Buyers usually begin with a rate comparison and end up surprised, in one direction or the other. The rate is the most visible number and the least predictive one. What determines whether an offshore engagement saves money is the total cost of getting working software delivered, and several of those costs never appear on an invoice.
The costs that appear on the invoice
The billable rate varies with seniority, engagement model and how specialised the skill is. It is worth asking what the rate includes: whether project management is billed separately, whether QA is a distinct line, and what happens to the rate when you need someone senior for a week.
The costs that do not
- Your own management time. Someone on your side has to set priorities and answer questions. Under-resourcing this is the most common reason offshore engagements disappoint.
- Specification effort. Distance raises the cost of ambiguity. Requirements that a colocated team would resolve in a corridor conversation need to be written down.
- Onboarding. Any team needs time to become productive in your domain and codebase. Treat the first weeks as an investment rather than delivery.
- Turnover. If people rotate off your account frequently, you pay onboarding repeatedly. This is worth asking about directly.
What actually moves the total
In most engagements the dominant variable is not the rate but rework, and rework is driven by clarity and feedback speed. A slightly more expensive team that asks good questions early will usually cost less over a year than a cheaper one that builds precisely what was written.
How to compare properly
Ask each vendor to estimate the same well-defined piece of work, and compare the total estimate rather than the rate. Ask what they would need from you to hit it. The vendor whose answer includes a list of decisions they will need from your side is usually the one who has done this before.
When offshore does not save money
Short engagements rarely repay their onboarding cost. Work that needs constant stakeholder access is difficult at distance. Being honest about this improves the odds on the work that does suit the model.



