TL;DR
- Core decision: Stop treating benefits as a single global package. Treat them as a stack of four layers, only some of which travel.
- When doing nothing is right: When headcount abroad is small, footprints are stable, and local teams report no retention problem. The honest default is restraint.
- What has to be true: You can describe, for every country you operate in, what the state provides, what you must add, and what you choose to add. If you can't, you're not yet in a position to design anything.
- How the options split: Three honest categories, direct local entity, employer of record, and hybrid. Each one has a real failure mode. None is a default.
- Decision rule: Decide the question, then choose the setup. People do it the other way round and explain the result afterwards.
- Outcome to expect: Cleaner cost, fewer accidental promises, and a package that holds together when a regulator, an auditor, or a new joiner pulls on a thread.
The package that travelled too far
A benefits lead in a mid-sized software company opens a spreadsheet in the second week of January. The home-country package looks generous on paper, and the chief executive has asked her to roll something similar out in three new markets by the end of the quarter. The columns are tidy. The assumptions underneath aren't. In Berlin, a competitor is offering a private pension top-up that doesn't even register as a perk, because the statutory scheme already covers most of what the lead assumed she was providing. In São Paulo, a health plan the home office considers standard is, in practice, a luxury tier that nobody local uses. In Singapore, the same wellness allowance that the home team praises in their engagement survey is constrained by how payments are classified for tax purposes, and the structure that made it work at home can't be reproduced.
But the meeting on Friday isn't really about a spreadsheet. It's about what the company believes it owes to the people it's asking to work in places where the contract of employment is built by someone else, mostly the state. The deeper question isn't which benefits to send abroad. The deeper question is how to tell, in any given country, what is already there, what must be added, what is genuinely left to compete on, and which of those decisions are worth getting the leadership team into a room for. The real issue isn't what we offer globally. It's what is already provided here, what we must add, and what is left for us to actually compete on.
When doing nothing is the right answer
Some readers don't need a project. They need permission to wait. The honest stages look like four recognisable companies.
Best tools for Global HR
The single-market exporter. Headcount in any foreign country sits at a handful of people, hired through an employer of record or a local partner that runs the statutory piece. There's no office, no local manager with benefits authority, and no real churn problem. The home-country package has not been demanded, and the joiners who have asked have done so because they assume parity, not because the market expects it. At this stage, sending the home-country summary document to international hires is more dangerous than not sending anything, because it creates an accidental promise. The right answer is to document that no international benefits position exists, write down the trigger that would force a review, and put the file away. That's not negligence. It's a defensible position if anyone later asks why nothing was done.
The friction stage. A few local hires are starting to compare notes with peers at competitors, and the comparison doesn't flatter the company. Quit rates among the second-year cohort aren't catastrophic, but the reasons given in exit interviews all sit in the same place. Recruiters are telling candidates that the package is broadly comparable, and candidates are doing the arithmetic and finding it isn't. The company doesn't have a crisis. It has a small but real liability, the kind that gets worse quietly. A modest diagnostic, a conversation with three or four credible local sources, and a written position that names the gap is usually enough to recover. The risk at this stage is overcorrecting and building a programme that costs more than the problem.
The real risk stage. The company is hiring in volume, often in a market where the state provides a great deal and a long tail of employers competes for the same small pool. The talent market is tight, and candidates are asking sharp questions about pensions, leave and family policy before they take a call. The internal equity question has started to bite: home-country staff have noticed that a colleague in a low-cost city is getting a different shape of package, and the explanation is being asked for. This is the stage where leadership is right to commission a review, and wrong to ask the review to do anything more than name the layers.
The edge case. The company has hit a regulatory or tax question it can't defer. A permanent establishment exposure has been raised by finance. A worker classification issue in California, a UK off-payroll question under the post-April 2026 thresholds, or an EU platform worker question has surfaced in a market the legal team has not previously been asked about. At this point, the benefits question is downstream of a compliance question, and the right move is to resolve the legal exposure first, then return to the package. Designing benefits around a structure that may need to change is a waste of everyone's time.
Five questions a benefits lead asks at 11pm
Are we even a global benefits employer yet? If you can't name, by name, every country where you've people on payroll, and can't tell finance how the package differs in each, the honest answer is that you're not. You're a domestic employer with some international hires. That's a real position. It just isn't a global benefits position. The mistake is to act as if the question has already been answered. The cost of pretending is that every later decision is built on a category error.
