Global capability
An offshore development centre that becomes yours
An offshore development centre is a dedicated engineering function in India, recruited to your specification and operating under your standards, with employment, facilities, IT and statutory compliance held by us. You get the team, the continuity and the intellectual property. Optionally structured as build-operate-transfer, so the people and the entity can move onto your books later at a fee agreed up front.
- Ten to one hundred-plus engineers, scaled in cohorts
- Build-operate-transfer with the transfer fee fixed at signature
- Open-book cost-plus pricing available, so the margin is visible
- Engineer attrition below 9 percent against an industry range of 15 to 25
At a glance
- Minimum viable size
- Around 10 engineers
- Minimum sensible horizon
- 30 to 36 months
- Time to first cohort productive
- 8 to 14 weeks
- Scaling rate after cohort one
- 6 to 8 engineers per month
- Locations
- Dehradun (primary), Bengaluru (secondary)
- Engineer attrition
- Below 9 percent annually
- Data residency options
- India, EU, UK, US, UAE, Singapore, Australia
- Enquiries
- nitesh@redcubical.com
Definition
What an ODC is, and what it is not
The term is used loosely, often to describe an ordinary outsourcing contract with a dedicated-sounding name. Here is the distinction that matters commercially, set against project outsourcing on one side and a captive global capability centre on the other.
ODC, project outsourcing or captive
An ODC sits between project outsourcing and a captive centre. Project outsourcing buys a defined outcome and ends. A captive or GCC means you incorporate an Indian entity, hire the people yourself and carry every statutory obligation. An ODC gives you the dedicated team, the continuity and the control of a captive, with the employment, compliance and facilities risk held by us.
| Dimension | Project outsourcing | Offshore development centre | Captive centre / GCC |
|---|---|---|---|
| What you are buying | A defined deliverable for a defined price | Ongoing dedicated engineering capacity and continuity | Your own subsidiary and your own employees |
| Who employs the engineers | The supplier, usually shared across accounts | Redcubical Systems Private Limited, dedicated to you and named in the contract | You, through an Indian legal entity you incorporate |
| Legal entity required in India | None | None. We are the entity | Yes. Private limited company, registrations, statutory filings, board |
| Time to first productive engineer | 2 to 4 weeks | 8 to 14 weeks for cohort one | 9 to 18 months including incorporation, premises and leadership hiring |
| Setup cost | Effectively nil | A mobilisation fee: typically USD 25,000 to 90,000 by size | USD 250,000 to 1,000,000 plus before the first engineer ships anything |
| Cost per engineer at steady state | Highest per unit for sustained work | Moderate. Our margin on a fully-loaded cost base | Lowest in theory, once you have absorbed management and compliance overhead |
| Who carries attrition and recruitment risk | The supplier, invisibly | Us, contractually, with replacement obligations | You, entirely |
| Who carries statutory and employment risk | The supplier | Us. Payroll, PF, ESI, gratuity, labour compliance, POSH | You, including termination law and litigation exposure |
| Control of technical standards | Limited. You review outputs | Full. Your standards, your tooling, your definition of done | Full |
| Employer brand | The supplier’s | Co-branded. We recruit in your name with your engineering story | Yours |
| Exit | Contract ends. Knowledge transfer | 60 days notice, or exercise a transfer option per the agreed schedule | Entity wind-down, retrenchment process, or sale. Slow and expensive |
| Honest best fit | Bounded work with a real end date | 10 to 150 engineers, multi-year horizon, no appetite for an Indian subsidiary | Above roughly 200 engineers, or where the centre is strategically permanent |
The comparison people skip is the middle-to-right one. A captive looks cheaper per engineer on a spreadsheet and frequently is, at scale. What the spreadsheet omits is nine to eighteen months of setup, an India leadership team you must hire before you have anyone to lead, and the fact that exiting an Indian entity is materially harder than exiting a contract.
Who does what in an ODC
You own
- The product roadmap, priorities and release decisions.
