CHURN IS DEAD
The Renewal Is Not Yours to Own. The Change Order Is.
11 min read · Strategy
Archive note: This issue predates the evidence ledger introduced in August 2026. Treat uncited benchmarks and examples as editorial analysis, not independently verified findings.
Thirty-one.
That's how many out-of-contract requests one enterprise CSM I know absorbed across eleven accounts in a single quarter. She counted them for me on a call, off a note she'd started keeping in Apple Notes because her manager kept asking why she looked exhausted going into renewals.
The custom Salesforce dashboard a VP of Ops wanted before a board meeting. The ad-hoc training session for a team that never showed up to the paid one. The integration triage when a customer's own engineer broke the webhook. The exec-briefing deck she built from a blank slide at 9pm because the sponsor had a QBR with *his* board the next morning and asked, nicely, if she could "just pull something together."
Every one of those was called "the relationship."
In professional services, every one of those is a priced change order. A documented variation to the statement of work. A line item someone signs.
CS took the "do whatever it takes" ethos from professional services and imported it wholesale. And left behind the only discipline that ever made that ethos survivable.
The 60-Second Version
The hard truth: The unpriced favors that make an account love you are the same favors that make your role economically indefensible and trivially outsourceable.
The framework: The Scope Line separates what the renewal actually pays for from the favor load you silently absorb, and routes every out-of-contract request to product, to priced services, or to a documented no.
This week: Start a favor ledger. Log every request that falls outside the contract for one quarter and total the hours per account. You cannot manage what you have never once counted.
The renewal was never the thing you owned
Here is the fantasy the whole category is built on: that the CSM owns the renewal.
You don't. You never did. The renewal is a procurement event decided by whether your exec champion still has budget and still has a job. It is settled in a finance review you are not invited to, against a pricing model you did not set, in a quarter shaped by a reorg no health score predicted.
I have watched green-scored accounts crater on a sponsor change and red-flagged ones save themselves because a new champion inherited a bigger budget. The renewal outcome lives in a political weather system the CSM influences at the margins.
So when Gainsight repositioned itself at Pulse this year as a retention-as-a-service business, telling the market that technology "is expected to deliver the outcome itself" and that it will "own the outcome, managing renewal motions end-to-end," boards leaned in. Not because Gainsight discovered something. Because CS never made its human labor legible, so the board had no way to tell the difference between a renewal a machine can run and one it can't.
The renewal is not yours to own. But the change order is. And you gave it away for free.
The broken state: unpriced generosity as a business model
Walk into most enterprise CS teams and you will find a workforce that is genuinely, admirably generous. They stay late. They build the deck. They absorb the integration mess. They wear it as a badge.
And not one of them can tell you what a single favor cost.
That's the difference between CS and the professional services orgs it borrowed its ethos from. In PS, uncontrolled scope has a name. It's called scope creep, and it is treated as a margin-destroying failure mode, not a virtue. A delivery lead who let thirty-one unpriced variations onto a project would be in a very uncomfortable meeting. In CS, that same behaviour gets you a shout-out in the team channel.
Benchmarkit's SaaS metrics work has put fully loaded CS cost-to-serve in the range of 10 to 15 percent of the revenue those teams support, and that figure only counts the work the org can see. The favor load is the part that never touches a system. It doesn't show up in the CRM, in the capacity model, or in cost-to-serve. It shows up as burnout, and eventually as a resignation.
Which is not a hypothetical right now. Forty-four percent of CS professionals have been at their current company two years or less, up from eighteen percent a year prior. Experienced people who know exactly what good looks like are walking out mid-reorg. And when they go, the favor load doesn't disappear. It gets inherited, uncounted, by the next person, who also doesn't measure it, until the account has been trained across three CSMs to believe that everything is free and nothing has a price.
Here's what that training does to your economics.
Every unpriced favor teaches the account two things at once. It teaches the customer that value has no cost. And it teaches your board that the value could have been produced by something cheaper than a senior human. You didn't build loyalty. You built a dependency that looks exactly like something a machine could run at a fraction of the price.
That is the trap. The most beloved CSMs are often the ones who have made themselves the easiest to replace, because they turned judgment work into a free utility, and utilities get automated.
What the real ledger showed
Back to the CSM with thirty-one favors.
Here is the honest part the LinkedIn version of this story would skip: after she wrote them down, she didn't march into her boss's office and reform the org. She showed me the note, we talked, and then the quarter kept moving. Nothing changed for a while, because there was no mechanism for anything to change. That's the actual problem. It isn't that CSMs love scope creep. It's that when a request lands, the only available action is to absorb it. There is no "route this" button. There is no change order to raise. Saying no feels like abandoning the customer, and saying yes is free to say and expensive to do.
When we finally tagged the thirty-one requests, the shape of the problem became visible for the first time.
About a third of them were product gaps. Things the platform should have done and didn't, so the human papered over the hole. Those weren't favors. They were unlogged bug reports and feature requests wearing a trench coat, and every hour spent on them was an hour product never learned it needed to fund.
Another chunk were genuine services work. Custom dashboards, bespoke training, integration builds. In a company with a productized services SKU, every one of those is billable. That model exists and it isn't exotic. Plenty of vendors now sell paid onboarding packages and CSM-as-a-service tiers precisely because they figured out that expert human time is a product, not a courtesy. CS teams that give it away are competing against their own company's price list and losing on purpose.