What does the state already provide here? Not in the abstract, and not in a brochure. In this country, for a worker of this kind, on this kind of contract, what is the statutory floor, and how is it administered. If you can't answer that for every market, the design will be wrong in ways you can't see from headquarters. The wrong answer is to assume the home-country model is the maximum. Sometimes the state is the maximum. Sometimes it's the floor. Most often it sits somewhere in the middle, and you need to know exactly where before you add anything on top.
Which of our existing benefits are silently mandatory here? A benefit treated as discretionary at home can be a statutory entitlement abroad. When that happens, the cost doesn't disappear. It just gets reclassified. The company is already paying, through payroll, through social contributions, or through the local partner that runs the package. Treating an existing obligation as a perk is a category error that compounds quietly. The wrong answer is to claim credit for what the state already does.
What is the candidate market actually rewarding? Not what the home-country team would value if they moved. What people in this market, comparing offers from employers in this market, actually pick. The wrong answer is to assume that the home-country differentiators travel. They rarely do, and the ones that do are usually the ones nobody at headquarters would have predicted.
Are we designing for the people we've, or the people we wish we had? The same question dressed up. The package should fit the workforce, the headcount plan, and the hiring pattern that's actually happening. Designing for a future where the company is twice the size in a market it has not yet entered produces a programme that costs money now and delivers nothing. The wrong answer is to build a structure that only pays off in a scenario the leadership team has not committed to.
The three honest categories
Direct local entity. The company incorporates in the country, employs people on its own payroll, and builds a benefits programme that sits on top of the local statutory floor. The legal employer is the company itself, which means obligations run to the company directly. This is the right answer when the headcount is large enough to absorb the cost of setting up and running a local operation, when the market is strategic and the company is willing to stay, and when the package needs to do more than the statutory minimum, because there's something genuinely worth competing on. The failure mode is fixed cost. A direct entity carries overheads that only make sense above a certain headcount, and benefits design is only one of the decisions being made in that structure. If the market turns out to be a three-year experiment, the local entity is a stranded cost, and the package is part of the wreckage.
Employer of record. The EOR employs the worker through its own local entity, running local payroll, statutory filings and employee-level obligations. From the worker's perspective, they're on a local contract with a local employer. From the company's perspective, the structure is a service. This substantially reduces the overhead of going into a market, and it changes what is variable and what isn't. The EOR is the legal employer and provides the statutory package through its own local entity, which constrains what the company can vary. You can top up. You can add a flexible layer. You can't redesign the floor. The right answer is when the headcount is too small to justify a direct entity, when the market is genuinely exploratory, or when speed matters more than customisation. The failure mode is the assumption that the EOR has absorbed the compliance question. It has absorbed a great deal, but not all of it, and not the parts that sit outside the employment relationship.
Hybrid. Some of the structure is run directly, some is run through one or more partners, and the split is usually drawn around the question of who is the legal employer in each country. This is the most common position for companies of the size the reader works in, and it's the one that produces the most mistakes, because the seam between the two structures is where promises get made twice and obligations get missed. The right answer is when the workforce is genuinely mixed, when some markets are strategic and others are opportunistic, and when the leadership team is willing to be honest about which is which. The failure mode is the attempt to run a hybrid without naming the rule that decides which side of the line a given country sits on. The rule is usually headcount and commitment. Write it down, and write down the date at which the rule is reviewed.
Five diagnostic questions you can answer
Can you name every country in which you've people on payroll, and what they get? Not a region. A country. Not a summary. The actual entitlements. If the answer is no, that's the first piece of work, and it has to be done before anything else is designed on top of it.
Do you know, for each country, which entitlements are statutory, which are common in the local market, and which are genuinely a choice you're making? To answer this, you need a source on the local market that you trust. That source isn't a benefits broker, who has an interest in the answer. It's a local employment law adviser, a peer benefits lead in the same market, or a credible industry survey from a body you would cite in a board paper. The way to check the answer is to ask two independent sources and to compare what they say.
Can you explain, in a sentence, why the package looks the way it does in each market? If the answer is "we copied the home-country package and translated the document", that's a real answer and a problem at the same time. The reason should be specific to the market, and it should be something the leadership team would still agree with if it were explained to them cold.
Have you asked the people in each market what they would change, and have you believed them? Local teams know what is missing. The home-country team often doesn't. The mistake is to treat the local team's view as one input among many, rather than as the input most likely to predict retention and engagement. The way to check the answer is to read the exit interviews from the last year and to see whether the same thing comes up more than once.
