A Polish company or an employer of record — which and when

The real question

Not which is better — how long you need the option

Cost barely separates the two in year one. What separates them is what each commits you to, and how quickly you can undo it.

So the question is not "which is better". It is how long do you need to keep the option of walking away? Answer that and the structure follows.

What it does

What an employer of record actually covers

It employs someone on your behalf. The provider is the legal employer — contract, payroll, social contributions, statutory obligations — and you direct the work.

That solves a real problem. You get a person on the ground in weeks, with no company, no local payroll and no filing obligations. If it does not work, the exit is a contract termination rather than a corporate one.

For testing demand with one or two people, nothing beats it.

What it does not

What a person employed through an EOR cannot do for you

This is the part that rarely appears in the comparison, and it is where entries stall.

• They cannot sign on your behalf. They are not your employee and hold no corporate authority. Every contract goes back to head office
• They cannot be appointed to a Polish board or as a commercial proxy, because there is no company to appoint them to
• You still invoice from abroad. Some customers will not accept a foreign invoice, and public tenders and large corporates frequently require a local entity outright
• You cannot hold assets, stock or a lease in your own name in Poland

None of this matters while you are testing. All of it matters the moment someone wants to buy at scale.

The triggers

When an entity stops being optional

Five triggers. One is usually enough.

• You need to sign local contracts in your own name
• You need to issue Polish invoices
• You are bidding for work that requires a local supplier
• You are holding stock or assets in Poland
• You are past a handful of people and the per-head EOR fee has stopped looking like a bargain

Four of the five are commercial rather than financial. The entity decision is usually forced by customers, not by a spreadsheet.

The numbers

What each one costs

Published ranges for a mid-sized company, first twelve months:

Those two ranges are not the same comparison. One is a single person and no company; the other is a team and a legal presence. Read side by side they suggest the wrapper costs €70,000, and it does not. Almost all of the gap is headcount.

Where the crossover actually sits

At the same headcount the two structures cost far less differently than the headline ranges imply. The shape is what differs:

• EOR: gross salary, plus employer contributions, plus a provider margin — per person. Every cost scales with headcount and nothing is fixed
• Entity: gross salary, plus employer contributions, plus accounting, payroll and statutory reporting — whether you employ one person or ten

So the crossover is arithmetic. At €300–600 per person per month, three people cost roughly €11,000–22,000 a year in provider margin alone — which is the region where the fixed cost of running a small Polish company stops looking expensive. With one or two people the EOR is almost always cheaper. Somewhere between three and five it usually stops being so.

Two numbers settle it: ask any provider for the margin per person at your intended headcount, and ask an accountant for the monthly cost of running a small company. Nobody selling either option volunteers both.

The asymmetry

What it costs to change your mind

From an EOR to a company: straightforward. Register, transfer the people across, keep trading. The EOR contract ends on notice. This direction is cheap, which is precisely why it should be the default order.

From a company back to nothing: expensive. Liquidating a Polish limited company takes at least six months because of a mandatory creditor protection period, and usually longer. Throughout it the company still files and still pays its accountant. Add severance on employment contracts and the remaining term on any lease.

The asymmetry is the whole argument. One direction costs a notice period; the other costs half a year and a closing bill.

The answer

How to decide

Take the shorter of two horizons: how long before you know whether this market works, and how often your board revisits the plan.

If that horizon is under a year, start with an employer of record and register when a trigger forces you to. You lose a little efficiency and keep the option.

If it is longer, or if one of the five triggers already applies, register now and stop paying for flexibility you have already decided not to use.

What does not work is registering because it feels more serious. It is the one version of this decision that costs money in both directions.

Who wrote this

Bartlomiej Sedek

Operating Partner based in Warsaw. I run market entry and CEE hub operations for international mid-market companies on fixed-term mandates, and I take no commission from any provider I recommend — which is why this piece does not have a preferred answer.

For 13 years I ran a foreign principal's Polish operation as the person accountable on the ground: 16 people to 400, 11 markets, and three separate operations in Moscow, Istanbul and Hamburg consolidated into the Warsaw base.

How the mandate works, what it costs, and who it is for →

Written for companies of roughly €20–250M in revenue. Thresholds and timelines change; check anything you intend to rely on.