On April 9th, I moderated “Buy, Sell, Hold: What Survives the AI Era?” at the Abacum Summit.

We tackled a simple question: What parts of finance are actually worth keeping — and what needs to go?

Alongside Chris Brubaker (Postscript), Fraser Hopper (PostHog) and Tania Secor (Stamford Health), we turned it into a live, audience-voted debate on what modern finance teams should buy into, sell off or hold on to as AI reshapes the function.

If you couldn’t make it to Gotham Hall, the full session is now available on demand. 

I’d like to buy you a beer if you live in the Bay Area on Thursday May, 14th.

Come hang with fellow Mostly Metrics readers.

I’ll be telling my favorite renewal rate jokes and I’ve asked the bar to play an eclectic mix of 80’s Joel and deep cuts from Creed’s second album.

Can’t wait to meet you IRL.

I’m convinced there’s no show on television that spans a wider age demographic than Netflix’s “Is it Cake?”. Both my 98 year old grandmother and my 4 year old daughter love it.

For those of you who haven’t seen it, master bakers compete to recreate household items in cake form and fool a panel of C list celebs (think: Ray J, but a rung lower). The dramatic event is when the host attempts to cut through whatever the panelists voted was the actual cake. Either frosting oozes out, or he’s awkwardly sawing a cowboy boot. It’s truly must-see tv.

We should do a spinoff. And call it Is it Revenue?

This post made the rounds in the startup community recently:

Scott is the co-founder and CEO of Spellbook, which makes AI for lawyers. He is (it appears) calling out a larger competitor, maybe Harvey or Legora, for counting their revenue creatively. The setup he describes is a multi year deal.

If true, this isn't new. Hell, I've personally contributed to the cARR epidemic in years past. Boy, did we look GREAT on paper, taking immediate credit for the exit values of five-year, heavily ramped contracts, riding a COVID seat-based SaaS explosion with net dollar retention rates that made your nose bleed.

And I had zero idea I was doing anything wrong.

Up until I was on an audit committee call with a famous CFO of a publicly traded company, where he eviscerated my revenue forecast. The gap (no pun intended) between what I was calling ARR and GAAP revenue had widened to nearly 40%, too gargantuan of a spread for even world class growth to account for. Woops.

Which brings me to something most finance people won’t want to hear: It’s an understandable mistake for many of these companies to make.

Why?

Remember - I was an FP&A Director with six years in-house, plus stints at PwC and a private equity firm, and I belly-flopped cARR vs. ARR badly enough to torpedo the forecast at a company approaching $100M.

So I have empathy for the technical founders building AI tools who don't yet know the difference between bookings and billings. They are engineering drop outs from Stanford, not CPAs.

Gary Tan, YC head honcho, chimed in on X with a guide. And got dunked on by a lot of people (which I think is the wrong move).

Remember- he’s not writing for the SEC accounting team at Datadog or CrowdStrike. He's writing for founders raising their first institutional check, who aren't going through a formal audit, who are just trying to prove enough people like the product to justify more capital.

This inspired me to come up with my own list. And before you try to dunk on me, like Gary, I’m doing this for early stage teams who are trying to get to their first few million in revenue. This is like a pre series B-ish stack ranking.

I have a confession to make: I’m probably responsible for 2,000 of these views. It’s gotten me through some tough times in life.

Also, if you take life advice from either of these two guys, you should get your head checked. Also, I’ve got some Florida swamp land I’d like to sell you.

Ok, let’s say you’re building something super hardware intensive. It’s hard to sell something you haven’t built yet. And you can’t build it yet because you don’t have capital. The next best thing you can do is collect indications of interest on paper.

I heard this story about Blake Scholl, founder and CEO of Boom Supersonic, the company trying to build commercial aircraft that can fly from NYC to London in two hours. I might be butchering the details (I listen to a lot of podcasts), but the gist is: hours before stepping on stage at a YC demo day, Blake got a non-binding letter of intent from Richard Branson's Virgin Galactic saying they'd be interested in buying a few planes if he could actually build them.

Is that as good as a signed contract with prepayment? Absolutely not. Branson can walk away tomorrow and owe Boom nothing. But it's a serious name attaching itself to a hard idea, in writing, before the thing exists. For a founder trying to prove commercial traction on something that takes a decade to build, that's worth a lot.

