Build-Operate-Transfer is sold as a way to get an offshore capability without the risk of setting one up. That is roughly true, and it hides where the actual risk sits — which is not in building or operating, but in whether the transfer was ever really going to happen.
The clause that decides everything
In genuine BOT, three things are fixed in the first contract: the transfer date, the transfer price, and the mechanism.
Where any of those is left to be negotiated later, you do not have BOT. You have outsourcing with an option you will price at the moment you have least leverage — because by then the team knows your systems, and walking away costs you everything they know.
Ask to see that clause before signing, not at the point you want to use it. A supplier who is comfortable fixing it is telling you something; one who wants it left open is telling you something too.
Build (six to twelve months)
The vendor recruits, onboards and stands up the unit under its own entity, licences and infrastructure. You are buying speed here: hiring in a market you do not operate in, without a local legal presence, is the part that would otherwise take a year.
Two things to insist on during this phase:
- You interview. If you cannot meet the people who will eventually be your employees, you are accepting whoever is available rather than whoever fits.
- The structure anticipates the transfer. Systems, documentation and access should be organised from the start so the handover is an assignment rather than a rebuild. Retrofitting that at the end is where BOT programmes lose months.
Operate (two to three years)
The unit delivers against your roadmap while the vendor carries payroll, compliance, facilities and attrition. This is the phase where the value is realised — and where the transfer either becomes achievable or quietly does not.
What separates the two is whether readiness is measured. Transfer readiness reviews against the criteria agreed at signature, run periodically through the phase, turn the handover into a tracked state rather than a judgement call in the final quarter.
The criteria worth agreeing:
- Documentation sufficient for someone outside the team to operate the systems
- No single point of failure — at least two people can carry any critical area
- Local leadership identified and already doing the job, not appointed at transfer
- Statutory and payroll processes documented and reproducible
Transfer
Incorporation of your entity, then assignment: contracts, assets, systems, and offers to the people.
The point most programmes underestimate is that employment does not move automatically. Individuals are offered roles in the new entity and accept or decline. Retention through the transfer is therefore a design problem — how the conversation is timed, what the new entity offers, whether people see it as a step up or as being handed over.
Getting this wrong is the most expensive way to fail at BOT: you complete the legal transfer and lose the knowledge it was for.
Or do not transfer
The option should stay an option. Circumstances change: the roadmap that justified a captive gets cancelled, the parent is acquired, the strategy shifts.
A good BOT agreement lets you extend the operate phase or walk away, at a price agreed in advance. What you are buying is optionality, and optionality that can only be exercised one way is not optionality.
When BOT is the wrong model
- The capability might be temporary. An entity is expensive to unwind — statutory closure in India is slower than incorporation. Start with an ODC; the transfer option can be added later.
- Below roughly twenty engineers. The fixed cost of an entity does not amortise, and a dedicated pod or ODC gives the same delivery for less overhead.
- You do not want to run an Indian company. That is what you are signing up for: payroll, statutory filings, an office, local employment law. Some organisations genuinely do not want it, and an indefinite ODC is a perfectly respectable answer.
The question to ask first
Not "how does BOT work" but "do we want to own this capability in five years?"
If yes, BOT has a destination the other models do not. If you are unsure, the honest sequence is an ODC now and the decision later — because moving up the commitment axis is straightforward, and moving back down is not.