n8n, Make and Zapier all connect apps and automate the steps between them, but they differ in three ways that matter to a recruitment agency: how they charge (per task, per credit or per run), where your data can live, and how they handle failures. There is no universal winner. The right choice depends on your volume, on who will maintain the workflows and on the data rules you work under. This guide compares the three using each vendor's own documentation, checked in October 2026, shows a worked usage example, and ends with a way to decide in an afternoon. Where a figure comes from a vendor's marketing page, the guide says so, and it contains no affiliate links.
The short version
| Tool | Tends to fit when | Tends to fit less well when |
|---|---|---|
| Zapier | You want the fastest start, a very large app catalogue and a simple step-by-step builder | Volume is high and every action is billed as a task, or you need EU-only data storage |
| Make | You want visual branching and data handling, moderate cost and a choice of US or EU hosting | You want to self-host, or your workflows are simple enough that the extra flexibility adds nothing |
| n8n | You want control over hosting, per-run billing and custom API calls | Nobody on the team can maintain it, or you want a fully managed service with no technical upkeep |
These are judgements based on how each tool is built and priced, not benchmark results.
How each tool charges
The billing unit is usually the biggest cost driver, because it decides how the bill grows as workflows get longer.
| Zapier | Make | n8n | |
|---|---|---|---|
| Billing unit | Task | Credit | Execution |
| What counts | Each successful action step. Triggers, Filter, Paths, Formatter and Delay steps do not count | By default one operation equals one credit, and some AI features use more | One run of the entire workflow, however many steps it has |
| Free option | Free plan with 100 tasks a month, limited to two-step Zaps | Free plan with 1,000 credits a month | Self-hosted Community Edition, where you pay for the server |
| Paid entry, as listed | Professional from 19.99 USD a month | From about 9 USD a month at 10,000 credits, per Make's own comparison page | Starter at 20 EUR a month billed annually, with 2,500 executions |
Prices are in different currencies, change often and depend on billing terms and tiers, so confirm them on the Zapier, Make and n8n pricing pages.
To see how the unit changes the bill, take one workflow that runs for every new candidate and performs six actions: create a record, add a tracker row, send two emails, notify a recruiter and update the ATS status. Assume the Make version uses a trigger plus six modules at one credit each.
| New candidates a month | Zapier tasks | Make credits | n8n executions |
|---|---|---|---|
| 300 | 1,800 | 2,100 | 300 |
| 3,000 | 18,000 | 21,000 | 3,000 |
On the same assumptions, Zapier's free 100 tasks cover about 16 candidates a month, and Make's free 1,000 credits cover about 142. Real counts will differ, because Zapier does not bill filter, formatter or delay steps and Make's counting depends on the module. Treat this as a way to see how the unit changes the shape of the bill, not as a quote.
Where your data lives and which contract applies
Candidate CVs and notes are personal data, so the location and the contract matter.
| Zapier | Make | n8n | |
|---|---|---|---|
| Hosting | Zapier's privacy page says data is hosted on AWS servers in the United States and that an EU-only storage option is not supported | You choose a US or EU data centre when you create an organisation, and cannot change it afterwards | n8n's cloud, or your own server in the region you choose |
| Contract | A data processing addendum that includes Standard Contractual Clauses. Zapier says it cannot sign other companies' DPAs | A data processing agreement naming you as controller and Make as processor, with Standard Contractual Clauses as the fallback for transfers | A data processing agreement forms part of the terms for Cloud customers. For self-hosted instances n8n says it is neither controller nor processor |
None of this makes an option unlawful. Transfers outside the EU can be lawful with the right safeguards, and each vendor describes its own. But if a client contract or your own policy requires EU-only storage, Zapier's documented position does not fit, while Make's EU region and a self-hosted n8n instance do. If a provider processes candidate data for you, Article 28 of the GDPR requires a contract that binds it to act only on your documented instructions. This is general information, not legal advice, and it is worth confirming your setup with a qualified adviser.
