Choosing the operating model before you choose the headcount
Key takeaways
- The operating model, captive, BOT, GCC-as-a-service, or managed, is decided before the org chart, whether you choose it or not, and it sets which roles you can credibly hire.
- A headcount can be signed off and still be unhireable in the model you picked: fundable is not the same as fillable, and a plan can be perfectly budgeted and quietly impossible.
- Skip the operating-model call and the market makes it for you, usually as a backbone shop that can't hire the judgment it later needs.
The operating model is the first decision, and the headcount plan is the second.
Most GCC build plans that reach us open with a number, roles, budget, ramp curve, and only implicitly answer the question that actually governs all of it: what kind of centre is this, and who owns it while it stands up? A captive that owns a product line, a build-operate-transfer arrangement that hands over in 18 months, a GCC-as-a-service partner running the centre on your behalf, and a managed backbone that executes volume at cost are four different organisations. They read as similar on a headcount sheet and behave nothing alike in the market for people.
The operating model sets the ceiling on the roles you can hire into. A managed backbone can staff throughput seats quickly; it struggles to attract the senior owner who wants a charter rather than a queue. A captive with real product ownership can pull that owner, but only if the plan named the model early enough for the pitch to be true. A BOT sits in between: candidates weigh who they're joining now against who they'll report to after transfer, and the offer has to answer both. Choose the headcount before the model and you end up recruiting against a story the centre can't yet support.
So the order matters, and it's the reverse of how most plans are built. The model decides who the centre can plausibly attract; the roster is downstream of that. When a plan arrives roster-first, our earliest useful contribution is often to send it back one step and ask what was assumed about ownership before anyone counted seats.
Each model implies a different first hire.
Walk the four models against the same opening role and they diverge immediately. A captive that intends to own a product wants an early leader who can hold a charter and grow a team under it: the pitch is autonomy and a P&L-adjacent mandate, and the seat only lands if that mandate is real on day one. A managed or backbone centre wants a delivery lead who can stand up process and scale reliable execution, a genuinely different person, recruited on a genuinely different promise.
BOT and GCC-as-a-service complicate the first hire further, because the person you attract has to be legible to two organisations at once. In a BOT, the leader you recruit may be technically employed by the build partner while being sold on the enterprise they'll transfer into; get that framing wrong and you either over-index on the partner and lose the ambitious candidate, or over-promise the parent and set up a bad handover. GCC-as-a-service shifts more of the early roster onto the provider, which changes who you hire and who does the hiring, and how much of the centre's identity is yours to offer at all.
The practical error is treating the first job description as model-agnostic. A “Head of Engineering” req reads the same across all four, and means four different jobs underneath. The candidate feels the difference in the first conversation, long before you do, and self-selects out of the ones where the story and the structure don't match.
Fundable is not the same as fillable.
Once the model is chosen, a second gap opens that plans rarely test: a headcount can be signed off and still be unhireable. A funded plan feels like a solved problem: the roles are approved, the budget is signed, the ramp is scheduled. But funding only answers whether the parent will pay for the seat. It says nothing about whether the person to fill it exists, at that price, in the model you chose, on the timeline you set. We regularly see plans that are immaculate on the budget line and quietly impossible in the market, because the seat was designed for a captive charter and funded at a backbone band, or scoped for an owner the chosen structure can't credibly offer ownership to.
The failure shows up as slipped ramps and open reqs that stay open, and it's usually diagnosed as a recruiting problem when it was a planning problem. The seat was never fillable as specified. No amount of sourcing effort closes the gap between an approved plan and a market that won't supply it on those terms; it just burns quarters proving the plan was wrong. Fundable is a conversation with your parent about appetite. Fillable is a conversation with the market about availability and price. The first doesn't settle the second.
A number can clear the board and still be impossible in the room where you have to fill it. Fundable is the parent's appetite; fillable is what the market will actually give you at that price, in that model.
Kalaiselvi Ponnurangam · Practice Head – Talent Consulting & Advisory · Recruise
| Operating model | Who owns the centre while it stands up | Roles it can credibly hire early | Where fundable-but-unfillable bites |
|---|---|---|---|
| Captive | The enterprise, directly, from day one | Charter-holding leaders, product owners, senior judgment seats | Owner roles funded but scoped without real mandate; the charter isn't there to sell |
| BOT (build-operate-transfer) | A build partner now; the enterprise after transfer | Leaders legible to both entities; delivery plus a transfer story | The senior hire who won't join a partner-badged role for an enterprise promise that isn't underwritten |
| GCC-as-a-service | A provider running the centre on the enterprise's behalf | Provider-sourced delivery and platform roles at pace | Roles the enterprise wants to own but has handed the provider the identity to offer |
| Managed / backbone | A managed-services operator, at volume and cost | Delivery leads, process owners, throughput seats | The judgment hire funded late, after the org is already shaped for volume |
Run both tests before the plan is signed.