- Technical standards, architecture direction and the definition of done.
- Role definitions and the final interview decision on every hire.
- All intellectual property, from creation.
- Your repositories, cloud accounts, data and environments.
- The right to require removal of any individual from the account.
We own
- Employment: payroll, statutory contributions, benefits, performance and career management.
- Recruitment throughput against your role definitions, and retention.
- Facilities, physical security, network, endpoints and access control.
- Indian statutory and labour compliance, including PF, ESI, gratuity and POSH obligations.
- Delivery governance: reporting, reviews, escalation and the steering committee.
- The consequences of attrition: replacement, overlap and context recovery.
Build, operate, transfer
The phased model and what transfers at each phase
BOT exists so you can start without an Indian subsidiary and still end up owning the capability. The critical commercial point is that the transfer fee schedule is agreed at signature, not negotiated in year three when your leverage has evaporated.
How BOT is phased
Build-operate-transfer runs in four phases. Build stands up space, IT and the first cohort in 8 to 14 weeks. Operate runs the centre for two to three years while it accumulates domain knowledge. Prepare puts transfer readiness in place over six months. Transfer moves people, processes and optionally an entity to you, at a fee fixed in the original contract.
-
Phase 0 — Weeks 1 to 4
Foundations
Master services agreement and ODC schedule signed. Space identified and fit-out started. Network segmentation designed with your security team. Role definitions agreed and the recruitment brief written. Your security questionnaire completed and any findings closed. Nothing transfers yet; what you get is a signed sizing plan, a named ODC manager and an agreed governance calendar.
-
Phase 1 — Weeks 3 to 14: Build
First cohort recruited, onboarded and productive
Recruitment runs against your role definitions with your engineers in the final loop. Space, badge access, segregated network and endpoints delivered. Tooling provisioned in your instances. First cohort onboarded, first increment shipped by week 14. Transfers on completion: the recruitment specification, the interview scorecards, the onboarding runbook, the environment and access model, and all code and IP from the first commit.
-
Phase 2 — Months 4 to 30: Operate
Scale, stabilise and accumulate domain knowledge
Scaling at six to eight engineers per month to the target headcount. Delivery governance running: weekly reports, monthly reviews, quarterly steering. Retention programme active. Domain knowledge documented deliberately rather than held in people. Transfers continuously: architecture decision records, runbooks, domain documentation, team performance data, and the full cost base under an open-book model.
-
Phase 3 — Months 24 to 30: Prepare
Transfer readiness, six months before the date
Transfer scope confirmed: knowledge only, employees, or entity. Employee consultation planned. Your India leadership hired if you are taking the team on directly, and shadowing our ODC manager. Vendor, lease and licence assignments mapped. A written transfer plan with dated milestones, signed by both sides. Transfers: the transfer plan itself, the asset and contract register, and the compliance file.
-
Phase 4 — Months 30 to 36: Transfer
People, processes, assets and optionally the entity
Employees offered transfer with continuity of service and protected terms. Payroll cut over on an agreed date. Leases, licences and equipment assigned or sold at agreed book value. Our ODC manager stays for 60 to 90 days post-transfer in an advisory role. Transfers: the employment relationships, the physical and IT assets, the operating processes, and full operational control.
-
After transfer
Advisory, or a clean end
Either a defined advisory retainer for two to four quarters while your own management settles in, or a clean end with a 90-day question window. What we avoid is the ambiguous arrangement where your team assumes we are still covering something we are not.
