Offshore Software Development Center (ODC): Models, Setup, and the Exit

Most guides to the ODC model explain what the acronym stands for and then walk you through five setup steps. That is useful for about ten minutes, which is roughly how long it takes to realise the decision in front of you is not “how do I set one up” but “should this be a center at all, and what happens to it in year four.” This guide answers those two questions first. It assumes you already know you need offshore engineering capacity and now have to choose the structure that will hold it for years.

Key Takeaways

• An offshore software development center is standing capacity, not a project. If your roadmap has an end date, you want a different model.
• The engagement model decides who employs the engineers and who keeps the domain knowledge. Settle that before comparing vendors.
• Setup is not the hard part. Retention and the handover are, and almost nobody prices them.
• Saigon Technology publishes its implementation rates at $22 to $46 per hour, against a market band of $25 to $70 and up.
• We have delivered 85+ offshore dedicated teams and transferred 40+ developers into client-owned entities when clients chose to bring the work in-house.

When an offshore development center beats the alternatives

Choose a center when the work is continuous, the domain takes months to learn, and you expect to still be building the same product in three years. Choose almost anything else when it is not.

That sounds obvious written down. It is still the most common mistake we see. An offshore software development center carries fixed overhead that never switches off: recruitment, infrastructure, a legal footing, and a retention budget. Those costs are entirely rational when they are amortised across years of continuous delivery, and they become irrational the moment you try to spread the same overhead across a single nine-month build with a defined end state. That gap is why so many companies conclude the model failed. It did not. They bought standing capacity to do finite work.

Three signals say a center is the right structure. All three should be true, not one.

If only the first is true, a dedicated development team gives you most of the benefit with far less overhead. If none are true, project outsourcing is the honest answer. Any vendor who pushes a center at you anyway is selling its own margin rather than your outcome.

There is a market signal worth reading here too, because the headline growth in outsourcing is not happening in the segment an ODC competes in. ISG, which measures signed commercial contracts worth $5 million or more in annual contract value rather than survey responses, reported that managed services grew just 2.7% year on year in Q2 2026 while cloud-based as-a-service spending rose 65%, and it forecast managed services at 2.1% for the full year (ISG Index, 9 July 2026). Classic managed delivery is close to flat. That is precisely why buyers are re-examining whether standing capacity they might one day own beats renting delivery indefinitely.

What is an offshore software development center (ODC)?

An offshore software development center is a dedicated, full-time engineering team in another country that works only on your product, under your processes and your technical direction. The vendor supplies the entity, the office, the payroll and the hiring pipeline. You supply the roadmap and the standards. The team is yours in practice while remaining the vendor’s employees on paper. That is the structural difference from how we staff and run offshore engineering teams on a project basis.

The term is used interchangeably with offshore development center, and both are usually shortened to ODC. The distinction that matters is not the name. It is the horizon. A center is priced per engineer per month and assumes an indefinite commitment, whereas project outsourcing is priced per deliverable and assumes an end.

That pricing difference decides where the domain knowledge ends up. Pay per deliverable and the vendor keeps the understanding, because for them the understanding is the competitive asset and handing it over would erode the reason you keep buying. Pay per engineer and it accumulates inside people you could eventually employ directly. Almost everything else here follows from that one structural fact.

ODC vs other engagement models

The contract decides your exposure more than the vendor’s logo does. Four structures cover nearly every arrangement, and each one draws the line between your risk and the vendor’s somewhere different. Read the table with one question in mind. If this goes wrong, whose problem is it?

An offshore software development center is the only one of the four that treats the exit as a designed feature rather than as an ending. That is its real differentiator. It is also the part most buyers never ask about until the moment they need it, which is the worst moment to discover it was never written down.

Two adjacent terms cause a lot of confusion. A captive center is one you set up and own yourself from day one, carrying the full legal and HR burden with no partner. A global capability center, or GCC, is the same idea at enterprise scale, usually spanning finance, operations and support functions rather than engineering alone, and usually justified by a headcount that no single product roadmap could ever absorb on its own. The scale gap between the two models is large: Zinnov and nasscom counted 2,117 GCCs operating in India in FY2026, employing roughly 2.36 million people (Zinnov and nasscom, India GCC Landscape 2026). Note that Zinnov also sells GCC setup services, so read that count as an interested party’s tally rather than as neutral data.

An ODC sits between the two. It is partner-operated, engineering-focused, and convertible later.

The four ODC models

Vendors describe their arrangements in dozens of ways. The commercial substance reduces to four.

Most engagements begin as dedicated resource and drift toward hybrid as the product matures. That drift is normal. Plan for it, because the contract you sign in year one is rarely the shape you want by year three, and renegotiating from inside a dependency is a materially weaker position than agreeing the mechanism up front while you are still choosing between suppliers. The model an offshore software development center starts in is not the model it ends in.

