How to have a 95% Gross Revenue Retention Selling SaaS to Small Businesses

In over 15 years of financing growing, private, companies, we have lent to folks selling to all sizes of customers, from mid-six-figure contracts with the very largest of enterprises, down to the four-figure annual subscriptions to the smallest of SMBs. We are extremely attuned to revenue retention – but we are also realistic about what is “good” depending on your customer profile.

For example, selling to enterprises like money center banks, insurance giants, or auto OEMs would carry an expectation of gross revenue retention (GRR; sometimes gross dollar retention or GDR) in the 90%+ range. Those institutions have long, considered sales cycles, lots of staying power, and are slow to change technology stacks.

Conversely, we might have a 75-80% GRR to be quite strong if you are selling to small businesses, which are more fickle, can change direction at an owner’s whim, and which frankly go out of business themselves. In other words: it’s no more required to have a 95% GRR than it is damning to have an 80% GRR, but it’s very context dependent.

That’s why I was puzzled to read the ServiceTitan IPO documentation and discover that they claim 95% GRR while selling software to the building trades, a sector well known for having lots and lots of small businesses and relatively high churn overall:

This figure is an almost unbelievably high GRR rate. It piqued my curiosity and drew my attention to the footnotes. Those footnotes direct the reader to dig into page 88 of their filing for more details. I’m quoting below from the ServiceTitan S-1 filing:

“To calculate our gross dollar retention rate as of a given quarter, we first calculate prior period annualized billings. We then identify the value of annualized billings from any customers whose billings were zero in the current period (excluding the impact of one-time credits), which we refer to as churn. We then divide (a) the prior period annualized billings minus churn by (b) the prior period annualized billings to calculate the gross dollar retention rate.”

Let’s piece this apart:

• First, we can effectively remove the word “annualized” in this methodology, because they are dividing “annualized” by “annualized.” The result therefore is quarterly, not annual in any way.
• Second, the figure quoted is not quarter-over-prior-quarter. If you read carefully, the figure is prior-quarter-over-prior-quarter, but excluding from the numerator those customers who completely ($0 billings) churn out in the next quarter. This alone isn’t problematic, as it appropriately excludes upsell from the gross retention figure, except that …
• Third, this methodology excludes downsells / downgrades, other than complete ($0) churns, from being deducted from GRR. This is also not entirely unheard of, but it’s a one-way ratchet that only increases GRR, and never decreases it.

It is worth noting that if the quarterly GRR is 95%, a naive compounding of it would result in 81.5% annualized GRR, which is … an entirely believable and altogether fine GRR figure for a company selling to the building trades.

It is left as an exercise to the reader to determine why top-left lead underwriter Goldman Sachs chose the non-standard methodology that yields the higher and markedly less plausible 95% figure.