Sizing
What ten, twenty-five, fifty and a hundred engineers each require
Governance requirement does not scale linearly with headcount. It steps. Each of these bands needs a genuinely different management structure, and getting that wrong is the most common reason an ODC stalls at around forty people.
| Size | Indicative annual cost (USD) | Lead time to full strength | Governance it needs | Structural risk at this size |
|---|---|---|---|---|
| 10 engineers | 0.75m to 0.95m | 10 to 14 weeks | One ODC manager doubling as delivery lead. Weekly report, monthly review. A single steering call each quarter | Fixed overhead is barely amortised. Honestly, a dedicated team is usually better value at this size |
| 25 engineers | 1.8m to 2.3m | 5 to 7 months | ODC manager, two delivery leads, a dedicated talent partner at 50 percent, shared IT and security owner. Formal quarterly steering committee | The first real management step. Teams above about eight people with one lead lose backlog discipline |
| 50 engineers | 3.4m to 4.4m | 9 to 12 months | ODC head, four to five delivery leads, an engineering manager layer, full-time talent partner, full-time IT and security owner, a QA and release function | Where most ODCs stall. Coordination cost grows faster than headcount unless service boundaries are made explicit |
| 100+ engineers | 6.5m to 8.5m | 18 to 24 months | ODC head with a leadership team, domain-aligned tribes of 20 to 30, a platform team, a dedicated compliance and audit function, monthly steering with executive attendance | Becomes a genuine organisational design problem. Above roughly 150 the captive question deserves a serious answer |
Costs are indicative fully-loaded annual figures at a typical mid-weighted seniority mix of roughly 20 percent junior, 45 percent mid, 25 percent senior and 10 percent lead, inclusive of facilities, IT, management and our margin. A senior-weighted team is 20 to 30 percent higher. Lead times assume recruitment throughput of six to eight engineers per month sustained, which is our practical ceiling without lowering the bar.
Support ratios we plan to
- Delivery lead: one per six to eight engineers. Above eight, backlog readiness and code review latency both degrade measurably.
- Engineering manager layer: introduced at around 35 engineers, one per three delivery leads.
- Talent partner: half a person per 25 engineers at steady state, one full person per 25 during a scaling phase. Recruitment is the constraint during build, not space.
- IT and security owner: shared below 25 engineers, dedicated from 25, a small team from 75.
- QA capability: one QA automation engineer per four to six developers, depending on how much of your regression suite already exists.
- Business analysis: one per 10 to 15 engineers where requirements originate offshore. Zero where your product organisation writes specifications itself.
How the ramp actually looks
Headcount plans drawn as straight lines do not survive contact with recruitment. The realistic shape for a fifty-person ODC:
- Months 1 to 3. Zero to eight engineers. Slowest phase, because there is no internal referral network yet and every hire is sourced cold.
- Months 4 to 8. Eight to twenty-eight. Fastest phase. Referrals from cohort one start producing, and the interview loop is calibrated.
- Months 9 to 12. Twenty-eight to fifty. Slows again as the remaining roles are the specialist ones, and as onboarding capacity rather than sourcing becomes the constraint.
- Steady state. Roughly four to five replacement hires per year at sub-9 percent attrition, plus growth.
We plan cohorts, not a smooth curve, because onboarding six people together is materially more effective than onboarding one person six times.
Cost transparency
Three pricing models, and where every dollar goes
How ODC pricing works
Three options. Per-seat is a fixed monthly rate per role, simplest to budget, with our margin embedded and not itemised. Cost-plus is open book: you see actual salary, benefits, facilities and IT cost and pay an agreed management margin of 12 to 18 percent. Fixed-fee management passes cost through at actual and charges a flat monthly management fee independent of headcount.
| Model | How it works | Best for | Trade-off you accept |
|---|---|---|---|
| Per-seat | A fixed monthly rate per role and seniority band. One line per seat on the invoice. Annual review against a published index | Simple budgeting, procurement that wants a unit price, ODCs under 25 engineers | You cannot see the margin. Salary inflation is our risk, which means we price for it |
| Cost-plus, open book | Actual salary, statutory cost, benefits, facilities, IT and management cost disclosed. You pay actuals plus 12 to 18 percent management margin | ODCs above 25 engineers, clients who want auditable transparency, anyone planning a future transfer | Administratively heavier for both sides. Your budget moves with Indian salary inflation rather than being fixed |
| Fixed-fee management | Costs passed through at actual with no mark-up. A flat monthly management fee, typically USD 18,000 to 45,000 by size, independent of headcount | Large ODCs where a percentage margin on a growing cost base becomes hard to justify | Our incentive is decoupled from your headcount, which is good, but the fee is payable whether the team is 45 or 55 people |
We recommend cost-plus above 25 engineers and per-seat below it. If you intend to transfer the centre later, take cost-plus from the start: it gives you the cost base you will need for the transfer business case, and it means the transfer is not the first time you see the real numbers.
