Zapier bills the action. Make bills the check. That is the whole comparison.
Zapier counts a task each time an action step succeeds; triggers and filters are free. Make counts an operation each time a module runs “to process data or check for new data”, so a scenario that polls and finds nothing still pays. Zapier is in production at Kindra; Make is not, so its column is documentation only.
We run one of these and not the other
Zapier is in production here, alongside n8n. Make is not, and never has been. Everything on this page about Make is its own pricing page and help centre, fetched on 3 September 2026 and linked. We have said this first because a comparison written by someone who uses one side is biased in what it can verify, and you should read the Make column knowing that.
What we can do is line the two billing models up against each other from the vendors’ own definitions, because that is where the decision actually lives. Feature lists converge. Both connect to thousands of apps, both do branching, both run on somebody else’s servers in another country. The bill does not converge.
If you have already decided the answer is neither — because volume is high, or the logic is complex, or the data has to stay on infrastructure you control — Zapier versus n8n is the page you want, and this one will only confirm what you know.
The unit of billing, in each vendor’s own words
Zapier: “A task is counted each time an action step in your Zap runs successfully. Trigger, filter, and certain built-in tool steps do not count as tasks” (Zapier help centre, fetched 3 September 2026). So a Zap that checks for new invoices every two minutes and finds none costs nothing. A Zap that finds one and writes it somewhere costs one task.
Make: “An operation is a single module run to process data or check for new data.” Every module after the trigger runs once per bundle, so a module that sends five emails is five operations; the trigger itself runs “only once to check for or retrieve data, regardless of the number of bundles returned” (Make help centre, fetched 3 September 2026). So a scenario that checks every fifteen minutes and finds nothing costs one operation per check.
Read those two definitions again, because they produce opposite cost curves. Zapier is expensive when things happen and free when they do not. Make is cheap per unit and never free, because looking is charged.
The arithmetic on one real workflow
Take a workflow that watches a job system for completed jobs and creates a draft invoice in Xero — three steps: a trigger, a lookup, an action. Volume: forty completed jobs a month.
On Zapier that is forty successful action steps, so roughly 40 tasks a month, or 80 if the lookup is a searchable action rather than a filter. The Professional plan starts at US$19.99 a month with 750 tasks and a two-minute polling interval (Zapier pricing, fetched 3 September 2026). You are nowhere near the ceiling.
On Make the same scenario, scheduled every fifteen minutes, runs 96 checks a day, or about 2,900 trigger operations a month before it has processed a single job. Add the forty jobs times two further modules and the month is about 2,980 operations. Make’s Core plan is US$12 a month with 10,000 operations and a one-minute minimum interval; the Free plan is 1,000 operations, two active scenarios and a fifteen-minute minimum (Make pricing, fetched 3 September 2026). Core is cheaper than Zapier Professional and the arithmetic is not close. Schedule the same scenario every minute, as Core allows, and polling alone is about 43,000 operations a month — four Core plans’ worth. The interval is the cost dial.
Published plans, side by side
Both vendors publish in US dollars by default. Zapier’s page offers a currency selector including AUD; Make’s does not. We quote what each publishes rather than converting, because a converted figure on the day we fetched it is wrong by the time you read it.
| Zapier | Make | |
|---|---|---|
| Billing unit | Yes A task: one successful action step. Triggers, filters and some built-in tools are free. | Partly An operation: any module run, including the trigger’s check for new data. Per bundle after the trigger. |
| A poll that finds nothing | Yes Costs nothing. | No Costs one operation, every time, on the schedule you set. |
| Free plan | Partly 100 tasks a month, two-step Zaps only, 15-minute polling. | Yes 1,000 operations a month, two active scenarios, 15-minute minimum interval, 5-minute execution limit. |
| Entry paid plan | Partly Professional from US$19.99 a month, 750 tasks, multi-step, 2-minute polling. 33% off annually. | Yes Core US$12 a month, 10,000 operations, 1-minute interval. Pro US$21, Teams US$38. |
| Overage | Partly Switches to pay-per-task at “a higher per-task rate” Zapier does not publish. Stops at three times the plan limit. | Not assessed Not assessed. Make’s help page on operations does not describe overage handling. |
| Where your data is processed | No “AWS servers located in the United States.” No other region offered. | No “AWS (EU/North America)” on every plan. No Australian region. |
| Assessed by us | Yes In production. We know its failure modes, including the replay queue and the private-app throttle. | No Never used in production. Documentation only. |
Where Make probably wins, and we would say so
Price per unit, decisively, for a workflow that fires often. Ten thousand operations for US$12 against 750 tasks for US$19.99 is not a comparison Zapier wins on any workflow where things actually happen. And Make’s one-minute interval on its cheapest paid plan beats Zapier’s two minutes on Professional and matches Zapier’s one minute on Team, which starts at US$69 (Zapier, fetched 3 September 2026).
