This guide covers one layer of a larger operating standard. For how retention fits with hiring, onboarding, workspace, devices and compliance, see the employer’s guide to professional remote work.

The number nobody puts in the business case

Offshore hiring decisions get made on a salary comparison. Someone builds a spreadsheet, puts the local cost in one column and the offshore cost in the other, and the gap is large enough that the decision looks obvious.

The spreadsheet is usually right about the salary and wrong about the total, because it assumes the person you hire is the person you still have in two years. In offshore support and technical roles, that assumption fails more often than it holds.

What the industry figures actually say, and how much to trust them

Two numbers get quoted constantly in this space.

The first is contact centre attrition, commonly cited in the range of 30 to 45% a year, with offshore voice operations often quoted higher and average agent tenure sitting around 14 to 15 months.

The second is the cost of replacing an employee, generally attributed to the Society for Human Resource Management and usually given as 50 to 200% of annual salary depending on seniority.

Both are worth knowing and neither is worth relying on.

Trace either figure and you find the same pattern: outsourcing vendors citing other outsourcing vendors, aggregator sites citing the vendors, and no primary research at the bottom of the chain. The SHRM range in particular is quoted almost universally without a locatable SHRM publication behind it. The direction of both numbers is not in dispute. The precision is entirely invented.

So use them as an order of magnitude and calculate your own.

The one benchmark that does survive tracing

There is a turnover figure your finance team can verify in a minute, and it comes from a government statistical agency rather than a vendor. The US Bureau of Labor Statistics publishes annual average quits rates by industry through its Job Openings and Labor Turnover Survey. For 2025 it recorded a quits rate of 2.3 for professional and business services and 2.2 for total private employment (BLS, JOLTS Table 22, annual averages, not seasonally adjusted, last modified 13 March 2026).

Those are monthly-equivalent rates, which is the detail almost everyone gets wrong when quoting them. BLS calculates the annual average as the sum of the twelve monthly quits levels as a percentage of the sum of the twelve monthly employment levels, so the annual rate is roughly twelve times higher. Across 2025, professional and business services recorded quits equal to about 28% of average employment. That is not 28% of people leaving, because one post can turn over more than once in a year, but it is the right order of magnitude for planning.

That is the number worth holding in your head, because it measures the buyer’s own domestic market, on the population closest to a professional offshore team, and anyone can check it at source. It reframes the question usefully: if an offshore arrangement is churning faster than the professional services average in your home market, the problem is the arrangement rather than the fact that it is offshore. For reference, twelve-month retention above 90% means annual attrition under 10%. The two are not strictly like for like: ours is headcount retention counting all leavers, BLS counts voluntary quits only, and the US labour market is far more liquid than Indonesia’s. The gap is wide enough to be worth stating anyway.

How to work out what turnover actually costs you

Four components. Only the first appears in most budgets.

1. Recruitment. Sourcing, screening, interviewing, offer, notice period. Four to eight weeks of elapsed time and a real cash cost, whether you pay an agency, a sourcing fee, or your own team’s hours.

2. Ramp. The period between someone starting and independently working at the standard of the person who left. For a Tier 1 support role, perhaps six weeks. For a Tier 2 engineer on a product with any real complexity, three to six months is honest. For a developer inside an unfamiliar codebase, longer.

3. Output quality during ramp. This is the component that never gets counted and is usually the largest. Slower resolutions. More escalations into engineering, which costs senior time at senior rates. Mistakes that reach customers. A backlog that builds while someone learns.

4. Management time. Yours, and your team’s. Interviewing, onboarding, reviewing, correcting, and absorbing the work the departed person used to absorb.

Add those together, multiply by your actual annual departure rate, and compare the result against the difference between the cheapest provider and the one that keeps people. In our experience the comparison is rarely close.

A worked example

Take a technical support specialist at USD 2,500 a month all-in, which is roughly what a senior one costs in Indonesia.

Recruitment might cost a month of that salary equivalent once you count sourcing and your own time. Ramp to full productivity on a complex product, call it four months at an average of 50% effective output, so two months of salary equivalent lost. Escalations into engineering during that period consume senior time that is worth considerably more per hour than the support salary. Management overhead adds more.

Without inflating anything, replacing that one person costs somewhere between three and six months of their salary. At 40% annual attrition across a team of five, you are replacing two people a year, every year, in perpetuity. That 40% is the contact centre case rather than the professional services one, and it is the rate a badly built offshore arrangement drifts towards.

The saving you captured by choosing a provider USD 300 a month cheaper is USD 18,000 over that period across the team. The turnover is costing more than that, and it is also costing you customer experience, which does not appear in any column.

What a resignation actually costs you in Indonesia

Most foreign employers assume that a departure in Indonesia triggers severance. On a voluntary resignation it does not, and the detail is worth knowing because it changes where the cost actually sits.

Under PP 35/2021, an employee resigning of their own accord has to meet three conditions set out in Pasal 36 huruf i: a written resignation request submitted at least 30 days before the resignation takes effect, no outstanding service bond (ikatan dinas), and continuing to carry out their obligations until the resignation start date.

Where those conditions are met, Pasal 50 entitles the employee to two things, and neither of them is severance:

• Uang penggantian hak, compensation of rights, defined in Pasal 40 ayat (4): annual leave accrued but not taken and not yet lapsed; the cost of transporting the employee and their family back to the place where they were hired; and anything else written into the employment contract, company regulations or collective labour agreement.

