ARR must be estimated. This is often done with data from the customer and is part of the sales cycle. For example, an e-commerce platform provider with UBP pricing based on revenue will likely know how much revenue their new customer has achieved historically and can apply their own pricing and estimate a booking number. In other cases, it’s not easy to estimate. For example, Snowflake may sign a large enterprise customer, yet it does not know with any certainty how much work will be moved to its platform or when.
In PLG (Product Led Growth) businesses, it’s particularly hard to estimate the annual revenue of a new customer when they first sign-up. For that reason, most PLG companies don’t report new bookings; they simply report the number of new customers, possibly segmented by type.
When SaaS companies must estimate bookings, they don’t mean as much as they do in the subscription world. In a subscription business, new bookings should directly translate into revenue, dollar for dollar, with only the implementation time lag as a variable. That said, sometimes implementations take a long time, and sometimes customers drop out during implementation. If that’s an issue for your company, track the ratio of new bookings to new revenue to see if and when your bookings deliver revenue. That ratio is (Increase in ARR from new customers in period Y) / (new bookings in quarter X). Period X should precede period Y by the standard implementation period. A ratio of less than one indicates revenue leakage or slippage.
Revenue growth is the biggest driver of a SaaS company’s valuation multiple, and it is mathematically the same for a usage-based business as a subscription business. That said, the revenue growth drivers for the two types of companies are different.
New bookings minus churn determine a subscription company's growth. There might be a few new modules sold into the installed base, or there might be a price increase, but expansion revenue is generally limited and typically offset by churn. This typically equates to an NRR (Net Revenue Retention Rate) of 100% or less for these companies.
For UBP companies, however, growth is more balanced between new bookings and expansion revenue. Expansion revenue can include selling new products to existing customers or price increases but is typically driven by:
The absolute best way to track revenue growth from period to period is a waterfall chart. See below for a waterfall chart explicitly designed for UBP companies.
A vital distinction UBP companies need to make is: what should be considered new revenue? Is it only “new” in the period a customer was added, or does new revenue continue through the ramp-up phase?
By way of example, Snowflake customers start at a low level of revenue and take at least a year to migrate into something resembling their fully deployed revenue. If “new” revenue was only the first month or quarter, 90% of a new customer's total revenue would be captured as expansion. This would be misleading as new bookings are the primary contributor to growth. More on Snowflake later, as this issue significantly impacts net revenue retention.
The best way to address the ramp-up problem is to treat new customer revenue as “new” for a standard period approximating when most new customers are fully deployed. For many companies, this would be a quarter or two from the first month of revenue recognition, but for others, it would be much longer.
NRR is the annualized change in revenue from a consistent cohort of customers. For subscription companies, churned revenue is offset by up/cross-selling, price increases, and growth in seats of retained customers. NRR for subscription SaaS is primarily driven by the loss of customers (logo retention). To that extent, it’s an “OK” indicator of the stickiness of a business’s revenue and shows to what extent cross/up-selling and price increases can offset the loss of customers.
Over an extended period, NRR across all subscription SaaS businesses has averaged about 95%, with lost revenue from churned customers slightly exceeding expansion revenue from retained customers. NRR is also highly correlated with average selling price (ASP). The higher the ASP, the higher NRR is.
NRR in usage-based SaaS is more complicated because there is more expansion revenue in general based on usage revenue generally going up over time. Because the NRR metric offsets churn with expansion revenue, it’s sometimes hard to tell what’s going on under the surface by only looking at NRR.
For example, the waterfall graphs shown here represent two very different SaaS businesses, both with 100% NRR. The company on the top lost 60% of its revenue to churn and contraction, while the business on the bottom retained 90% of its starting revenue. Which one would you like to run or fund?
Ramp-up revenue from new customers discussed in the prior section also profoundly impacts NRR. Snowflake characterized much of this ramp-up revenue from new customers as expansion revenue contributing to an impressive 160% NRR in 2021. But when new bookings slowed, and there was less ramp-up revenue, NRR went down (a relationship that should be uncorrelated). Wall Street was confused and unhappy as it assumed the business would grow at 60% without any new bookings.
GRR is typically the “go-to” metric when trying to understand the underlying stability of a subscription business's customer base and revenue stream because it does not contain the “noise” you find in NRR from outlier customers or price changes. The GRR calculation includes losses from churn and contraction but does not include any expansion revenue.