Are the things you offer on top of the statutory floor the same things in every market, or have you decided, deliberately, that they should be? This is the question that separates a global benefits position from a domestic one with a translation layer. The right answer isn't necessarily that they should be different in every market. The right answer is that you've decided, with eyes open, and that the decision is written down.
Four Layers of a Global Package
Statutory entitlements you must provide
This is the floor. In every country where you employ people, there's a set of entitlements that the state requires, that local law enforces, and that no amount of internal policy can replace. They cover the things people actually need: leave, sick pay, social contributions, pensions where they exist, family and parental entitlements, notice periods and severance where applicable, and any locally required insurances. The work isn't in designing this layer. The work is in knowing it precisely, in every market where you've people, and in being honest about the fact that what looks generous at headquarters may already be the minimum locally.
Where this layer falls short is that it's the floor, and treating it as the package is the fastest way to lose credibility with both local staff and the regulators who administer it. The other failure mode is treating a statutory entitlement as a perk. Doing so buys no goodwill, because the worker isn't receiving something extra. It also creates a real risk the day the statutory scheme changes, because the package as described no longer matches reality.
Near-universal market practice that is technically optional
In most markets, there's a layer above the floor that almost every credible employer offers, that candidates expect without being told to expect it, and that's not strictly required by statute. This is the layer where the home-country team most often misreads the situation, because what is "market" abroad reads as "generous" at home. A private medical top-up, an additional pension contribution, a defined number of well-being days, a flexible benefit allowance: these are all things that may sit in the second layer in some markets, even when they read as differentiators in the home market.
Where this layer falls short is that competing on it's a slow route to nowhere, because everyone is offering roughly the same thing, and the cost is largely fixed. The mistake is to over-invest in this layer, in the hope that it functions as a differentiator, when in practice the workforce has already priced it in.
Genuine differentiators
This is the layer that actually wins candidates and keeps people. It's also the layer that's hardest to design, because by definition it isn't what everyone else is doing. Genuine differentiators look like specific, concrete things that are tied to a particular workforce in a particular market, and that the home-country team wouldn't have arrived at on its own. They might be a parental leave policy that goes further than the local norm, a sabbatical that the workforce can actually use, a learning budget that's large enough to matter, a relocation package that's honest about what the company is willing to do.
Where this layer falls short is that differentiators only work if they're credible, and credibility is local. A benefit that reads as progressive in one market can read as strange or performative in another. The other failure mode is the differentiator that nobody uses, because the home-country team designed it for a workforce that doesn't exist in the local market.
Global-consistency benefits that only work if offered everywhere
Some benefits only make sense as global promises, because the workforce notices the seam when they're not. Equity, in some form, is the obvious example, but it isn't the only one. The principle is the same: if a benefit is offered in one country and not in another, and the reason is operational rather than principled, the workers on the wrong side of the line notice, and the package as described becomes harder to defend.
Where this layer falls short is that global consistency is expensive, and the cost is often higher than the goodwill it produces. The other failure mode is the global benefit that's genuinely offered everywhere but functions badly in the markets where the underlying infrastructure isn't there to support it. The answer is usually to be honest about which benefits belong in this layer and which don't, and to write down the reason.
The Decision Table
| Situation | Scale | Setup | Primary Pain | Recommended Starting Point |
|---|---|---|---|---|
| Exploratory market, headcount under twenty, no clear plan to grow | Small | EOR or local partner | Designing a programme nobody will use | Document the statutory floor, write the trigger for a review, do nothing else |
| Established market, headcount between twenty and one hundred, real retention pressure | Medium | Hybrid, partner for payroll, direct on top-up | Inconsistency between what is promised and what is delivered | Build the country benefits map, identify the second layer, decide what to add |
| Strategic market, headcount over one hundred, leadership committed | Large | Direct local entity | A package that does not match what the market actually rewards | Full four-layer design, with the differentiators identified by the local team |
| Workforce of mixed contract types, including independent contractors in regulated markets | Mixed | EOR plus classification review | Worker classification risk in California, the UK, or the EU | Resolve classification first, then return to package design |
| Workforce mostly through an EOR, but with one or two senior hires on a bespoke local contract | Small with edge cases | EOR plus direct for the edge cases | The seam between the two structures, and the accidental promises made across it | Write down the rule for who sits on which side, and the date it is reviewed |
| Recently reorganised, finance raising a permanent establishment question | Any | Existing setup, under review | Tax exposure that may force a structural change | Defer benefits design until the legal question is resolved |
| Remote workforce spread across many countries, none with a significant headcount | Distributed | EOR everywhere, or some mix | Equity between countries that is not actually being delivered | Decide what is genuinely global, write down what is not, and stop promising the rest |
| A new market being entered through acquisition, with legacy staff on legacy terms | Any | Inherited complexity | Promises made by the previous employer that may or may not be honoured | Inventory what the legacy staff actually have, then decide what to standardise |
Building a Country Benefits Map
A country benefits map is a working document, not a brochure. It sits in a shared folder, gets updated when something changes, and is the first thing anyone in the benefits function opens when a new question comes in. For every country the company operates in, the map has to answer four questions in a way the leadership team could read and challenge: what does the state provide, what does the company add, what is genuinely left to compete on, and what has the local team said is missing. The map isn't the design. It's the input to the design, and the design that doesn't start from a map is usually a guess dressed up as a strategy.