How to set up an offshore software development center

Setting up an offshore software development center runs to five steps. The first two decide whether it works. The last three are execution, and execution is the part vendors are genuinely good at.

1. Choose the location

Location fixes your talent pool, your cost band and your time-zone overlap, and those three trade against each other. Judge a country on the depth of its mid-to-senior engineering pool rather than on its graduate output. A center does not fail for want of juniors. It fails when there is nobody to promote into a lead role.

2. Choose the partner and the model

This is the step that decides the outcome, and the questions that matter are commercial rather than technical. Ask who employs the engineers. Ask what happens to the team if you later want to take it in-house. Ask for the attrition rate on their existing centers, and ask for it as a number rather than an adjective. A partner who has never transferred a team has never tested the exit they are describing to you.

3. Stand up infrastructure and security

Development environments, access control, device management and network segregation. If your sector carries a compliance regime, this is where it lands, and retrofitting it later costs materially more than building it in now. Saigon Technology is certified to ISO 9001 and ISO 27001 by BSI, which is the sort of baseline worth asking any partner to evidence rather than assert.

4. Settle the legal entity and the IP

Two questions, both far cheaper to answer before signing than after. Who owns the code as it is written, and by what mechanism does that ownership pass to you? What is the written path if you later want to employ the team directly? A good partner hands you the answer as a clause. A weaker one offers reassurance.

5. Plan HR, payroll and retention

Offshore development center setup is usually treated as finished once the team is seated. It is not. Retention is the recurring cost that decides whether the center compounds knowledge or quietly leaks it, and it belongs in the model from the first month rather than from the first resignation.

What an offshore development center costs to run

Published rates are rare in this market, which makes honest comparison harder than it should be. Saigon Technology publishes its implementation rates. The market band around them is wide.

For context on the other side of that gap, the US Bureau of Labor Statistics put the median annual wage for American software developers at $135,980 as of May 2025 (U.S. Bureau of Labor Statistics). In Vietnam, ITviec’s 2025 survey of 1,839 IT respondents put senior back-end developers at a median of 54.9 million VND per month (ITviec Vietnam IT Salary Report 2025-2026); that is a self-selected online panel rather than a probability sample, so treat it as indicative. Both figures are salaries rather than billing rates, and the delivered cost gap is materially narrower than the salary gap once employer contributions, management overhead and vendor margin are added.

Which is the real point. The hourly figure is the smaller half of the calculation. What actually decides whether an offshore software development center pays back is the set of costs that never appear on a rate card at all: the ramp period before a new engineer is productive, the rework that follows every departure, and the recruitment cycle each replacement sets off behind it. A center at the top of the band with low attrition is routinely cheaper across three years than one at the bottom with high attrition. The rate applies to hours. Churn applies to knowledge.

Attrition at that scale is measurable, and the large offshore providers disclose it. For the quarter ended 30 June 2026, Infosys reported last-twelve-month voluntary attrition in IT services of 13.0% (Infosys Form 6-K, filed 28 July 2026) and Wipro reported 13.9% on a trailing twelve-month basis (Wipro Form 6-K, filed 16 July 2026). Those are India-headquartered global firms rather than Vietnamese centers, so read them as an industry reference point rather than a like-for-like benchmark. On a twenty-person team, losing 13% a year means roughly one departure every five months, each carrying months of undocumented context out of the door.

So ask any prospective partner for two numbers alongside their rate. First, attrition on comparable accounts. Second, average time from a developer joining to delivering independently. Vendors who track those numbers will tell you what they are. Vendors who do not track them will tell you something else instead.

Vietnam as an offshore development center location

Our own centers are in Vietnam, so treat this as a practitioner’s account rather than a neutral survey.

One official figure is worth more than the rest of the country marketing put together. Vietnam’s Ministry of Science and Technology reported that Vietnamese digital technology firms earned $11.5 billion in overseas revenue in 2024, up 54% year on year (Ministry of Science and Technology, 15 January 2025). That is specifically revenue from foreign markets, which is the offshore delivery economy rather than the domestic or hardware-manufacturing one, and it is the number that tells you whether a country does this work at scale.

The practical case then rests on three things. Engineering salaries sit well below US and Western European levels, while the mid-to-senior pool is deep enough to promote from internally, which is the constraint that actually binds in year three. Time-zone overlap with the United States runs a 10 to 12 hour offset, so structured handover works well and same-day synchronous collaboration takes deliberate scheduling. Retention also behaves differently from the region’s higher-churn markets, and for the reasons in the section above that matters more than the headline rate does.