| Cost line | Monthly amount | Share | What it covers |
|---|---|---|---|
| Gross salary | 2,850 | 42% | Cash compensation for a mid-level engineer, 3 to 6 years, at Dehradun and Bengaluru market rates |
| Statutory employment cost | 410 | 6% | Provident fund, ESI where applicable, gratuity accrual, professional tax, employer statutory contributions |
| Benefits and insurance | 235 | 3% | Group medical cover including dependants, personal accident and term cover, wellness allowance |
| Recruitment, amortised | 190 | 3% | Sourcing, agency fees where used, interview time and referral bonuses, spread over an expected 40-month tenure |
| Facilities | 360 | 5% | Desk, meeting space, power, redundant UPS and generator, cleaning, security guarding, rates |
| IT and connectivity | 295 | 4% | Laptop amortised over three years, dual monitors, redundant leased lines, MDM, endpoint detection, VPN, standard licences |
| Learning and certification | 95 | 1% | Cloud certifications, conference and course budget, internal training time |
| Delivery and ODC management | 640 | 9% | Delivery lead, ODC manager, talent partner, IT and security owner, apportioned per seat |
| Corporate overhead | 425 | 6% | Finance, legal, HR, compliance, audit, insurance, ISO-aligned control operation |
| Sub-total, our cost | 5,500 | 81% | What one seat genuinely costs us before margin |
| Management margin at 15 percent | 825 | 12% | Our profit. Disclosed rather than embedded |
| Contingency for attrition and overlap | 475 | 7% | Funds replacement recruitment and the two-week paid overlap when someone leaves |
| Total invoiced | 6,800 | 100% | The number on your invoice for one mid-level engineer |
Indicative figures for illustration at a specific seniority and location mix, not a quotation. Two honest notes. First, the margin line is 12 percent of the invoiced total, which is a normal services margin and materially below the 25 to 40 percent typical of large consultancies. Second, the contingency line is real money that funds replacement and paid overlap. Suppliers who omit it either absorb attrition out of margin, which eventually degrades quality, or bill you for the replacement.
Recruitment, retention, facilities and security
How we hire for you, keep them, and secure the place they work
Recruitment and retention
We recruit against your role definitions with your engineers making the final call, using co-branded employer messaging so candidates know whose product they will build. Retention runs on four levers: above-market pay reviews, real technical progression, exposure to your engineers, and stable teams. Engineer attrition is below 9 percent annually against an industry range of 15 to 25 percent.
How we recruit for your ODC
- Your role definition, not ours. We write the brief with your engineering manager and it names the actual codebase, stack and problems. Generic requisitions attract generic applicants.
- Co-branded employer messaging. Candidates are told which client and which product from the first conversation, with your consent. This is the single largest lever on offer-acceptance rate we have found. Engineers accept a specific interesting product over an anonymous "leading global client" almost every time.
- Your engineers in the final loop. Every hire clears a technical interview with one of your engineers or ours acting on an agreed rubric, plus the six-stage funnel described on the hire developers page.
- Sourcing mix. Roughly 45 percent referral, 30 percent direct outbound, 15 percent inbound applications, 10 percent agency. Referral share rises after cohort one, which is why cohorts two and three are faster and cheaper.
- Location strategy. Dehradun as the primary hub gives us a materially lower attrition profile than Bengaluru, because there is far less competitive poaching. Bengaluru is used for specialist roles where the Dehradun pool is genuinely thin, notably senior ML and specialised data engineering.
