Why ODC Setup in India Loses 40% of Staff in Year One

The call that every India expansion lead dreads

You are four months into your offshore development centre build. The CFO signed off on 35 seats. You have filled 28. This morning you received resignation number twelve.

The exit interview cited 'lack of career clarity' and 'disconnect from the core product vision'. Salary was not mentioned. Your recruiter says the market is simply hot. Your EOR provider reminds you that 90-day attrition is normal in India and suggests you over-hire by 20%.

Both are missing the point.

Attrition in the first 18 months of a captive or ODC setup in India is not a labour-market problem. It is a design problem, and it begins the day you sign the entity paperwork or the employer of record services agreement, not the day someone resigns.

This article unpacks why so many India operations lose 35 to 50% of their inaugural cohort before the end of year one, and what a retention-first setup model actually requires.

You hired for skill but onboarded for compliance

Most new ODCs run onboarding like a three-day checklist. IT setup, HR policy walkthrough, compliance training, meet the team on Zoom. Then the new joiner is handed a Jira backlog and wished good luck.

That works if the employee already understands the product, the customer and the commercial reason the feature exists. In month one of a greenfield ODC, no one does.

The result is capable engineers building features they cannot explain to a friend, QA analysts testing workflows they have never seen a user touch, and data analysts producing dashboards no one in India has clearance to interpret. The work is real. The connection is abstract.

Within eight weeks, a competitor offers the same salary plus the promise of 'more interesting problems'. Your developer accepts, because interesting is the only variable left that differentiates the role.

Retention-first onboarding inverts the sequence. Compliance can wait until week two. Week one is reserved for context: a recorded walkthrough of the product by the VP of product, a customer win story narrated by sales, and a diagram that shows where this team's output lands in the value chain. If the role involves customer data, show a real (anonymised) support ticket the work will help close faster.

One fintech client we supported in Pune had lost 9 of their first 16 developers within five months. Attrition was running at 56% annualised. We rebuilt onboarding so that every new joiner spent their first afternoon pair-programming with someone in London via Zoom, then received a weekly 15-minute check-in with the product owner for the first 90 days. Twelve months later, attrition had fallen to 11%, and two of the original seven hires were leading squads of five.

Context is not a luxury. It is the difference between someone who codes and someone who stays.

You delegated line management to someone who has never managed remotely

In the scramble to hire fast, many firms promote the first senior engineer who joins into a 'team lead' role by month three. That person is often excellent technically and has never run a one-to-one in their career, let alone across time zones with a reporting line that zigzags through an EOR contract.

The new lead does what feels natural: they focus on delivery. Standups become status updates. One-to-ones become ticket triage. Career conversations do not happen because no one has explained what the progression framework looks like, or whether one even exists.

Meanwhile, the employee who joined expecting mentorship finds themselves in a coordination meeting. After 90 days, they update their LinkedIn profile to 'open to opportunities'.

Retention-first setup means hiring or appointing managers who have run distributed teams before, or investing in a four-week bootcamp that teaches them how. That bootcamp should cover how to run a career conversation when you cannot promote someone for 18 months, how to escalate a performance issue when the employee sits under an employer of record services contract, and how to make someone feel seen when you meet twice a year.

We worked with a logistics SaaS operator who hired 18 QA engineers in Hyderabad through an EOR. Eleven quit within 90 days. The reason, surfaced in exit interviews, was that no one had explained how their work connected to the end customer. We introduced a monthly all-hands where the CEO walked through one recent customer win that QA had de-risked. Attrition dropped from 61% annualised to 19%, saving roughly £140,000 in rehire and lost-productivity cost over eight months.

Management is not optional. It is the architecture that keeps people longer than the salary does.

You treated the India team as a satellite, not a capability

Many ODC setups begin with the assumption that India will execute work defined elsewhere. Requirements come from London. Priorities come from New York. Innovation stays in San Francisco.

That model works for six months. Then your best engineer asks when they will be invited to shape the roadmap, and you realise you have no answer.