The correction is to run the model check and the fillability check together, at plan time, not after the reqs are open. Beside every funded role, put two questions. Does the operating model we chose actually let us offer what this role needs to be filled: the ownership, the reporting line, the story a strong candidate will believe? And does the funded band match what that capability clears at today, in this location, for this kind of centre? For the deep, plentiful roles the answers are yes and the check is quick. For the scarce, senior, charter-dependent seats, the ones the mandate usually rests on, the check is where the plan gets honest, and where a role gets rescoped or a band gets adjusted before it's committed rather than defended after it fails.
This discipline is small and its payoff is large. It also intersects with geography, because the same role can be fundable and fillable in one talent market and impossible in another; that's a decision in its own right, and we treat it separately in why location is a talent-supply decision. Bring the model, the location, availability, and comp into the planning room next to headcount, and the plan you approve is one you can actually staff. Leave them out and you've approved a spreadsheet that describes a workforce you can't build.
The default, when unchosen, is the one that's hardest to leave.
When the model is left unstated, centres drift toward the managed backbone. It's the fastest to stand up and the easiest to fund, because the work is legible and the roles are familiar. The plan simply defaults into a backbone because that was the shape that funded cleanly. The trouble arrives a year or two later, when the parent wants the centre to own judgment rather than execute instructions. By then the org is shaped for volume, the senior seats were never created, and the people who could deepen the charter have no reason to join a place that reads as a queue.
The firms that avoid this trap treat the operating-model call as a leadership decision, and they revisit it deliberately when the mandate changes. Naming the model early doesn't lock the centre into one shape forever: a BOT is designed to become a captive, and many managed centres do graduate to ownership. It just makes sure the first hires are consistent with the shape you're trying to grow into, so that when the mandate deepens you aren't starting the senior recruitment from a standstill against your own history.
Frequently Asked Questions
Why choose the operating model before the roles?
Because the model sets the ceiling on who the centre can credibly attract. A captive with real product ownership can pull a charter-holding leader; a managed backbone staffs throughput seats quickly but struggles to win that same senior owner. The roles you can plausibly fill are downstream of the operating-model call, so deciding headcount first means recruiting against a story the centre can't yet support. Choose captive, BOT, GCC-as-a-service, or managed first, then write the roster the model can actually deliver.
What does “fundable is not the same as fillable” mean?
Fundable means the parent has approved and budgeted the seat. Fillable means the person to fill it exists, at that price, in the model and location you chose, on your timeline. A headcount can be signed off and still be unhireable, a plan can be perfectly budgeted and quietly impossible, usually because a role scoped for a captive charter was funded at a backbone band, or scoped for an owner the chosen structure can't offer ownership to. Test both at plan time, not after the reqs stay open.
How does the first hire differ by operating model?
A captive wants an early leader who can hold a charter, sold on a real mandate. A managed or backbone centre wants a delivery lead who can stand up process and scale execution, a different person on a different promise. A BOT needs a leader legible to both the build partner and the enterprise they'll transfer into. GCC-as-a-service shifts much of the early roster onto the provider, changing who you hire and who does the hiring. A single “Head of Engineering” req means four different jobs underneath.
What happens if we don't choose an operating model at all?
The market chooses for you, and it defaults to the managed backbone because that's the shape that funds most cleanly. A year or two later, when the parent wants the centre to own judgment, the org is already built for volume, the senior seats were never created, and the leaders who could deepen the charter have no reason to join. It's the hardest default to leave. Make the operating-model call a deliberate leadership decision and revisit it when the mandate changes.
One hiring pattern worth knowing, every ten days.
The Mandate Desk is our read on the senior GCC talent market — one signal that moved, the read behind it, and one thing worth doing. Written from live placement data.
More from Recruise Insights.
Workforce Management Most contract staffing decisions are made on the wrong number
Bill rate against salary prices two different instruments as though they were one. The question that decides it is whether the work has a real end date.
Workforce Management Most hiring compliance failures are process failures, not legal ones
The policy is usually right. What breaks is consistency across interviewers, a record nobody owns, and the exception granted verbally under time pressure.
Workforce Management The ninety-day notice period is a planning input, not an inconvenience
US hiring plans are built on a 2-week notice assumption that doesn’t hold for senior India hires. Most of the damage happens after the offer is signed.
Have a senior seat to fill?
Tell us the mandate — the role, the level, the market. We’ll come back with what the market is really doing on it, and how we’d run the search.