How each tool handles failures
Every automation fails sometimes. The question is whether anyone finds out and what happens to the failed record.
| What it offers | |
|---|---|
| Zapier | Replay of errored steps on every plan. On paid plans, autoreplay retries failed steps up to five times, and replays must happen within 60 days. Replayed steps use tasks. Custom error handling adds an error path to a step, on Professional and above |
| Make | Error handlers on a module's error route. Options include replacing a failed output so the run can continue, storing a failed run as an incomplete execution to review and retry, and committing or rolling back changes in apps that support transactions |
| n8n | An error workflow that starts with the Error Trigger node and can alert you, plus a Stop and Error node to fail on purpose when data is bad |
Whichever tool you choose, decide before going live what happens to a failed candidate record: who sees it, how it is retried and how duplicates are avoided. A silent failure in a recruiting workflow usually shows up as a candidate who never got a reply.
Fit with recruiting work
| Situation | Likely fit | Why |
|---|---|---|
| Connecting mainstream tools through ready-made connectors | Any of the three | Check the specific connector for each of your tools. Zapier cites more than 9,000 apps on its pricing page and Make's comparison page cites over 3,000, but vendors count differently, so list your tools and check each one |
| High volume of runs with many steps each | n8n, or Make with careful design | Per-run billing and credit use grow differently from per-action billing |
| Complex branching and reshaping of data | Make or n8n | Both are built around a canvas for routing and data handling, though Zapier also offers Paths and Formatter steps |
| A non-technical owner who wants no servers | Zapier or Make | Both are managed cloud services |
| EU-only storage requirement | Make's EU region, or self-hosted n8n | Matches the documented hosting options above |
| A tool with an API but no connector | n8n or Make | All three can call APIs, but check how much custom work each one needs |
A way to decide in an afternoon
- List your five most valuable workflows and the tools each one touches.
- Check the connector for every tool in each platform, and note where you would need a custom API call.
- Count steps and monthly runs, then run the usage arithmetic above for each platform.
- Write down your data rules. Is EU-only storage required? Which contract applies? Who is the processor?
- Decide who maintains it. If nobody technical will, favour a managed cloud service.
- Build the same simple workflow in two tools on free plans, using made-up candidate data. Compare the time to build, how clear the errors are and the cost at your volume.
- Choose for the next twelve months and write down what would make you switch.
Mixing tools
Many teams use more than one tool, for example a simple connector in one and the heavy workflows in another. The cost is more places to look when something fails, so keep an inventory of every workflow, its owner and its failure alerts. Because workflows are built in each tool's own format, moving later means rebuilding, so make the choice before you have built hundreds.
The same comparison works outside recruiting
The four criteria here (billing unit, data location, failure handling and who maintains it) apply to any business automation platform choice, whether the workflows handle candidates, invoices, client onboarding or support requests. Run the same usage arithmetic, read the same contract terms and build the same test workflow, and the answer usually becomes clear.
Common mistakes
- Choosing by app count. What matters is whether your own tools have good connectors.
- Comparing headline prices. Convert each plan to your own monthly runs and steps first.
- Treating vendor comparison pages as neutral. Each of these vendors publishes comparisons that favour itself, so check the claims against pricing and documentation pages.
- Ignoring failure handling. The first silent failure is an expensive way to learn it.
- Ignoring data location. It is easier to choose a region at the start than to move later.
- Assuming a free plan covers production. The free tiers are small, and Zapier's free plan is limited to two-step Zaps.
- Building before deciding. Rebuilding hundreds of workflows in another tool is a project in itself.
Next step
Pick your five workflows and run the afternoon test this week. If you would like an outside view of a tooling or automation decision in your business, whether in recruitment, sales or operations, you can book a short call. For more guides, see what n8n is and which recruiting workflows fit it, twelve processes you can automate in a recruitment agency, what to automate first and what to keep human, and the rest of the blog.
Sources
- Zapier, Plans and Pricing
- Zapier Help, How is task usage measured in Zapier
- Zapier Help, What is replay
- Zapier Help, Set up custom error handling
- Zapier, Data Privacy Overview
- Make, Pricing and Subscription Packages
- Make Help Center, Credits
- Make Help Center, Overview of error handling
- Make Help Center, Organizations
- Make, Data Processing Agreement (May 2024)
- Make, Make vs n8n in 2026 (a vendor comparison page, used only for Make's listed entry price and app count)
- n8n, Plans and Pricing
- n8n Docs, Handle errors gracefully
- n8n Docs, Privacy and security
- Regulation (EU) 2016/679, Article 28 (Processor)