The narrative that India is 'delivery only' becomes a retention tax. High performers leave for roles where they influence outcomes, not just implement them. Mid-tier performers stay but disengage. Within 18 months you have built a team that ships on time and contributes nothing to product strategy.

Retention-first India expansion starts with the assumption that by year two, at least one strategic initiative will be owned end-to-end in India. That does not mean handing over the entire platform. It means giving the ODC one feature, one market experiment, or one technical spike where they define the problem, propose the solution, and present the outcome to the board.

Signal that intent in the first all-hands. Reinforce it by rotating one India team member into a quarterly sprint in the home market. Reward it by promoting the first India hire who ships a capability that generates revenue or reduces cost.

When people believe they are building a centre of excellence, not just a cost centre, they stop surveying the market every six months.

You optimised the contract but neglected the first 90 days

Most procurement and finance teams spend weeks negotiating EOR fees, IP clauses, and notice periods. Then they spend zero hours designing what happens between offer acceptance and day 90.

The contract governs the exit. The first 90 days govern whether an exit ever happens.

Retention-first setup allocates a named buddy (not the line manager) for every new joiner, schedules a welcome video from the C-suite within 48 hours of start date, and creates a 30-60-90 day plan that includes at least one visible win the new hire can point to in their first quarterly review.

It also means auditing the tools and access barriers that make new joiners feel like outsiders. If it takes three weeks to get Jira access, two weeks to get added to Slack, and four weeks to get read rights on the roadmap, you have just told someone they are not trusted. They will remember that when the recruiter calls.

You have no answer to the question 'What happens after two years here?'

In mature markets, employees tolerate ambiguity about progression because they can see the org chart. In a 20-person ODC that reports into a 300-person HQ, there is no chart. There is a reporting line to someone in another country, a vague promise of 'growth', and a salary review cycle no one has explained.

Without a credible answer to 'What does good look like in 18 months, and what role could I grow into?', your highest performers leave before you have a chance to promote them.

Retention-first ODC design includes a published progression framework by month six, even if it is a simple one. Individual contributor to senior individual contributor to tech lead to engineering manager. Associate to analyst to senior analyst. Each level should have three to five observable behaviours, a salary band, and an example of someone who has made the jump.

If you cannot promote anyone for two years because the headcount is frozen, say so. Then explain what mastery looks like at their current level, what stretch projects they can own, and what skill development you will fund. Clarity is a retention lever even when opportunity is constrained.

Frequently asked questions

How much does attrition cost in the first year of an ODC setup in India?

For a technical role at mid-level, expect four to six months of loaded cost (recruiter fee, onboarding time, lost productivity, backfill delay). A 30-person ODC losing 12 people in year one can easily incur £250,000 to £400,000 in attrition cost, depending on seniority and speed of replacement.

Should I use an employer of record or set up my own entity to reduce attrition?

Attrition is a function of onboarding, management quality and career narrative, not contract structure. An EOR can be faster and lower-risk for the first 18 months, but retention requires the same design discipline either way. If you plan to scale past 50 heads, a captive entity offers more control over benefits, culture and long-term EVP.

What is a realistic attrition target for a new ODC in India?

Industry baseline for captives under 50 people is 25 to 35% annualised in the first two years. Retention-first setups achieve 12 to 18%. Anything below 10% in year one usually signals over-investment in retention at the expense of performance management, or a market where competitors are not hiring.

What this means for your team

If you are planning an ODC, EOR engagement or centre of excellence build in India, the design choices you make before the first hire joins will determine whether you are filling the same seats twice or building a team that compounds in capability.

Expante Global Consulting works with US, UK and European firms to design retention into entity setup, EOR onboarding and governance models from contract signature onward. We have supported more than 60 ODC and captive builds, and the common thread among the ones that retain talent is that they treated the first 90 days as strategic, not administrative.

If you are evaluating an employer of record services provider, scoping a GCC or captive setup, or inheriting an India team that is haemorrhaging talent, the questions worth asking now are: What does onboarding look like on day one? Who owns the career conversation? And what story are we telling people about why this role matters?

Those answers shape attrition more than salary ever will.