Acme Corp's most senior CSM was, on paper, the team's best performer: highest renewal rate, cleanest QBR decks, the health score dashboard everyone else's numbers got compared against. What that title hid was that she was spending roughly half of every week rebuilding the health score model, fixing broken CRM fields, and writing the reporting queries the rest of the team quietly relied on, work nobody had asked her to own, work that had simply drifted to whoever was competent enough to do it without complaint.
When she gave two weeks' notice, the manager discovered the actual size of the problem: no one else on the team knew how the health score was built, three dashboards had silently stopped updating months earlier, and the "best CSM" title had been quietly subsidizing a second, unstaffed job the whole time.
The two jobs hiding inside one CSM title
"Customer success manager" bundles two genuinely different kinds of work under one job description. The first is relationship-facing: renewal conversations, expansion pitches, escalation handling, the judgment calls that come from knowing an account's history and the people inside it. The second is operations-facing: building and maintaining the health score, keeping CRM fields clean, writing the reporting that shows leadership what's actually happening across the book, configuring whatever tooling the team runs on.
At a small team, one person doing both is not a design flaw, it's the only option, and it works fine as long as the operations half stays small enough to fit around the relationship work rather than compete with it. The trouble starts when the operations half stops being occasional maintenance and becomes a standing, second job that nobody scheduled time for.
Why it holds together at five and breaks at twelve
The operations workload doesn't scale linearly with headcount, it scales with the number of systems and the number of people relying on the outputs of those systems. A five-person team usually runs on a lean tech stack, one spreadsheet-based health score, and one person who happens to be good at both halves of the job. A twelve-person team is running the same core systems under load from three times as many accounts, three times as many CSMs asking for report changes, and a leadership team that now expects dashboards to be reliable rather than best-effort.
That's the point where the informal arrangement, whoever's most comfortable in spreadsheets does the ops work on the side, stops being sustainable, because the ops work has grown into a role-sized job that's still being absorbed as an unpaid side project by someone whose actual title and comp plan is built entirely around account coverage, not systems maintenance.
Three mistakes teams make when they finally split the role
Once a team recognizes the split is overdue, the failure mode usually isn't inaction, it's doing the split badly and creating a second, differently broken arrangement.
Splitting along account tier instead of work type
The instinctive first cut is to make one person "senior CSM for our biggest accounts" and call that the split, but that just creates two relationship-facing roles with an unequal workload, not a relationship role and an operations role. The health score, the reporting, and the tooling still have no dedicated owner, they've just been quietly assigned to whichever half got the smaller accounts.
Promoting the best CSM into ops, not the person best suited to it
The CSM who's been informally doing the ops work is often the obvious internal candidate, and sometimes that's the right call, but not automatically. Being good enough at spreadsheets to keep a health score alive under pressure is a different skill from wanting to own systems, data hygiene, and process design as a full-time job. Moving your best relationship-builder into a pure ops seat because she happened to be the one holding it together can cost the team more in lost account coverage than the split was supposed to save.
Splitting without changing who reports on what
The split only works if the reporting lines and the KPIs split with it. If the new ops hire still gets evaluated on the same account-health metrics as the CSMs, or if CSMs are still expected to know the internal mechanics of a health score they no longer maintain, the org chart changed but the actual accountability didn't, and the same person ends up doing both jobs again within a quarter.
"RetainSure helped Mailmodo's CS team crack upsell at scale. By zeroing in on high-potential self-serve accounts and providing personalised email drafts, the team saw a 20x ROI from their very first month on the platform."
Sanjana Shankar, Head of Customer Success · Mailmodo
What the split actually looks like in the first thirty days
A clean split starts with writing down, in one document, every recurring task the informal ops work currently covers: which reports get sent weekly, who maintains the health score formula, who owns CRM field hygiene, who fields tooling questions from the rest of the team. That inventory alone usually reveals two or three tasks nobody remembers deciding to own, the kind of drift that made the original CSM's unofficial second job invisible in the first place.
From there, the CS Ops hire takes explicit, documented ownership of that list, with their own goals tied to system reliability and reporting accuracy rather than account retention, resolving the same automate-versus-hire question the team had been putting off, and the remaining CSMs get their full account time back. The transition period matters more than the org chart: for the first month, the outgoing "informal ops owner" should be reviewing the new hire's changes before they ship, not just handing over a folder of spreadsheets and disappearing back into a full account load.
Split the role without splitting your reporting stack into two versions of the truth.
RetainSure keeps the health score, tooling, and reporting layer consistent no matter who's staffed to maintain it.
How to tell if your team is ready to split
The clearest diagnostic doesn't require a formal audit. Ask each CSM to estimate, honestly, how many hours in a typical week go to something that isn't a customer conversation, an account plan, or direct prep for one. If that number is climbing past a day a week for more than one person on the team, the operations half of the job has already grown past what a generalist role can absorb, whether or not anyone's given it a name yet.
The second signal is who gets asked when a dashboard looks wrong. If there's already an informal answer, one person everyone pings first, that person is already doing the CS Ops job without the title, the comp plan, or the protected time that would let them do it well. The split, at that point, isn't creating a new role. It's just finally paying for the one that already exists.
Six weeks after the promotion, Acme Corp's new CS Ops hire had rebuilt the health score from a shared, documented formula instead of one person's private spreadsheet, and every CSM on the team could see exactly why an account's score had moved without asking anyone. The relationship work hadn't gotten easier. It had just, for the first time, stopped competing with a second job nobody had agreed to give it.