Why “% of revenue” may no longer be the right analysis
Probably because they have nothing better to do, in May 2014 FASB decided to introduce ASC 606, which changed the rules around revenue recognition. Most notably for software investors, it meant that software sold on-premise could no longer be recognized linearly over the length of the contract. For example, let’s say you signed a 3 year deal with a customer worth $1M/year. With ASC 605, the previous standard, you’d just book $1M/year in revenue, but with ASC 606, you now book 80% of the value of the contract up front ($2.4M in this case), and then recognize 20% of the value of the contract linearly over the 3 year license term ($600k/3 = $200k per year). And it doesn’t matter if the contract is paid annually — this is independent from the payment terms.
As you can see in the overly simplistic example of a business with just the one $3M contract, this creates an annoying dynamic where there’s effectively a “revenue pull-forward” effect. You can see how this in effect screws up all the calculations of OpEx as a % of revenue.
In a growth business, you don’t really notice it because you keep booking more contracts that keep revenue high. You also start to see old contracts come back for renewal, which softens the effect. It’s also common for businesses to not book a mix of 3 and 1 yr contracts, and the 1 yrs don’t have this effect, so it’s blends out.
That all being said, it does create this Ponzi-like effect where when you stop booking these contracts, you really appreciate how little of the revenue you are recognizing in the tail.
The implication of this being…
I was in a board meeting last year discussing a $400k/year CMO hire for a company with ~$20M in ARR and ~$26M in run rate revenue (due to the ASC 606 effect of the revenue pull-forward). We were discussing this in the context of our S&M spend as a % of revenue, showing a bunch of comps and how we were below their spend levels.
In reality, that analysis wasn’t really apples to apples. We needed to think about this $400k hire as a percent of our $20M in ARR (i.e. 2.0% of the avg annual cash generation of the business, not the 1.5% of the $26M in revenue). By looking at this % of revenue analysis, it understated the scale/impact of the expense.
Is this going to change how we do spend analysis?
Well it’s certainly something to keep in mind when looking at an on-premise software business selling multi-year contracts, particularly when it’s a bit of a slower grower. Fortunately it doesn’t have much of an impact on other types of software companies!