Your Complete Guide to Quote to Cash: Billing

How much cash is your billing process leaving on the table?

Every extra day between delivering value and sending an invoice is another day you’re waiting to get paid.

And as pricing gets more complex, with usage, tokens, credits, commitments, and custom terms, billing lag compounds fast.

Tabs is built to eliminate that lag by turning contracts and usage into accurate invoices automatically, without the spreadsheets and manual work.

The impact? Faster billing. Faster cash. Less work for finance.

So what is billing lag costing your business?

Use the Tabs Billing Lag Calculator to get your number. 

Gee wiz. I run a recruiting company now. Life comes at ya fast.

We do Strategic Finance and FP&A hiring. And we place at the manager, director, and VP levels. Is it time to find your right hand person?

Before we get into billing, I was searching for "Bill" memes and found an entire sub genre of this stick figure named Bill.

I hope after learning about the Quote to Cash process we can all be more like Bill.

Anywho!

In this three part series we're going through the 8 layers of the Quote to Cash Process:

Billing has changed because the proxy for value has changed.

The unit of work is no longer a human. A pricing that's anchored to a person named Bill does not work anymore.

Ten years ago the dream biz model was you just charge $100 a month and call it a day.

Today every company seems to have seats, usage, credits, overages, platform fees, AI consumption, and AI credits. 

It used to be seats were a really good proxy for value. If you think about a world where you're using your favorite AI software, you could charge one user $20 a month and they could run five prompts in that month and they're doing something simple. But you charge another user the same $20 a month and that user is running hundreds of thousands of prompts and effectively eliminating a workflow a very large team used to do.

From a unit economics perspective, as the seller, that dog don’t hunt.

If you think about the value realized between those two people, it's very different. While they both paid the same price, the company that offered this service incurred a very different cost profile for the users.

Hell, even charging them is now harder. It used to be a simple back office function attached to the ERP.

$100K for life… send invoice!

Software was originally sold on a perpetual licensing basis, so you sent them a one time bill. But when we got into this mode of recurring subscriptions it brought some pricing model complexity with it (hence the rise of companies like Zuora and Chargebee).

Now you start to have more nuanced use cases, like free trials or billing the customer in advance.

The more recent wave is usage based billing, which is attached to the product (and therefore struggles to simply live in the ERP). Real time metering and ingestion needs to be built in. It's a design concern that touches product, finance, accounting, and rev ops.

While subscription has a P (price) and a Q (quantity), and usage also has a P and a Q, getting to the P and the Q is very complicated.

On the surface, metering sounds like oh how hard could it be? Taxi cabs do it. But it's very hard to meter.

First, what are you metering? Is it the number of API calls? Is it a proxy metric? And at what level? Because then you have the pressure of ingesting tons of data.

For AI companies the unit that shook out is tokens, which behave more like a currency than a measure of work. The same token is worth different amounts depending on context, the way a water bottle costs one thing at the grocery store and another inside a World Cup stadium. Output tokens cost more than input tokens because reading your prompt happens in one parallel pass and writing the answer happens one token at a time.

If you're on the buying side, the nice thing is that when everyone coalesces on the same unit you can line vendors up and make them explain the difference (and then negotiate). Public cloud went through this already with the virtual core per minute, then per second. And eventually the services became largely commoditized (we’ll hit that point too if inference reaches a “smart enough for all your tasks” saturation point).

Even once you figure out the metering, the pricing logic is quite hard to reason through.

If you think about a classic subscription, you bill $20 for the first year and then ramp them up to $50 in the second year.

Usage based billing will say you have 100 included credits and then your next 1,000 credits are at $3 and then the next 1,000 after that we'll give you a discount, but first you need to commit to 5,000 credits. And if you ever breach this threshold we'll cut you off because we don't want you to stiff us.

And once you get to the revenue story your finance folks have a pretty meaningful reconciliation challenge. Sometimes there's late arriving usage because the metering pipeline is doing it in arrears… so how do we count it?

This quickly becomes a real time data and pricing problem, where subscription was mostly an invoicing problem.

What does a company need to get right architecturally before it even considers usage based billing?

You can't YOLO it. It takes forethought.

You'll have to meter and you'll have to count something. What you are counting are called product "events".

And not to be Captain Obvious, but you need to ensure the product you built can actually emit those events that you want to count. That means on day zero when you are building your product you need to have your monetization model in mind because your product needs to emit the right events in a way that those events can be tagged, understood, and ID'd.

Then… they can then be metered.

Metering is taking raw events and turning them into usage records.

And then… they can then be rated.

Rating is turning those usage records into dollars.

While you are def responsible for emitting your own events, you can either choose to build or buy a product that does the metering and rating.

(oh, and you need to think about entitlements: who is accessing the product and what do they have access to, which we talked about in part 1)

Finally, then billing turns your rating charges into invoices. Whew!

Finally, finally, vFinal you need to make sure your billing system can flow down into your finance system so you can prove you sold this usage, turned this usage into billable usage, and its actual revenue. And check that no usage leaked somewhere or was stolen.

TL;DR: You need to think about the chain of events: Metering, Rating, Entitlements, Finance and then figure out the portions you want to own in house vs have a third party do for you.

We mentioned that billing was in many ways a module within the ERP and backoffice. Today it feels like it's not only a finance system, but part of the product.

The modules that are leaving the ERP, like billing, are the modules that need flexibility or more product awareness.

To be clear, the ERP is not going to be replaced. But the ERP is getting unbundled because it's a batched system and there are some systems that are very monetization facing that need to be real time. This is because of everything we talked about with usage based billing and entitlements.

The general ledger and the close aren't going anywhere. Those are ERP first functions.

Rev rec has largely moved out already. AR automation is going too, especially now that an agent can handle the invoice number being wrong.

If ERP is the system of record, billing is becoming an operational layer, or an orchestration layer on top, and adjacent.

The functions that need to move at product speed will move to the billing system.

Our dollar's been billed. It hasn't been taxed, collected, or recognized.

(And you do have to pay the tax authorities, so you don't go to jail… they did my man Ja Rule dirty).

Say bye to Bill for now. We’ll cover the rest of the Quote to Cash process next week.

I've been in the lab. And this time I brought a friend.

Me and Ali Hussain of Tabs fame are kicking off a four part series on how AI creates a new economic system inside companies. Whoah!

They say accidents happen at intersections, and we're here to play finance crossing guard as we transition to this brave new world. In this new system...

This is part 1. Rumor has it people were lining the streets like a Jordan shoe drop in anticipation.

We like you a lot though - you are special - so you can skip the queue and download it NOW wherever you get your podcasts.

Wishing you a product that can emit events,

CJ