• Uang pisah, separation pay. PP 35/2021 does not set an amount: Pasal 50 points to the employment contract, company regulations or collective labour agreement, so your own documents govern it. Where those documents are silent on uang pisah, there is no fixed statutory sum to apply.

Uang pesangon (severance) and uang penghargaan masa kerja (long-service pay) are not payable on a voluntary resignation. Those attach to employer-initiated termination, where the ground applies a multiplier of 0.5 to 2 times the severance scale in Pasal 40 ayat (2), with long-service pay carrying its own separate scale under Pasal 40 ayat (3). Our guide to severance pay in Indonesia works through every ground.

Two practical consequences follow. The statutory exit cost of a resignation is small, so turnover in Indonesia costs you almost entirely through the replacement model above rather than through the exit payment. And the two variable items, repatriation and uang pisah, are both governed by documents you control: a generous company regulation drafted years ago is still binding on the day someone resigns.

That 30 days is a statutory minimum for the resignation to qualify, not a cap: employment contracts routinely set longer notice. It is the main reason a replacement realistically starts at least four to six weeks after you make an offer, even when the candidate is otherwise available, and it is why the ramp clock in the model above starts later than most plans assume.

Why offshore support churns, and what actually stops it

The causes are consistent, and pay is rarely the first one.

Isolation. Someone working alone from home for a company in another timezone, with no colleagues nearby and no local employer, has very little holding them. This is the most common cause and the least discussed, because it does not show up in an exit interview as clearly as money does.

No progression. A support role that leads nowhere loses its best people to roles that do. The strongest performers leave first, which is the expensive way round.

No local HR presence. When something goes wrong in someone’s employment and the only route is a support ticket to a platform in another country, small problems become resignations.

Fixed-term contracts. Employing people on rolling fixed-term contracts to preserve flexibility tells them exactly how much commitment to give back.

What works is unglamorous and mostly structural:

• Permanent employment. In Indonesia that means PKWTT contracts rather than fixed-term PKWT. It is also the compliant instrument for ongoing work.

• A physical place to work. Dedicated workspaces solve the isolation problem more directly than any engagement programme.

• Local HR who know the person. Someone reachable in their timezone, in their language, who can resolve a problem before it becomes a resignation.

• Benefits above the statutory floor. In Indonesia, meaningful supplementary cover above BPJS.

• Somewhere to go. Professional development that makes the role a step rather than a destination.

• Colleagues. Contact with other professionals in the same situation, which is the part most remote-first arrangements never provide.

What this looks like in practice

MixWork employs people permanently on PKWTT contracts, provides dedicated workspaces in Jakarta, includes Total Care 360 with every Employer of Record engagement at no extra cost, runs monthly check-in calls with both the employee and the client, offers Flex Benefits above the statutory floor, holds cultural activities with other professional remote staff, and runs professional learning sessions.

Twelve-month retention across our placements runs above 90%.

That figure deserves a caveat rather than a flourish. The 30 to 45% attrition benchmarks quoted earlier describe contact centre agents, typically on high-volume voice work and often on shifts. MixWork places permanent professional staff into product teams. Those are different populations doing different jobs, and part of the difference is the role rather than the model.

What it does demonstrate is that the churn most buyers treat as an unavoidable cost of offshore hiring is largely a consequence of how the arrangement is built. Build it differently and the number changes.

The question to ask a provider

Not “what is your attrition rate”, because everyone has an answer ready.

Ask these instead:

• What contract type do you employ people on, and why?

• Where do they physically work?

• Who in your organisation knows this person by name, and where are they based?

• What happens, specifically, when someone gives notice at month nine?

• What is your average tenure, not your attrition rate?

The answers separate providers who have thought about retention from providers who have thought about pricing.

Working out your own number

If you are building the business case, the cost calculator will give you an all-in monthly figure for a specific role, which is the input the calculation above needs. Our guide to Employer of Record in Indonesia covers the employment mechanics, and the article on hiring a technical support team in Indonesia covers what these roles cost and how to structure them.

Or book a free consultation and we will work through the numbers for your specific roles, including the ones that make the comparison less flattering to us.

Who we place, and who we do not

We hire top tier professionals with degrees from Indonesia’s top universities. The QS World University Rankings 2026 place Universitas Indonesia at 189 globally, Universitas Gadjah Mada at 224 and Institut Teknologi Bandung at 255, with 26 Indonesian universities ranked overall. Our placements average six years of professional experience, mostly inside multinationals and global agencies, and run from individual contributors through team leads to managers.

That is a different standard from a conventional call centre or seat based operation, and a different arrangement from a contractor, freelancer or virtual assistant. Those give you shared attention and no continuity. We place permanent employees who work only for you, with twelve-month retention above 90%.

Turnover hurts most where accumulated context is the product. That is why retention matters when you hire IT developers in Indonesia.

Disclaimer: This article is general information, current as at 15 September 2026, and is not legal advice. The Indonesian provisions referred to are in Government Regulation PP 35/2021 and are fact-specific. Confirm your own obligations with qualified Indonesian legal counsel before relying on them. MixWork provides EOR, HR, recruitment and managed IT services, not legal advice.