Each provider gets our score from 0 to 10 for each service it offers: a weighted mean of five criteria. This page sets out the criteria and their weights, what stays out of the score, how prices and coverage are shown, and how every list on the site is ordered.
Short answer
- Our score is a weighted mean of five criteria: compliance 30%, coverage 20%, price 20%, platform 15% and support 15%. Each service a provider offers is scored on its own.
- Ratings on G2, Capterra and Trustpilot are shown next to our score, with the date we checked them and a link. They are not part of it.
- Prices are shown as each provider publishes them, in their own currency and unit. We never convert currencies.
- Catalogs and lists are ordered by our score. Where an editor can fix a place, this page says so, section by section.
On this page
How we score
We score each service a provider offers separately, so a provider with an Employer of Record service and a contractor payouts service has two scores. Each of the five criteria gets a value from 0 to 10 and a short reason. Our score for the service is the weighted mean of the five values, rounded to one decimal.
| Criterion | Weight | What it covers |
|---|---|---|
| Compliance | 30% | The provider’s legal footing: its entities on company registers, whether the service agreement is public, its own security certifications and its licences. |
| Coverage | 20% | Where the service works: the countries it documents and claims, against the widest provider of the same service, and for employment services its own registered entities. |
| Price | 20% | What the service costs as published: the “from” price ranked among providers of the same service, then setup fees, deposits, minimum terms, termination fees and billing terms. |
| Platform | 15% | The software you use: API, integrations, self-serve onboarding, the documents it produces and, for contractor services, payout methods. |
| Support | 15% | How the provider helps you: live support channels and hours, a dedicated manager, stated onboarding times and a stated response commitment. |
The weights add up to 100%. Each value is computed from the facts we hold on the provider, by fixed rules set out below. A provider page shows, for each service, every criterion with its weight, its value, its contribution to the score and the reason. The header of a provider page shows the score of its main service: Employer of Record when the provider offers it, otherwise the first service it lists.
This page describes method version 2026-10c. It reads the weights and the thresholds from the same code the score is calculated with, so the two cannot differ.
How each criterion is scored
Each criterion reads only facts we record from a source, such as the provider’s pricing page, a company register or its security page. Price and coverage compare the service with every other provider of the same service. The points per criterion:
| Criterion | Inputs | Points |
|---|---|---|
| Compliance | Registry-evidenced entities (own entities in the service’s countries plus the provider’s own register entry); the published contract; named security certifications; licences with a register link or a named issuing authority. | Entities, employment services: 6+ → 3 · 3+ → 2.5 · 1+ → 2 · none → 1.5; other services: a registered contracting entity 3. Service agreement public 2, only general terms or nothing public 0, not checked 1. SOC 2 and ISO 27001 1.25 each, one other standard +0.5; one linked only to a homepage counts 50% of that; none 0. Licences: 2+ → 2 · 1+ → 1.5 · none → 1. |
| Coverage | Sovereign states on the service’s country list with register or provider-site evidence, plus the countries it claims; for employment services, own entities on a company register. | Documented countries plus 0.5 × each claimed-only country, as a share of the widest provider of the same service: 80%+ → 10 · 60%+ → 8.5 · 40%+ → 7 · 20%+ → 5.5 · above 0% → 4; no list or claim 5; claims only, capped at 7. Employment services: 0.75 × that value plus own registered entities as a share of documented countries, 50%+ → 2.5 · 20%+ → 2 · 5%+ → 1.5 · above 0% → 1.25, none 1. |
| Price | The published “from” price (or “up to” price) with its unit and currency; setup fee, deposit, minimum term, termination fee; annual billing. | No published price 3. Custom terms, a price in several recurring or per-transaction parts, or fewer than 5 other prices of the same unit to rank against, 5. Otherwise by the share of providers priced lower: up to 20% → 9 · up to 40% → 8 · up to 60% → 7 · up to 80% → 6 · up to 100% → 5. Then each term stated as zero +0.5, a charge −0.5 (a deposit over a month or a minimum term of 12 months or more −1), a price that needs annual billing −0.5. |
| Platform | API; number of integrations; self-serve onboarding; contracts, IP assignment and closing documents in the platform; payout methods (contractor services). | API 2.5; integrations 50+ → 2.5 · 20+ → 2 · 5+ → 1.5 · 1+ → 1, none 0; self-serve 2, sales-led 0.5; documents 0.5 each; payout methods 3+ → 1.5 · 2+ → 1 · 1+ → 0.5. A part not disclosed earns half its maximum. The sum is scaled to 0–10. |
| Support | Live support channels (chat, email, phone, messengers, video, tickets); support hours; a dedicated manager; stated onboarding times; a stated response target (an SLA, or an aim such as “we aim to respond within 24 hours”). | Live channel kinds 4+ → 3 · 3+ → 2.5 · 2+ → 2 · 1+ → 1.5 · none listed → 1.5. Live 24/7 3, other stated hours 2, not stated 1.5. Dedicated manager 2, not stated 1. Onboarding up to 5 days → 1 · up to 14 days → 0.75, longer or not stated 0.5. Response target stated 1, not stated 0.5. |
Facts not disclosed
A fact the provider does not disclose gets a neutral value, not a zero, and the criterion’s reason lists it. Two gaps a buyer can see for themselves score low instead: no published price, and no named security certification. Each reason ends with how many of the facts the criterion looks for were disclosed.