- We will not lower the bar to hit a headcount date. If a role cannot be filled to standard we tell you and re-plan. A wrong hire on an ODC costs about five months.
Retention, and why sub-9 percent matters
- The arithmetic. On a fifty-person ODC, 9 percent attrition is about four and a half departures a year. At 20 percent it is ten. Each departure costs three to five months of productivity in recruitment, onboarding and lost context. That is 20 to 25 engineer-months a year of pure waste, which is roughly two extra headcount you pay for and never see.
- Compensation reviewed twice a year, not annually. Indian salary inflation in engineering has run high enough that an annual cycle guarantees you are below market for six months of every year, which is when people leave.
- Technical progression that is real. A published band structure with capability expectations, and promotion decisions made quarterly against evidence rather than tenure.
- Exposure to your engineers. The strongest non-cash retention factor we measure. Engineers who present demos to your team, join your architecture discussions and get reviewed by your seniors stay considerably longer than engineers who receive tickets.
- Team stability. We do not rotate people between accounts for internal convenience. Context is the asset.
- What we cannot fix. If the work is genuinely unrewarding, no retention programme survives it. An ODC given only maintenance backlog and no design ownership will have an attrition problem, and that is a joint problem rather than a supplier one.
Facilities, network security and access control
The physical and network layer of an ODC is where security questionnaires get detailed, so here is the specification rather than an assurance.
- Dedicated, segregated floor space for the ODC, physically separated from other client work. Not a shared coworking arrangement and not hot-desking
- Badge-controlled entry to the ODC zone with an access list you approve, quarterly access reviews, and same-day revocation on any personnel change
- CCTV on all entry and exit points with 90-day retention, plus visitor logging with escort requirements in client-data zones
- Clean-desk and no-paper policy in regulated engagements, with locked disposal and no printing capability enabled by default
- Segregated VLAN per ODC with no lateral routing to other client zones, and firewall rules reviewed quarterly with your security team
- Two independent leased-line providers with automatic failover, plus UPS and a diesel generator sized for full-floor operation
- Full-disk encryption, MDM enrolment, endpoint detection and response on every device, with no local administrator rights and patch compliance reported monthly
- Role-based least-privilege access to your systems through your identity provider, with just-in-time elevation and audited session recording for production access
- Virtual desktop option for regulated data: no local storage, clipboard and USB blocked, screenshot prevention, and all data remaining in your jurisdiction
- Mobile device restrictions in high-sensitivity zones, implemented as a physical policy rather than a written aspiration
- Annual penetration test of our own network and an internal control review, with the report shareable under NDA
- Your right to audit on 14 days notice, on site, once per year, plus an unlimited right to audit remotely on documentation
Legal, data and governance
India employment and statutory context, data residency, and how the ODC is governed
Legal and governance structure
You do not need an Indian entity for an ODC. We are the employer of record, which means Indian employment, payroll and statutory obligations sit with us: provident fund, ESI, gratuity, professional tax, labour law compliance and POSH obligations. Your data does not have to reside in India. Governance runs on three tiers: weekly delivery reporting, monthly delivery review, quarterly steering committee.
India entity, employment and statutory considerations
At a high level, and only at a high level, this is the landscape an ODC operates in.
- Employer of record. Engineers are employees of Redcubical Systems Private Limited, a private limited company incorporated in India. You contract with us commercially and have no employment relationship with any individual.
- Statutory contributions. Employer provident fund contributions, Employees' State Insurance where wage thresholds apply, gratuity accrual after five years of service, and state professional tax. All held and administered by us and visible in the cost base under an open-book model.
- Termination law. Indian employment termination is more procedural than at-will jurisdictions, with notice, documentation and in some cases state-level approvals for larger reductions. This matters most in an exit or downsizing scenario, which is why the exit clause is drafted in detail rather than left generic.
- Shops and Establishments registration under Uttarakhand and Karnataka rules for our respective premises, covering working hours, leave entitlement and holiday provisions.