A smaller set were things that should have been a polite, documented no.
And a genuinely small handful were the real thing. The favors worth keeping.
The point of counting was never to eliminate generosity. It was to find out how much of it was actually generosity and how much was just an org quietly running professional services for free through people it pays to manage renewals.
The Scope Line
The fix is not more ownership. It's a line. Here is The Scope Line, and how each part actually runs.
1. In-contract: name what the renewal pays for
Most CSMs cannot articulate, in one sentence, what the contract actually entitles the customer to. Neither can the customer. That ambiguity is where every favor is born, because when nobody has stated the edge, everything feels included.
Write it down. The specific entitlements. The response commitments. The number of reviews, the scope of support, what "onboarding" concretely covers and where it stops. Say it out loud in the kickoff, not defensively, just clearly. Both sides should be able to point at the same edge.
A line everyone agrees on is not a cold line. It's a fair one. The coldness comes from surprises, and surprises only exist where the edge was never drawn.
2. The favor ledger: count for one quarter
You cannot manage what you have never counted. For one quarter, log every request that falls outside that in-contract line. Not the big obvious ones. All of them.
Make the ledger something a VP can screenshot and hand to the team on Monday. Real columns, real tags:
1. Date and account
2. The request in one plain line
3. Hours it actually took, including the context-switching tax
4. Tag: `product-gap` / `priced-services` / `relationship-investment` / `should-be-a-no`
An example row from that real ledger: *Mar 4 | [Enterprise Ops account] | Built custom exec dashboard for board meeting | 6 hrs | priced-services.*
Six hours. Not on any invoice. Now multiply that across a book and a quarter and you have the number your cost-to-serve model has been hiding. That total, the sum of the `priced-services` and `product-gap` hours, is your favor load: the hours of unscoped, unpriced work a CSM absorbs per account per quarter. It is the single most useful number CS leadership does not currently track.
3. Route, don't absorb
The entire game is here. Every favor gets routed, never silently eaten.
- `product-gap` goes to product, logged, with the customer's name attached, so the roadmap finally sees the cost of the hole its own users are being made to plug by hand.
- `priced-services` goes to your services team or a scoped statement of work. This is the change order CS forgot how to raise. It doesn't have to be adversarial. "That's a great project, and it's exactly the kind of thing our services team scopes. Let me get you the right person." That sentence protects margin and treats the request as serious rather than as a chore squeezed between check-ins.
- `should-be-a-no` becomes a documented, gracious no. Not a fight. A boundary, stated once, held consistently.
- `relationship-investment` is the only bucket the CSM keeps. See below.
The routing is the mechanism the CSM with thirty-one favors never had. Give someone a place to send the request and they stop being a landfill for scope.
4. The relationship line: the favors worth keeping
This is the part every scope-discipline conversation gets wrong by omission, so let's not.
Drawing a line does not mean becoming transactional. It means being deliberate about the small set of unpriced moves you keep on purpose, because they build something an agent structurally cannot.
An AI agent can send the reminder, generate the report, deflect the ticket, run the health check. What it cannot do is walk into a boardroom and rebuild a coalition after your champion gets promoted out. It cannot read the room in a tense renewal and know that the CFO's real objection is political, not financial. It cannot give a new sponsor, three weeks into inheriting your account, the honest heads-up before a hard product change so that person looks smart in front of their own boss instead of blindsided.
Those are the relationship-investment favors. The private warning before bad news. The introduction that makes your champion look good internally. The five minutes of candid advice that isn't about your product at all. They are unpriced on purpose, and they are worth every unbilled minute, because they build the human coalition that decides renewals in exactly the political weather no dashboard tracks.
The discipline is not to stop giving. It's to stop giving away the services and the product gaps for free, so you have room left to give the coalition-building that only a human can.
The fixed state is not a nicer CSM. It's a legible one.
The difference the Scope Line makes isn't that your CSMs work less, though they might. It's that the work becomes visible.
When product sees the true volume of gap-patching, the roadmap gets a business case it never had. When services sees the priced-candidate hours, that's a revenue line and a hiring case. When the CFO sees the favor load quantified, the conversation about cost-to-serve stops being a vibe and becomes a number with buckets, and each bucket has an owner who is not the exhausted human eating all of it.
And when the board hears a vendor offer to "own the outcome" of your renewals, you can finally answer the question that has been sitting unanswered under this whole debate. Which accounts are outcome-ownable by an agent, and which need a human who can walk into a boardroom. You can answer it because you have data, not because you have loyalty.
Professional services firms figured this out decades ago, and they didn't do it because they were cold. They did it because they intended to still be in business next quarter.
CS is at the same fork now. The generosity is real and it matters. But generosity without a change order isn't a relationship strategy. It's an unfunded liability with a smiling face, and boards have started noticing that it looks a lot like something they could buy cheaper.
Count your favor load this quarter. Route what isn't yours to carry. Keep the handful that a machine will never be able to fake.
The renewal was never yours to own. But the line, the one that decides whether a human was necessary at all, still is. For now.
By Kuber Sethi · All issues · Subscribe