For UBP companies, however, GRR can be heavily depressed by even modest variability in month-to-month revenue as GRR captures the downside of the normal variability but not the upside. It might be good for internal trending but not for external benchmarking or fundraising.
Logo retention is simply the percentage of customers you had a year ago that you still have today. It’s not influenced by expansion revenue or even the size of the customer. To the latter point, logo retention is a good indicator of the health of a business with large customers that skew revenue-based metrics. If a company has strong logo retention, all the other retention metrics will also be strong. The inverse is not true. It’s a good metric for usage-based companies because it cuts through some of the noise on NRR without understating retention, like the GRR calculation.
This has become the most widely used metric to track how efficiently a SaaS business can grow its revenue. CAC Payback is calculated as (customer acquisition cost for the period)/((ARR of the acquired customer's/12) x’s gross margin percentage). Basically, how many months does it take for new customers to repay their cost of acquisition? CAC typically includes all fully loaded sales and marketing costs.
This is a straightforward metric for subscription businesses, with the only real obstacle being matching the right cost with the right revenue using the average sales cycle. Most companies selling larger products lag the acquired revenue by three months from the sales and marketing spend, but it’s company-specific. The metric should be calculated on a rolling basis monthly and include at least one quarter of CAC and one quarter of new bookings.
By way of example, let’s say a SaaS company’s fully loaded sales and marketing expenses were $500,000 in Q1; its average sales cycle is 60 days, it had new bookings of $700,000 March-May, and its gross margin is 80%. The newly booked customer cohort will contribute $47,000 ($700,000/12)*.8) monthly to the business and pay back its acquisition cost in ten and a half months ($500,000/$47,000).
Usage-based businesses face the added challenge of determining the ARR for newly acquired companies. It’s the same challenge they face when trying to determine a new bookings number, and the solutions are the same. As discussed in the New Bookings section above, your approach will be driven by the nature of the UBP revenue.
Options 2-4 are all efforts to estimate the normalized future annual revenue of the cohort of acquired customers so CAC Payback can be calculated quickly. Another approach is to wait and keep track of the actual revenue. Every cohort will generate a contribution margin each month, and as you add those dollars over time, they will eventually exceed the cohort's acquisition costs. The month that happens is the CAC Payback period.
In this example, it’s month eleven.
I call this the wait-and-see or cumulative approach. The downside to the approach is that the metric will often be finalized 9 to 24 months in arrears and might be useless to management in adjusting their go-to-market spending. That said, at a minimum, this calculation should be done to validate the estimated approaches listed above. The metric could lead to fundamentally flawed management decisions if the estimates are wrong.
Like the CAC Payback ratio, the cash burn ratio is an efficiency metric. CAC Payback focuses explicitly on the efficiency of new customer acquisition, while the cash burn ratio is more comprehensive. The metric answers the question: How much money did it take to create a new dollar of ARR? The metric is calculated as (Net New ARR/Cash Burned).
This metric is a good reality check to see if cash-flow-negative businesses are creating value. If a cash-burning business is not growing ARR, it is destroying value.
Like all metrics that use ARR, this one is complicated by consumption pricing. But because this is a higher-level metric, if calculated on a trailing twelve-month basis, the distortions of UBP tend to fade away. I prefer to use EBIT to define the company’s cash burn.
The Rule of 40 is calculated as (revenue growth rate + operating profit margin) and has equal applicability to subscription and UBP businesses. There are no fundamental distortions to this metric brought about by usage-based pricing.
The metric is called the Rule of 40 because private equity firms developed it as a benchmark for their portfolio companies. Growth plus profitability for their portfolio companies was targeted to exceed 40%. (Not easy.)
The metric is a good but simplistic reminder that profits should be considered when evaluating a SaaS business. It’s less instructive, however, in early-stage companies where the growth rate can dominate the calculation. The metric also suffers from its intrinsic assumption that profits and growth are valued equally. From a valuation perspective, this is not true—growth rate drives valuation multiples more than profit margins. As of Q3 2023, revenue growth had a .4 R-squared correlation to revenue multiples, while the Rule of 40 R-squared was .19.
The same is true in the private markets. PE firms know how to make a high-growth business profitable, but it's much harder to turn a profitable business into a high-growth company. As a result, they will pay more for a growth business.
SaaS metrics based on ARR are difficult to calculate for businesses with variable revenue streams. But with some thought and careful estimation, variable revenue streams can be normalized or estimated to support metrics that are helpful to management teams and investors.