- POSH compliance. An Internal Committee constituted under the Sexual Harassment of Women at Workplace Act, with mandatory annual training and annual filing.
- Permanent establishment risk. A properly structured ODC where we are the employer and you are a customer should not create a permanent establishment for you in India. The structure matters: heavy direct control over individuals, contracts concluded in India on your behalf, or a fixed place of business in your name all change the analysis.
- Transfer pricing. Relevant if you later transfer the centre into a subsidiary and it becomes a related party. Worth modelling before, not after.
- Zero-rated export of services. Services exported from India to an overseas client are generally zero-rated for GST with the appropriate declarations, so Indian indirect tax should not be a cost to you.
| Client market | Primary residency option | Cross-border mechanism for engineer access | Practical note |
|---|---|---|---|
| United Kingdom | UK region (London), or EU region if you prefer | UK International Data Transfer Agreement or the UK Addendum to EU standard contractual clauses, plus a transfer risk assessment | Access from India is a restricted transfer even when the data never leaves the UK region. The IDTA covers the access, not a copy |
| Germany and Netherlands | EU regions (Frankfurt, Ireland, Amsterdam) | EU standard contractual clauses with supplementary technical measures, plus a transfer impact assessment | Where a transfer impact assessment is uncomfortable, a virtual desktop with no local storage keeps personal data inside the EU region entirely |
| United States and Canada | US or Canadian regions | Contractual data processing terms. Sectoral rules apply: HIPAA business associate agreement for health data, state privacy laws such as CCPA | Canadian public-sector work frequently requires Canadian residency and sometimes Canadian-only access. We will tell you if that rules us out |
| United Arab Emirates and Saudi Arabia | UAE or Saudi regions, or EU where a local region is unavailable for a service | Contractual terms under UAE federal data protection law or Saudi PDPL, with local advice on sector rules | Financial and government sector rules can mandate in-country residency and in-country processing. That constrains offshore access materially |
| Singapore and Australia | Singapore or Sydney regions | PDPA transfer obligations in Singapore, Australian Privacy Principle 8 for offshore disclosure | Australian government and health work often requires onshore storage. Offshore access remains possible with the right controls |
| India | Indian regions (Mumbai, Hyderabad) | Domestic. DPDP Act obligations apply as data fiduciary and data processor | Sector rules bite: RBI requires payment data storage in India, and IRDAI and SEBI have their own localisation expectations |
An ODC does not require your data to be stored in India. In most engagements it is not. Engineers in India access data hosted in your jurisdiction over controlled paths, which is a cross-border access question rather than a storage question, and it is solved contractually and technically rather than by moving your data. Where a regulator prohibits offshore access outright, we will say so and decline the work rather than construct a workaround.
The governance model
Three tiers, each with a different question. Weekly asks whether delivery is on track. Monthly asks whether the team is healthy. Quarterly asks whether the ODC is still the right shape.
| Forum | Cadence | Attendees | Decisions it owns |
|---|---|---|---|
| Delivery stand-up | Daily, inside the overlap window | Your product owner, our delivery leads, the team | Day-to-day blockers, reprioritisation within the sprint |
| Delivery report | Weekly, written | Issued to your engineering and product leadership | No decisions. Progress, risks, blockers awaiting your input, and the burn against plan |
| Delivery review | Monthly, 60 minutes | Your engineering manager and product lead, our ODC manager and delivery leads | Velocity and quality trends, team health and attrition, recruitment pipeline, process changes |
| Steering committee | Quarterly, 90 minutes | Your executive sponsor, our ODC head and Nitesh as Head of Global Engagement | Headcount plan, budget, commercial changes, escalated risks, ODC scope and shape |
| Security and compliance review | Half-yearly | Your security and compliance leads, our IT and security owner | Access review outcomes, control changes, audit findings, penetration test remediation |
| Transfer readiness review | Quarterly, from month 18 on a BOT contract | Your executive sponsor and India lead, our ODC head | Transfer scope, timing, dependencies and readiness gaps |
Escalation runs on a defined ladder with response times: delivery lead within four working hours inside the overlap window, ODC head within one business day, and our Head of Global Engagement within one business day for anything commercial or relationship-level. The path is documented in the ODC schedule with names and contact details, not described in principle.