The information to build the map rarely sits in one place. The statutory floor comes from local employment law, and the source is either a credible law firm in the market or a peer benefits lead who has actually been through the process. The market-practice layer comes from recruiters who place people in the same market at a similar level, from local industry bodies, and from candid conversations with people the company has already hired. The differentiator layer comes from the local team, not from headquarters. The global-consistency layer comes from the company's own values and from a willingness to be honest about which of those values will actually cost something in practice.
| Source | What it tells you | What it does not tell you |
|---|---|---|
| Local employment law adviser | The statutory floor, recent changes, common mistakes | What the market actually rewards |
| Local recruiter at the right level | What candidates compare, what they ask about, what they walk away from | The statutory floor in detail |
| Peer benefits lead in a non-competing firm in the same market | What the second layer looks like, what has stopped working | Anything specific to your workforce |
| Local team, in a structured conversation | What is missing, what is misunderstood, what is being promised informally | Anything the local team is not yet in a position to see |
| Existing staff, in an exit interview or a stay interview | The seam between the package as described and the package as experienced | Anything that the workforce has stopped mentioning because they assume it will not change |
The map is only useful if it's updated, and it's only updated if it has an owner. Pick a person, write down the review cadence, and write down what triggers an off-cycle review. The triggers are usually a regulatory change, a competitor doing something notable, a churn spike in a particular market, or a finance question about cost. Confirm every entitlement locally before publishing it, because the cost of being wrong on a statutory entitlement isn't a footnote.
Equity and Global Consistency
Equity is the promise that's hardest to keep uniformly, because the floor under it's different in every market. Equalising total cost in every country produces a package that's generous in low-cost markets and inadequate in high-cost ones, and the workforce notices both directions. Equalising total value produces a different problem, because the conversion from one country's benefits into another's isn't a real conversion. A pension top-up in one country isn't a health top-up in another, and pretending that it's creates a category of promise the company can't keep.
The honest position is that some benefits are global promises, because the workforce expects them and the company is willing to fund them everywhere, and some are local choices, because the underlying entitlement is shaped by the state and the company's role is to add on top. The line between the two is the design question, and it's one the leadership team has to make explicitly.
| Type of benefit | Travel well? | What to watch |
|---|---|---|
| Core health provision, as the local market defines it | No, by design | A global standard set too low leaves high-cost markets exposed |
| Parental and family leave, beyond statutory | Sometimes | The market comparison is more important than the absolute number |
| Learning and development budget | Yes, in principle | The actual delivery has to be possible in each market |
| Equity, in some form | Yes, in principle | Tax and securities law make this harder than it looks, take local advice |
| Well-being allowance, as a flexible credit | Yes, in principle | Tax treatment varies, and the shape of what is permissible is not uniform |
| Retirement or pension top-up | No, by design | The local statutory scheme is part of the picture, and any top-up is layered on top of it |
| Death and disability cover, beyond statutory | Sometimes | Local insurance markets are not interchangeable, and the same named benefit can mean very different things |
| Remote working stipend | Yes, in principle | Tax treatment is the trap, confirm locally |
The most common failure on global consistency is the benefit that's offered everywhere in name and nowhere in practice, because the local infrastructure isn't there to support it. The second most common is the benefit that's offered in the home market and translated into local terms without checking whether the translation works. The way to test either failure is to ask a worker in a market the home-country team has not visited recently, in their own language, what they would change.
