Colour theme

Comparison

Which automation platform should you use?

There is no overall winner on this page, deliberately. Which platform fits depends on your budget, your technical comfort, and the shape of the workflow — and those pull in different directions. This page organises evidence we already tested or documented elsewhere on this site by the actual decision you are making, rather than by which tool scores highest on a made-up scale.

Last updated Prices and terms last checked
How this page is funded

No link on this page is an affiliate link. We are in the programme below, so we could earn a commission elsewhere on this site — but not from anything on this page. We recommend free tools where they are the better answer, including ones that earn us nothing.

  • Make.com: 35% of subscription payments for 12 months; extra-operation purchases do not earn commission (terms verified 2026-09-24)
What we actually ran

Three synthetic scenarios each on Make, n8n Cloud and Pabbly’s free plans, on 2026-09-22 — the same tests behind the billing-units comparison and the failed-runs comparison. We have never run Zapier. This page cites those same three tests, not new ones.

How to use this page

Skip to the section that matches your actual constraint — budget, technical comfort, or workflow shape — rather than reading top to bottom for a verdict. There isn’t one.

0. Check whether you need a platform at all

Before comparing any of the four: native integrations, Google Forms→Sheets, and Apps Script triggers solve a large share of what people search “automation platform” to solve, for free and with nothing to subscribe to. The full free-first checklist is on the free alternatives page.

1. By budget

PlatformFree tier realityDid we run it?
MakeReal free plan; we hit no paywall across three scenariosYes — tested
n8nCloud is a time-limited trial, not permanent. Self-hosted Community has no cost or execution cap.Cloud trial — tested. Self-hosted — never run.
PabblyFree plan; our account showed 100 tasks allottedYes — tested
ZapierFree tier exists per its own pricing page; we have no figures to addNo

We tested itA free tier is spent by failure, not just success: a deliberately broken run consumed 21 of our 100 free Pabbly tasks in about ten minutes, because automatic re-execution is on by default and retried it for us without asking.

DocumentedSelf-hosting n8n Community removes the licence cost entirely — the running cost becomes your own time (updates, backups, uptime) instead of a bill. We have never run the self-hosted edition, so we cannot tell you how much of your time that actually takes. (n8n documentation, accessed 2026-09-22)

2. By technical comfort

Make, n8n Cloud and Pabbly are all hosted — no server to run. Self-hosted n8n Community is the one genuinely free-forever option among these four, and the one that asks the most of you technically: you own uptime, backups and security yourself, a trade we have not made ourselves and cannot describe from experience.

3. By workflow type — this is where it actually matters

The single clearest, most evidence-backed differentiator we have is how each platform bills a workflow that fans out into many items per run — an order with several line items, a batch of records, anything that multiplies.

We tested itOn the same ten-line-item order: Make cost 13 credits, Pabbly cost 11 paid tasks, and n8n cost exactly 1 execution. n8n billed the same 1-line-order and 10-line-order identically — the only one of the three where line-item count made no difference to the billed unit.

If your workflow’s cost depends on something your customers control — order size, list length — that is worth weighing heavily. The full breakdown, including why the units do not convert between platforms, is on the billing-units comparison.

4. By failure-recovery needs

PlatformWhat we saw on a forced failure
MakeNo automatic re-execution; recovery was manual in our test
n8nNo automatic re-execution in a verified ~11-minute window; per-node retry exists but is off by default and untested by us
PabblyRetried automatically on its own — the only one of the three that did — but that same behaviour silently burned a fifth of our free monthly allowance on one broken run
ZapierNot tested — no claim either way

Pabbly’s automatic retry is genuinely double-edged: a strength if you want unattended recovery, a real risk on a free plan if a failure repeats. The full run log is on the failed-runs comparison.

5. By debugging depth

We tested itIn our tests, n8n surfaced the most detail: the failing node highlighted, downstream nodes visibly skipped, and the exact failing item shown beside the output. Make’s error panel named the origin and offered to build an error-handler route. Pabbly’s per-iteration selector was the most useful of the three for finding which item of a fan-out failed. We cannot compare any of this to Zapier — we have not run it.

6. What we cannot tell you about Zapier

Zapier is the best-known name in this category, and the one we cannot monetise — its affiliate programme is gated and its rate is undisclosed publicly.

DocumentedZapier’s own legal terms state its commission rate only as “set forth in the Partner Portal” — not published, and not something we can quote a figure for. (Zapier legal terms, accessed 2026-09-22)

⚠ We could not verify thisWe could not verify anything about how Zapier actually bills a fan-out workflow, how it handles a forced failure, or how much debugging detail it surfaces, because we have not run it. Its own pricing and documentation are the authority until we have tested evidence to add — see the Zapier tool page for exactly what that page does and does not claim.

There is still no winner — here is where to start instead

Limitations and failure cases

  • We tested Make, n8n Cloud and Pabbly on three synthetic scenarios each, on one day (2026-09-22). That is not a survey of the market, and behaviour can change with vendor updates.
  • We have never run Zapier or self-hosted n8n Community. Every claim about either is explicitly marked as not tested.
  • We did not test any of the four at real production volume, under concurrency, or with alerting configured — see each tool's own page for its full limitations list.
  • This page repackages evidence already published elsewhere on this site; it does not introduce new testing.

Still deciding whether the task is worth automating at all, on any platform? The Automation Decision Sheet works through that question first, free and ungated.