Exit, and the honest constraint
Transfer and exit mechanics, and when an ODC is the wrong structure
How you leave an ODC
Exit takes 60 days notice and a structured transition of 90 to 180 days. Three exit routes exist: wind-down, where we redeploy the people and hand over documentation; transfer to another supplier, with our cooperation contractually required; and transfer to you, where employees are offered continuity of service and assets are assigned at agreed book value.
| Route | Notice and duration | What happens to the people | What you pay |
|---|---|---|---|
| Wind-down | 60 days notice, 90-day transition | Redeployed to other Redcubical accounts where possible. Their employment continues with us and is not your liability | Fees to the end of transition, plus any unamortised mobilisation cost per the agreed schedule. No termination penalty |
| Transfer to another supplier | 60 days notice, 90 to 120-day transition | They remain our employees. We cooperate with the incoming supplier and do not obstruct them approaching individuals | Fees to the end of transition, plus a transition support fee at standard rates for effort beyond the included handover |
| Transfer to you (BOT) | 6 months notice, 90 to 180-day transition | Offered transfer with continuity of service, protected terms and no loss of accrued benefit. Individuals may decline; historically most accept | The transfer fee formula fixed at signature, plus per-employee transfer fee, plus assets at depreciated book value |
Employee transfer, in the detail that matters
- Consent is required and cannot be assumed. Indian employment does not transfer automatically on a contract novation. Each engineer receives an individual offer and chooses. We are contractually obliged not to discourage acceptance.
- Continuity of service is preserved in the transfer offer, which matters for gratuity eligibility at five years and for leave accrual. Breaking continuity is the fastest way to lose half the team at transfer.
- Terms protected or improved. The transfer offer cannot reduce cash compensation or benefits. Where your own benefit structure differs, the gap is bridged rather than left as a reduction.
- Retention incentive at transfer, typically two to three months of salary paid twelve months after transfer, funded by you and administered by whichever entity employs them at the time. Transfers without a stay incentive lose 25 to 40 percent of the team inside a year.
- Provident fund and gratuity transferred or settled per the mechanism agreed at signature. This is the detail most BOT contracts leave vague and it is the one that generates disputes.
- No non-solicitation obstruction. We commit in the original contract not to use non-solicitation or non-compete provisions to impede a transfer we have agreed to.
- Assets, leases and licences. Equipment sold at depreciated book value on a schedule fixed at signature. Leases assigned where the landlord permits, otherwise a sub-lease for a defined period. Software licences assigned or re-procured, with the list maintained throughout the engagement rather than reconstructed at exit.
- Management continuity. Our ODC manager stays for 60 to 90 days post-transfer in an advisory capacity, and your India lead should shadow them for at least a quarter before the date.
An ODC is a three-year commitment, and below ten engineers it is the wrong structure
This is the section we would remove if we were optimising for conversion. It stays because getting this wrong is expensive for you and unpleasant for us.
- The mobilisation cost needs time to amortise. Space fit-out, IT provisioning, the first recruitment wave and management set-up run to USD 25,000 to 90,000 depending on size. Over three years that is noise. Over twelve months it is a 5 to 9 percent premium on everything.
- The fixed overhead does not scale down. An ODC manager, a talent partner and an IT and security owner cost roughly the same whether they support eight engineers or twenty-five. At eight, that overhead is 20 to 25 percent of your invoice. At twenty-five it is 9 percent.
- Recruitment throughput has a floor. Standing up a dedicated hiring pipeline, calibrating an interview loop and building a referral network takes four to six months regardless of how many people you want at the end. On a small team you pay for that machinery and barely use it.
