Hi, it's CJ Gustafson and welcome to Looking for Leverage.

Today's term: recurring revenue, and whether a customer who has paid you every month for three years without signing anything actually counts toward it.

I’ve always called that kind of revenue reoccurring. It comes back, sometimes with incredible consistency, but nobody really has to come back.

(I’m a reoccuring customer of dunkin donuts. A very strong reoccuring customer now that I have kids, who have added munchkin expansion revenue to my iced coffee orders of 20 years)

This distinction sounds like SaaS semantics (which would be a great name for an 80’s hair metal band) until you’re using that “ARR” number to feed your hiring plan… and eventually your valuation. Then you care quite a bit about what’s actually inside it.

A while back I had Todd Rakow, the CFO of Gladly, on Run the Numbers to talk about usage-based business models.

Gladly sells customer support software, and its revenue comes in two flavors. There are seats for support agents, which customers pay for whether their agents log in 10 times or 10,000 times. Then there’s telephony, where the bill rises and falls with call volume.

Both show up on the invoice every month, but only the seats are contractually “required to”.

We got into how aggressively a CFO should count the usage piece inside ARR, and Todd gave me a great line:

You can come up with a pretty compelling explanation for why something belongs in ARR. That does not mean the buyer on the other side of the table is going to agree with you.

You already know the textbook version. A customer signs a $120K annual software contract. Congrats, you have $120K of ARR (or $10K in MRR).

Investors like that revenue because, barring churn, it should show up again next year without your sales team having to go win the customer from scratch.

But what if the customer never signs a contract and has historically paid you $10K every single month for the last three years?

Now it gets kinda philosophical.

Dave Kellogg has a great piece on this exact problem called The Mental Mapping from Annual to Monthly and Usage-Based SaaS Metrics.

He uses a guy ordering a $17 martini. A tourist who walks into a bar once obviously isn't MRR. But what about the local who has shown up every Thursday for two years? Dave calls that empirical MRR (and maybe a drinking problem).

Then there’s the guy who just moved into the neighborhood and tells you this is now going to be his Thursday bar. That’s intentional MRR.

And finally there’s the club member paying an annual membership fee with a quarterly food and beverage minimum. That one is contractual MRR.

Dave's argument is that “recur” never actually meant “contractual.” Traditional annual SaaS contracts didn't obligate the customer to renew forever either. We called the revenue recurring because customers tended to come back.

I think that’s right. I also think there’s a point where an operator needs a little more precision than that.

I wrote about this distinction a few years ago in Mostly Metrics in ReCurring vs ReOccurring Revenue: What's the Diff (A Lot).

The way I think about it is pretty simple: recurring revenue is your foundation. Reoccurring revenue can sit on top of it.

The second bucket can still be awesome revenue. You just need to remember nobody signed a piece of paper promising it would be there and therefore it has some additional risk to it in the eyes of any acquirer (or budgeter).

Let's go back to Gladly.

Say you sell a customer $500K of contracted agent seats, and historically that same customer generates another $200K per year of telephony usage.

Can you tell me the customer is worth $700K per year? Sure. They quite literally are.

Can you tell me they're producing $700K of ARR? Ehhhhhhhh.

Once the entire $700K goes into ARR, that number starts making its way around the company. It goes into the quarterly board deck, then the budget, then the sales comp plan, then the lender presentation, and eventually someone sticks a multiple on it when it’s time to sell.

It’s hard to put the ARR toothpaste back in the tube.

Let’s go a bit deeper on a few use cases of this number.

For private companies, ARR is a management-defined, unaudited metric.

And that's fine. There is no ARR police (although I’ll be running for office in 2028).

The important part is defining it, applying the definition consistently, and making it obvious what sits inside the number. AKA footnotes!

The problems usually come from definition drift rather than somebody making one giant aggressive decision. Something like this: You launch a usage product and annualize the latest quarter. Then month-to-month customers start counting after they've been around long enough. Then overages get added at the current run rate. And now the ARR number is massive.

From my experience, nobody sits in a dark conference room with a master plan to juice ARR this year. You just make five individually reasonable decisions and suddenly the metric means something very different than it did 18 months ago.

What I’ve found is you can debate the philosophical definition of recurring revenue until the cows come home. But the diligence team is going to rebuild it from scratch anyway.

They’ll take the contracts and invoices and start splitting revenue into buckets:

Then they’ll look at how each bucket behaves (which is absolutely critical).

If you've been reporting a headline $40M to get them interested in acquiring you, and they rebuild to only $33M, you now get to spend the next few weeks defending your definition and why there shouldn’t be a change in price. And I would wager the purchase price eventually moves closer to their number than yours.

Public markets eventually force the issue too. If a company reports ARR as a KPI, it generally spells out how it calculates it, because the market needs to know that $1 of ARR this quarter means the same thing it meant last quarter.

Say you're reporting $40M of ARR, made up of:

You claim ARR = $40M.

Underneath that, your contracted floor is really $30M. Assume 90% gross retention and you’d expect roughly $27M of that to make it through next year's renewal cycle before you sell anything new.

Now look at the usage piece.

The $7M of usage above minimums turns out to be $5M when you use trailing twelve-month actual spend instead of annualizing Q4.

Add the $3M of month-to-month customers and your reoccurring layer is $8M.

Still good revenue. I'd rather have it than not have it.

But if that layer falls 25% in a softer year (buyer’s always take a conservative approach), $8M becomes $6M. Your downside revenue base is now roughly:

(Oh, and if you built the hiring plan off $40M, you have $7M of fixed-cost support that just went poof.)

Regarding valuation, let’s use completely illustrative multiples. You pitch:

The buyer says the $30M contracted piece deserves 7x, while the $8M reoccurring piece gets 3x:

That’s a $46M difference.

If you have meaningful usage or month-to-month revenue, I wouldn't stop reporting ARR. I'd just make the ingredients super obvious.

Recurring and reoccurring revenue can both be great revenue. But the person buying your company is still going to care which one is which.

Wishing you more contractually committed club members than passionate Thursday regulars,

CJ

Looking for Leverage breaks down one PE term, clause, or mechanic each week, written for the CFOs and finance leaders who actually have to live with these things. If this got forwarded to you, subscribe at lookingforleverage.com. If there's a term you want broken down, reply and tell me. I read everything.

Thanks for reading, and make sure to check out our sponsor, Netsuite.

Readers of Looking for Leverage can try NetSuite Next for free.