Cost to Serve: The Line Item You’d Rather Ghost Than Model

You ever watch someone calculate margin while actively lying to themselves?

It’s a spiritual experience.
There’s this faint look of serenity—like a monk in a burning building—while they insist that yes, this customer is profitable, and no, of course the CSM isn’t secretly doing onboarding, support, and three calls a week of emotional labor because “they’re strategic.”

Let me save you some time.
You’re not profitable.
Not really.

You’re underpriced, over-serving, and modeling your customer base like it’s a row of neat little dollar signs instead of what it actually is:
A swamp of operational complexity in a trench coat.

And the whole illusion lives and dies in one place:
Cost to Serve.

The number no one models.
Because it doesn’t look good in a board deck.
Because it doesn’t respect department boundaries.
Because it has the gall to be emotionally messy.

It Starts With ARR. And Delusion.

ARR is such a beautiful, stupid number.
It floats. It shines. It lets founders whisper things like “predictable growth” while their implementation team eats lunch at 4pm because the latest customer “just needed a quick call.”

You fall in love with revenue the way teenagers fall in love with summer—loud, blind, and absolutely convinced that this time it’s different.

But it never is.

Because you didn’t sell software.
You sold support.
Time. Bodies. Edge cases. Jira tickets. Explainers. Renewals that require wine and calendar gymnastics.

And you didn’t plan for any of it.

The Absurd Quiet Around It

Let’s walk through the org chart, shall we?

CS doesn’t own the full number.
Product pretends it’s GTM’s fault.
Finance can’t track it.
Ops doesn’t want to be rude.
And the CEO—well, they’re still bragging about “land and expand” like it’s a free dessert, not a second mortgage.

So Cost to Serve becomes this ghost.
You feel it. You dodge it. You definitely pay for it.
But you never look it in the eye.

Instead, you write “High-margin segment” in the plan.
Then pay six people to hold that margin together with Slack threads and personal favors.

Why No One Builds It In

Because building it in means admitting how weird it all is.

It means modeling that 20% of engineering time is not product work—it’s customer maintenance theater.
It means mapping how support load scales non-linearly with your most “valuable” accounts.
It means asking your Head of RevOps what “enterprise” actually means—and watching their soul leave their body halfway through the answer.

There’s no spreadsheet function for denial.
But most FP&A models try.

The Heroic Workaround Economy

Let me guess.
You’ve got “lean” GTM.
“Efficient” growth.
“Tech-touch” motion.

And also:

A Slack channel called #urgent-customer-needs
An engineer named Dave who hasn’t done roadmap work since Q1
A renewal pipeline built on good vibes and Google Docs

This is not a business model.
It’s improv comedy.
Except no one’s laughing and someone just made up a new pricing plan live on a demo call.

You call it nimble.
Investors call it sticky.
Your team calls it Friday night.

And none of it’s in the forecast.

What It Actually Costs

Let’s get real:

It costs time. Quietly.
It costs credibility, when your “one-hour onboarding” becomes a three-week ordeal.
It costs margin, morale, and runway.
And it always costs more than it looks like.

But your model shows improvement.
Because your CAC’s going down.
Because expansion’s up.
Because your churn rate is “decent.”

Meanwhile, one customer costs more to support than three engineers’ salaries, but you haven’t seen it yet because the headcount plan is organized by team, not trauma.

The Model You Swore Was “Solid”

You want to scream, right?

Because technically, everything’s in there.
Sales targets. Hiring plan. Ramp time. CAC. Payback. Churn.

It’s tight.
It’s smart.
It’s the Mona Lisa of misdirection.

But it doesn’t include:

Manual QBRs
Support-led product QA
CFOs playing payments helpdesk
Renewals that require three execs and a prayer

You’re playing jazz with a calculator.
And calling it strategy.

Why Founders Can’t See It

Because you don’t want to.

Cost to Serve is the opposite of founder energy.
It’s slow. Humbling. Specific.
It doesn’t scale well. It doesn’t present well.
And it doesn’t give you a chart to send to Andreessen.

But it’s what determines whether your 10M in ARR is a business…
or just a glorified group project held together by legacy API calls and good intentions.

Every company hits this wall.
The trick is noticing before your VP of CS leaves, your product team revolts, and your cash turns weirdly squishy around month nine.

The Only Real Fix

You don’t outsource this.
You don’t OKR it.
You sit in it.

You map time per customer.
You trace the weird workflows.
You model support as a function of complexity, not headcount.

You tell the truth about what it takes to keep revenue alive.
Then you charge for it.
Or you walk away.

Either way—you finally see it.
Not as overhead.
Not as noise.

As a line item that holds the entire company together when the forecast stops pretending.

We’re not raising this quarter. Just charging double for “low-touch” accounts.

DM me if you’re ready to put Cost to Serve in the model—before it puts you in therapy.