- Domain knowledge is the actual return, and it accrues slowly. An ODC becomes better value than a dedicated team somewhere around month fourteen to eighteen, when the team knows your business well enough to challenge requirements rather than implement them. Exit before then and you paid for a structure whose main benefit you never collected.
- Below ten engineers, take a dedicated team. Identical engineers, identical vetting, identical overlap commitment, no mobilisation fee, 30 days notice instead of 60, and 10 working days to first commit instead of 8 to 14 weeks. See hire developers. It is a smaller contract for us and the right answer for you.
- Above roughly 150 engineers, take the captive question seriously. At that scale the arithmetic of your own Indian subsidiary starts to work, and we would rather help you get there through BOT than argue you out of it.
Other cases where we would advise against an ODC
- Your engineering management is already stretched. An ODC does not reduce management load, it changes its shape and adds a quarterly planning obligation. It will expose a management gap faster and more expensively than any other model.
- The work is a single programme with a defined end. That is fixed-scope or capped time and materials work. See engagement models.
- A regulator prohibits offshore access to the data. Not "makes it complicated" but prohibits it. In that case an ODC in India cannot do the work and we will say so at the first call.
- You need the team in a specific country for contractual or security-clearance reasons. We deliver from India. We do not hold security clearances in any jurisdiction and we will not imply otherwise.
Answers
Offshore development centre questions
What is an offshore development centre?
A dedicated engineering function operating in India on your behalf, staffed with engineers recruited to your specification, working in your tooling, under your technical standards and your governance. We hold the employment, facilities, IT and statutory compliance. You hold the roadmap, the technical direction and the intellectual property.
How is an ODC different from just outsourcing a project?
Project outsourcing buys a defined outcome for a defined price and ends. An ODC buys ongoing capacity and organisational continuity. You are not procuring a deliverable, you are standing up a team that accumulates domain knowledge and stays. That is worth a great deal after month nine and worth nothing in month two.
What is the minimum viable size for an ODC?
Around ten engineers. Below that the fixed overhead of dedicated space, an ODC manager, a talent partner and an IT and security owner is spread across too few seats, and a dedicated team gives you identical output for less money. We will tell you if you are below the threshold.
How long does it take to stand up an ODC?
Eight to fourteen weeks to the first cohort of engineers being productive, for a team of twenty-five. Weeks one to four cover legal, space and IT foundations. Weeks three to ten run recruitment for the first cohort. Weeks eight to fourteen cover onboarding, tooling and the first delivery increment. Scaling past the first cohort adds roughly six to eight engineers per month.
What does build-operate-transfer mean in practice?
We build the team and operate it for an agreed period, typically two to three years, then transfer it to you. Transfer can be of processes and knowledge only, of the employees through a documented transfer mechanism, or of a whole entity. The transfer fee schedule is agreed at signature rather than negotiated later from a weak position.
How do you price an ODC?
Three options. Per-seat pricing is a fixed monthly rate per role, simplest to budget. Cost-plus is open book: you see actual salary, benefits, facilities and IT cost, and pay an agreed management margin. Fixed-fee management passes costs through at actual and charges a flat monthly management fee. Cost-plus is the most transparent and the most administratively demanding.
What is your attrition rate and why does it matter?
Below 9 percent annually, against an Indian IT services industry range of roughly 15 to 25 percent. On a fifty-person ODC the difference between 9 and 20 percent attrition is about five and a half extra departures a year, each costing three to five months of lost productivity in recruitment, onboarding and context loss.
Where does our data live and who can access it?
Data residency is your choice: India, EU, UK, US, UAE, Singapore or Australia. An ODC does not require data to be stored in India. Access is role-based and least-privilege, engineers work in a segregated network zone, and regulated data can be confined to a virtual desktop with no local storage or clipboard egress.
Model an ODC against your actual headcount plan
Send us the roles, the target headcount and the horizon. You get a sized plan with a fully-loaded cost model, a realistic ramp curve, a governance structure and a transfer fee schedule. If the answer is that you should take a dedicated team instead, that is what the document will say.