The strategic thesis
Operating model choice determines more of a capability center's outcome than location, incentives or even initial talent strategy. It sets speed to first delivery, control over IP, cost trajectory and the realistic ceiling on mandate scope.
The market has consolidated around four viable models — direct captive, build-operate-transfer, managed capability and hybrid — with hybrid taking a growing share above the 200-seat threshold.
The decision is rarely permanent. Most enterprises transition models at least once, and the ones that plan the transition contractually at the outset avoid the value leakage that makes transfers painful.
What the data says
Direct captive, BOT, managed capability and hybrid dominate current practice.
Threshold above which hybrid models typically outperform pure options.
Time to first delivery under BOT or managed models.
Time to first delivery under direct captive setup.
Common BOT transfer window written into contracts.
Fastest-growing structure for mandates spanning core and non-core work.
Strategic context
Operating model choice determines time-to-launch, talent quality, governance and long-term value capture. Most enterprises now adopt hybrid models that combine speed-to-market with ownership control.
Build–Operate–Transfer (BOT) remains the dominant entry path for mid-market enterprises; captives dominate above 200 FTEs.
Four operating models — what each optimises for
01 · Direct captive
Maximum control and IP security; slowest to stand up and highest fixed commitment.
02 · Build-operate-transfer
Partner builds and runs, enterprise takes ownership on a defined date and formula.
03 · Managed capability
Partner retains employment and operations; enterprise buys outcomes with scope flexibility.
04 · Hybrid
Core IP work held captive, elastic and specialised scope run through partners.
Operating model comparison for India capability centers
| Model | Time to first delivery | Control | Cost profile | Best fit |
|---|---|---|---|---|
| Direct captive | 12–18 months | Highest | High fixed, lowest at scale | Core IP, long horizon, 500+ seats |
| Build-operate-transfer | 6–9 months | Deferred, then high | Premium then normalising | Enterprises new to India wanting eventual ownership |
| Managed capability | 3–6 months | Contractual | Variable, scope-linked | Uncertain demand, specialised skills |
| Hybrid | 6–9 months | High on core | Blended | Mixed core and elastic scope above 200 seats |
Start from the work, not the structure
The decisive question is what proportion of the mandate touches core intellectual property, regulated data or long-horizon architecture. Work of that kind belongs under direct employment; everything else can legitimately sit in a partner structure.
Enterprises that begin with a structural preference — 'we only do captive' — routinely overpay for elasticity they never use, or delay entry by a year in a market where competitors have already secured the senior talent.
Transfer terms decide whether BOT works
A BOT arrangement is only as good as its transfer clause. The terms that matter are the valuation formula, employee transfer mechanics, IP assignment at each stage, and what happens to the partner's platform tooling on transfer day.
Where transfer economics are left open, enterprises discover late that the effective cost of ownership exceeds a direct build. Negotiate the exit before signing the entry.
Fixed method agreed upfront, not a market-value negotiation at transfer.
Named-role continuity commitments with retention incentives funded in advance.
Continuous assignment from day one rather than a single event at transfer.
Where each model actually costs less
Direct captive has the lowest unit cost at scale but carries a high fixed base and a slow ramp; below roughly 200 seats that fixed base rarely amortises well.
Managed models cost more per unit but convert fixed cost to variable, which is worth paying for when demand is uncertain or the skill is scarce enough that a captive would be unable to hire it.
Plan the second model while implementing the first
Most centers change structure once within five years. Building transition provisions into the initial contract — data portability, tooling licences, employee continuity — turns a disruptive event into an administrative one.
The reverse case matters too: captives that need to shed elastic scope find it far easier when partner frameworks were pre-negotiated rather than sourced under pressure.
What to do now
- →Classify the mandate by IP and regulatory exposure before evaluating structures.
- →Negotiate BOT transfer valuation, employee continuity and IP assignment before signing.
- →Use managed capability for scarce skills and uncertain demand rather than for cost alone.
- →Assume at least one structural transition and build portability into every contract.
- →Revisit the model at each doubling of scale, not on a fixed annual cycle.
The decade ahead
Hybrid structures will keep gaining share as enterprises separate IP-critical work from elastic capacity rather than choosing one model for the whole center.
BOT contracts will standardise around clearer transfer formulas as more enterprises complete a full cycle and price the exit correctly.
What matters most
- 1Model choice follows the work profile, not enterprise preference.
- 2Hybrid dominates above roughly 200 seats.
- 3BOT value is decided by transfer terms agreed at entry.
- 4Portability provisions make the inevitable transition cheap.
Frequently asked
Which operating model is fastest to first delivery?+
Managed capability at three to six months, followed by BOT at six to nine; direct captive typically takes 12 to 18 months.
When does a direct captive make economic sense?+
Above roughly 500 seats with core IP work and a long horizon, where the low unit cost at scale outweighs the high fixed base.
What is the main risk in a BOT contract?+
An open transfer valuation, which can make eventual ownership more expensive than building directly from the start.
Why are hybrid models growing?+
They let enterprises hold IP-critical work under direct employment while keeping elastic and specialised scope variable.
How often do centers change model?+
Most change structure at least once within five years, usually at a scale inflection point.