Unverified facts
A fact we could not confirm on a public page, for example one read only in a search snippet or stated to us privately, stays on record but is not counted in the score and is not stated on our pages until a public source confirms it. Claims that a provider takes on liability, guarantees or indemnifies are never scored, and neither are regulations such as GDPR, which are laws rather than certifications.
Editor adjustment
An editor can move one criterion of one service by at most 0.5 points either way, with a written reason and a source. Several adjustments on one criterion add up to no more than that. The reason is shown with the criterion on the provider page.
Limited data
When fewer than 50% of the facts the five criteria look for are on record for a provider, across all its services, its scores carry a “Limited data” note. Such scores lean on neutral values and may change as facts are added.
What is not in the score
Ratings from G2, Capterra, Trustpilot and other review platforms are not part of our score. Neither are Reddit threads or the AI summary. We show each platform rating next to our score with its number of reviews, the date we checked it and a link to the platform’s page. We copy each figure by hand and never combine ratings from different platforms into one number.
On alternatives pages, a provider rated higher than the page’s provider on a platform gets a line saying so, but only when that platform rates both on the same scale and shows at least 50 reviews for each. That line does not change the order.
Prices
We show each price as the provider publishes it: the amount, the currency and the unit, such as per employee a month, per contractor a month or a percentage of each payout. Where a provider ties its lowest price to annual billing, the price says so. A provider that publishes no price shows “Custom quote” and stays in every list it qualifies for.
We never convert currencies. Prices are compared only within one group:
- per person a month, in one currency;
- a percentage of the payout, whatever the currency;
- other pricing terms, which are listed by name, not by amount.
When a catalog is sorted by price, the groups come in that order, US dollar prices before other currencies, and providers with no published “from” price come last. A comparison page marks the lower price only when both prices share the currency, the unit and the billing term. A best-of list ordered by price includes only prices per person a month in the list’s currency.
On a country’s Employer of Record page, the EOR fee in the cost table is the median “from” price of the providers that price that country in its own currency. Providers that price it in another currency are left out of the median, not converted.
Coverage
A provider’s country list and the countries we can document are not the same thing, so for each service we show two numbers:
- Claimed: the number of countries the provider states.
- Documented: the countries on its list backed by a company register entry or by the provider’s own website, each with a link and the date we checked it. A country listed without such evidence is shown but not counted.
For each country we also show the entity model where the provider states it: its own entity, a partner’s, or both. Where it does not say, we show “Not disclosed”, never a zero.
On country pages, a provider whose Employer of Record service covers the country is marked “Works in” that country for its other services too (Contractor of Record, Agent of Record, contractor management, contractor payouts and global payroll), even where its list for that service does not name the country; such a mark carries a note that it is inferred from the Employer of Record coverage.
When a provider’s country list is longer than the number it claims, for example a payout network that lists every country it can pay into, the provider page header shows the provider’s own claim with a plus, such as “150+”, instead of our count.
How lists are ordered
Catalogs
Each service has a catalog: Employer of Record providers on the providers page, other services on their own pages. Every provider that offers the service is in it. The default order is our score for that service, highest first; equal scores go by name. You can sort by price, within the price groups above, or by name, and filter, for example by country, entity model, price band or reach. The filter keeps its state in the page address. The catalog template lets an editor fix a provider at a given place; the others then move down one place.
Best-of lists
A best-of list starts from every provider that offers its service and keeps those that meet its scenario: company size, hiring pattern, industry, country or a maximum price. A provider whose data does not show that it meets a condition is left out. The rest are ordered by our score, or by the lowest published price for a price list. A list can set its own weights for the five criteria for its scenario; it then shows them next to ours. An editor can fix a place in a list and add a “why here” line. The list names fixed places under “How this list is built”.
Alternatives
The alternatives to a provider are the other providers of its main service, ordered by our score, highest first; equal scores go by name. Platform ratings never change the order. On an alternatives page an editor can move providers to the top, and the page names them above its table. The right-hand column on provider and reviews pages shows the first three.
Country rankings
A country’s Employer of Record page ranks every provider whose EOR service covers that country, by our EOR score, with equal scores by name. Editors can add a note to a card but do not choose who is ranked or in what order.
Reviews pages
A provider’s reviews page shows what each review platform publishes about it, quoted with the date we checked and a link: the rating and the number of reviews and, where the platform shows them, the star distribution, the share of recent reviews, reviewer company sizes, sub-ratings, the platform’s own labels and warnings, and what it shows about where reviews come from and how the provider replies to negative ones.
Review themes and Reddit threads are our selection, not all reviews. Each theme names the platforms it comes from, and each thread is linked with its date. We do not judge whether individual reviews are genuine and draw no conclusions about why they were written.
AI summary
Some provider, reviews and comparison pages carry an AI summary: short answers to questions buyers ask, written by an AI model from the data on our pages. Each set shows the date it was written, and an answer can note what the data does not show. The AI summary is not our score, not a recommendation and not a quote from any AI assistant.
Sources and dates
Figures such as contribution rates, minimum wages, notice periods and provider prices carry a numbered footnote. The sources are listed at the end of the page, each with the date we checked it. Platform ratings, Reddit readings and coverage evidence carry their own checked dates. The line under each page title shows when the page was last updated.
Rules and prices change. A figure is correct as of the date shown next to it; check it with the source or the provider before you rely on it.
Corrections and updates
If you find an error, email corrections@employer-of-record.io with the page address, the fact that is wrong and a source. Many pages end with a “Report an error” link that opens this email with the page address in the subject.
Providers who want to update their data can write to hello@employer-of-record.io. Changes go through the same method as everything else on this page. See also About and Contact.