You could say: We have $100K in non-binding LOIs from customers X, Y, and Z.

You could not say: We have $100K in revenue.

Is it cake? No. Is it revenue? No. Is it the right kind of evidence at this stage? Yeah, actually.

GMV is the big number flowing through your marketplace.

You get to keep what is left over, traditionally called your take, or your rake.

In this case, you sell $20 worth of burritos and deliver them on a bicycle, but charge the customer only $12.

Your rake is negative $8. This does not work at scale. You are destroying value to onboard customers, betting that you can fix the unit economics later through some combination of price increases, scale, and squeezing the supply side.

BTW: Amazon, Uber, and Doordash all started off as negative gross margin GMV businesses.

You could say: We're processing $1M of GMV per month at a negative 40% take rate, and we have a path to positive contribution margin by doing xyz by date 123.

You could not say: We're a $12M Revenue business.

Not all marketplaces figure it out.

Same burrito, extra guac. You sell $20 worth of food, deliver it for a $24, plus a $5 service fee, and clear $9 in net revenue. That's a 45% effective take rate.

This is a real business. The marketplace is generating cash on every transaction, the supply side is getting paid, and the rake is real net revenue.

However, GMV still isn't your “revenue”. Your revenue is the take. Saying you're a "$100M Revenue company" when you keep $20M would be dishonest and misleading.

A booking is a signed contract. The customer has committed, in writing, to pay you some amount over some period. This is real. It's a step beyond an LOI, beyond GMV, beyond every level above this. Someone has agreed to give you money.

Bookings reflect the demand for your products or services and the sales team's ability to secure new business.

However, Bookings do not necessarily translate to immediate cash flow or revenue recognition. There are often timing differences between when a contract is signed and when the actual services are delivered or goods are shipped.

The math gets exposed eventually…. There will come a quarter when your investors look at your reported GAAP revenue, then look at your cARR, and scratch their heads at the chasm between the two. If you're growing fast and stacking multi-year deals, that gap only gets wider. You become a victim of your own creative accounting. And jumped the shark (shirt).

Imagine being at $102M in "ARR" and then dropping back down to $86M purely because of how you were adding it up? That would suck.

This is real revenue. Nice.

You sold something, the customer paid, the cash hit your account. Nobody's contesting the dollars. Good work.

The mistake is what you call it. We’ve all seen hustlers on insta calling their creatine gummy revenue ARR.

ARR stands for Annual Recurring Revenue. The recurring part is so key, because it implies that you have paid once to secure a contractually recurring revenue stream that we can expect to renew at the same time and at (least) the same price next year.

This mix up shows up most often in three places:

Polymarket can annualize their monthly revenue and call it an annualized run rate. But what if it’s an election cycle or the Superbowl? Would you feel comfortable putting that revenue on the same level as a contract with a business with a renewal date?

This is the prize.

A customer signed a contract. They've started paying. They will keep paying on a regular and predetermined cadence (monthly, quarterly, annually) and unless something goes wrong, they'll renew when the term is up. The revenue is recurring not because you hope it will or because you have solid empirical evidence that they purchase on a regular basis (this guy always buys a 12 pack of Miller Lites on Fridays!), but because the contract says it will and the customer keeps writing the check.

This is the only level where you've earned the right to use ARR without a prefix, a footnote, or a creative interpretation. It's legit ARR.

You could say: We're at $50K MRR, or $600K in ARR, with 95% gross retention over the last twelve months.

You could say it on a slide. Nice. You could say it to investors. Wahoo. You could say it on a podcast. Like mine!

And you could say it to the publicly traded CFO on the audit call without sweating through your stupid ass shark shirt.

Is it cake?

No. It's revenue.

(And you can buy a lot of cake with real revenue).

And why stop at revenue? If you want, you can take this canonical Lead Edge memo and trace it all the way to cash profits.

TL;DR: Median Multiples are UP week over week.

The overall tech median is 3.4x (UP 0.2x w/w).

What Great Looks Like - Top 10 Medians:

Wishing you trade at a high revenue and EBITDA multiple,

CJ