The free plan is more useful. A thousand operations and two live scenarios is enough to run something real for a sole trader; Zapier’s hundred tasks and two-step limit is a demonstration.
And Make publishes an EU processing option — “AWS (EU/North America)” — where Zapier is US only. Neither is Australia, and for most trade businesses neither matters. If an EU location matters to you for some reason, only one of these offers it.
Where Zapier probably wins
Idle is free. A Zap that waits for a rare event — a new customer, a failed payment, a form that gets filled in twice a week — costs nothing between events. On Make the same watch costs an operation every interval whether or not anything happened, so a rarely-firing scenario checked every minute burns roughly 43,000 operations a month doing nothing. Webhook triggers avoid this on both platforms, but only where the source system can send one.
Legibility, when somebody else will maintain it. Zapier’s failure modes show up in a UI a business owner can read. That is why it is usually what we hand over when a client will run the automation themselves after we leave — the reasoning is on the Zapier integration page.
And we know it. We know that held runs replay at one a second, that workflows stop above three times the plan limit, and that a private app on Professional is capped at 100 requests a minute. We know none of the equivalent things about Make, and neither does anyone who has only read its pricing page.
What neither of them fixes
Both are plumbing. Three problems survive a move in either direction, and each has a different page.
- KNDR·01A process that changes every monthBoth will run an unstable process on a schedule and charge you for it. What not to automate is the page to read before either pricing page.
- KNDR·02Fifteen chained steps with error handling bolted onThat is the point where a platform stops being cheaper than a small purpose-built service, on either platform. Where that line sits is its own page; most small businesses are nowhere near it.
- KNDR·03Data that has to stay in AustraliaNeither offers it. The options are self-hosted n8n on infrastructure you or we control — where things actually run — or a custom build. Not a different subscription.
What we have not assessed
Make, as a product. We have not built a scenario, hit its rate limits, watched its error handling or restored a failed run. Everything in its column comes from two Make pages fetched on one day. Make’s overage behaviour in particular is not described on the page we worked from, so the table says so rather than guessing.
Per-connector reliability at scale on either platform. The capability register publishes what each vendor documents; observed reliability is the column of it still empty.
Currency. Zapier bills in AUD if you choose it and Make does not appear to; we have not compared the AUD figures because Zapier’s page renders them only after selection and we would rather link the page than quote a number that changes with the exchange rate.
The questions people actually ask.
01So Make is cheaper. Why do you use Zapier?
Because most of what we hand to clients waits for rare events, and idle is free on Zapier. Because the failure modes are visible to a non-developer. And because we know its limits from running it, which is worth more to a client than a lower list price on a platform whose behaviour under load we would be guessing at. If your workflow fires constantly, Make’s arithmetic wins and we have said so.
02Can I avoid Make’s polling cost with webhooks?
Where the source system can send one, yes, on both platforms — a webhook trigger runs only when something happens. Many small-business systems cannot. Tradify has no API at all, MYOB has no webhooks, and ServiceM8’s webhooks carry no data so each one costs a follow-up call. Check the source system before assuming webhooks solve it.
03Prices change. How current is this?
Both pricing pages and both help articles were fetched on 3 September 2026 and are linked under the table. The billing definitions have been stable for years; the dollar figures move. Check them before deciding, and if either vendor changes what counts as a unit, tell us and we will change the page.
Stop bending.
Start shipping.
Bring the one flow that keeps costing you a Tuesday afternoon. We map it live and tell you what we'd build, what we wouldn't, and what it costs.