There is an honest limitation. It is scale at the very top end. If you need two hundred engineers on one product inside twelve months, the larger regional markets will absorb that demand faster than Vietnam will, and a partner who tells you otherwise is quoting a hiring plan rather than a delivered capability. Saigon Technology runs three delivery centers with 400+ developers, which suits teams in the five-to-fifty range rather than the several-hundred range. The specifics of standing one up here are on our Vietnam Offshore Development Center page.

Owning the center: the build-operate-transfer exit

We have transferred 40+ developers into client-owned entities under a build-operate-transfer center model. The clearest case ran for six years with a Netherlands client, scaling from 2 engineers to roughly 50, and finished with a complete transition into their own entity with no delivery disruption across the handover. A shorter worked example is our logistics center engagement.

That last clause is the entire point. Most discussions of build-operate-transfer describe the mechanism, and the mechanism is not hard to describe. What is hard is executing it without a delivery gap, because at the moment of transfer the people doing the work change employer, the tooling changes owner, and the institutional memory has to survive both of those at once while releases keep shipping. A partner who has done it can tell you where it strained. A partner who has only modelled it will describe something clean and frictionless. That is the tell.

Three things make the exit from an offshore software development center survivable, and all three have to be in place years before you use them.

“Clients negotiate the rate hard and sign the exit clause without reading it. Four years later the rate has saved them a few percent, and the exit clause decides whether they own anything at all. I would rather lose a deal on price than write a clause I know we cannot execute.” – Thanh (Bruce) Pham, Chief Executive Officer at Saigon Technology

Where ODC engagements actually fail

An offshore software development center rarely fails visibly. Three quieter patterns account for most of it.

Attrition outruns documentation

The center keeps its headcount and loses its memory. Nothing in the reported metrics moves, because headcount is the thing that gets reported and it has not changed. The symptom surfaces months later, as slowing delivery on work the team has already done before. By then the cause is a year old.

The center becomes a second organisation

It develops its own standards, its own review culture and eventually its own view of the roadmap. This is the failure mode that looks healthiest on a dashboard, because a self-directing team reports well right up to the point where its priorities and yours have quietly separated.

Nobody owns the relationship on your side

A center needs a counterpart with authority to decide, not a monthly status call. Where that role sits unfilled, decisions queue up, and the team defaults to whichever interpretation lets it keep moving. That is rational behaviour. It is also how a roadmap drifts without anyone choosing to change it.

Each of the three is recoverable on its own. They matter because they compound, and because the remedy for all three is unglamorous: a named owner on your side, documentation treated as delivery, and succession named before it is needed rather than after.

Frequently asked questions

1. How long does it take to stand up an offshore development center?

Plan on three to six months from signature to a team delivering independently. Recruitment is the long pole for senior roles, and the ramp to independent delivery usually takes longer than the hiring did. Vendors quoting a few weeks are describing a team seated, not a team productive.

2. What does an ODC cost per developer?

Saigon Technology publishes senior-led implementation rates of $22 to $46 per hour, against a market band of $25 to $70 and up. Model the ramp period and the expected attrition alongside the rate, because those two together usually move the three-year figure more than the rate itself does.

3. What team size makes a center worth it?

Below roughly five engineers the fixed overhead is hard to justify, and a dedicated team is usually the better structure. An offshore software development center earns its overhead once the team is large enough that internal promotion becomes possible, because that is the point at which it stops depending on external hiring for continuity. Our mobile product team engagement shows how that plays out at a mid-size team.

4. Who owns the IP and the source code?

You should, from the moment it is written, through an assignment clause rather than through custom or good intentions. Confirm the mechanism in writing. Confirm too that it covers work in progress and not only delivered releases, because the gap between those two is where disputes actually happen.

5. How does an ODC differ from a dedicated team?

Mostly in horizon and in exit. A dedicated team is continuous capacity with no assumption about eventual ownership. A center assumes a multi-year commitment and is structured so that ownership can transfer to you. The daily working experience is similar. The contract is not.

6. Can we take ownership of the center later?

Under a build-operate-transfer arrangement, yes, and it should be agreed at signing rather than negotiated at the point you want to use it. Ask any partner how many transfers they have actually completed. We have moved 40+ developers into client-owned entities.

Choosing your next step

Structure matters more than destination here. Work out whether your roadmap genuinely needs standing capacity rather than finite delivery, decide now who you want employing these engineers in year four, and get the exit written down while you still have enough negotiating room to get it written well rather than merely written. A partner willing to discuss the handover before you have signed anything is telling you something useful about how that handover will actually go. Ask early.

If you want to talk through which model fits your roadmap, or what an exit clause should actually say, get in touch with our team.