The cost of getting this wrong
The cost that appears on an invoice is the cost the company is at least prepared to defend. The second-order costs are the ones that arrive later, in places the original design never imagined. There's the cost of the accidental promise, where a candidate was told something in a recruitment conversation that the benefits lead has never heard of, and that promise is now part of the implied contract. There's the cost of the well-meaning copy, where a document was translated without being checked, and the translation says something the original didn't. There's the cost of the legal exposure that was inherited along with a market entry, where the package as designed assumes a structure that the regulator is no longer willing to accept. And there's the cost of the workforce that has quietly stopped mentioning the things that are wrong, because they have already decided that the company isn't going to change.
So the cost of getting this wrong is the moment, two years from now, when a regulator, a competitor, or a single determined employee pulls on a thread, and the package comes apart in public. The cost of getting it right is the absence of that moment, which is harder to see and easier to underfund. The real question isn't what we can afford to spend on benefits this year. The real question is what we are quietly committing to in every market where our name is on a contract, and whether we are willing to be honest about it.
When you are ready to go further
If the picture above matches the question you're sitting with, the next step is comparison work, not vendor selection. HROpsLab is a review publication, not a vendor, and we sell nothing. Our independent comparisons exist to help reward and people leads build a shortlist on the basis of fit, not on the basis of who showed up first in a search. We don't supply software, we don't supply consulting, and we don't supply legal advice. What we do is read the materials, ask the questions, and publish the answers in a form the leadership team can challenge.
When you're ready, the place to start is the comparison work itself. Read it cold, take it into the room where the decisions are made, and decide what to do with it.
Frequently Asked Questions
Should we offer the same benefits everywhere?
No, and the longer you try, the more the package will drift from what each market actually rewards. Offer the same overall approach everywhere, in the sense that you can explain, in each market, why the package looks the way it does. What the package contains will vary, because the floor under it varies. The mistake is to treat consistency of appearance as the goal, when the goal is consistency of method.
How do we find out what is mandatory in a market we are entering?
Ask a local employment law adviser, in writing, and ask them to name the statutory entitlements by category, not by document. Cross-check with a peer benefits lead in the same market at a similar headcount. Confirm the answer with a local recruiter who places people at the level you're hiring. None of the three sources is sufficient on its own, and all three together are still not a substitute for your own document. Confirm every entitlement locally before publishing anything that depends on it.
Can an EOR package be topped up?
Yes, in principle. The EOR is the legal employer and provides the statutory package through its own local entity, which constrains what you can vary on the floor, and you can add a flexible layer on top. The trap is the assumption that topping up makes the EOR the answer to every question. It's the answer to the employment-law question, within limits. It isn't the answer to corporate tax, and it doesn't replace the need to take local advice on anything that sits outside the employment relationship.
What should we do about pensions?
Treat pensions as a market-by-market decision, with the local statutory scheme as the floor and any company contribution as a layer on top. The shape of that layer depends on the local market, the tax treatment of the contribution, and what candidates at the level you're hiring actually compare. The mistake is to design a global pension promise and then discover, market by market, that the structure that delivered it at home isn't available abroad. Confirm the position locally before committing.
How do we handle equity across countries?
Decide, first, whether equity is a global promise or a local choice. If it's global, fund it everywhere, and take local advice on the structure in each market, because tax and securities law make the mechanics harder than they look. If it's local, write down the reason, and accept that the workforce in markets where the package is more limited is going to ask why. The honest answer is usually some combination of the two, and the design is only as good as the willingness to be specific about which is which.
Should we equalise total cost or total value?
Neither, in pure form. Equalising total cost produces a package that's inadequate in high-cost markets and excessive in low-cost ones, and the workforce notices both directions. Equalising total value requires a conversion that doesn't really exist, because the underlying entitlements aren't interchangeable. The honest position is that some categories are global promises, and within those categories, the company funds to a level that's credible in the highest-cost market, and the rest is local choice. Write down which is which, and write down the reason.
What do we do when the state already provides something we would otherwise offer?
Stop offering it as a perk. The state provision is part of the floor, and treating it as an extra buys no goodwill, because the worker isn't receiving something additional. The right answer is to redirect the budget that would have gone to that benefit into something the state doesn't provide and the workforce actually values. The mistake is to keep the old line in the document because it's easier than rewriting it.
