Shopify Flow vs Zapier vs n8n: Pick Your Automation Layer

The Quick Verdict

Here is the answer by workflow shape; everything after it is the arithmetic underneath.

If one row describes your situation, that is usually the whole answer — the rest of this article shows the arithmetic, names the boundaries that decide it, and handles the case where two rows both apply. Two scoping notes before anything else. Every price here is a United States list price in US dollars, with Zapier and n8n figures verified on 28 July 2026 and Make on 1 August 2026. And the volumes priced here run to 10,000 billable units a month, because that is the band in which all four vendors actually publish numbers.

Four Layers, Not Four Competitors

Most comparisons of these tools line them up as rivals and score them on features. That framing is what produces the two most expensive mistakes merchants make here: paying a monthly subscription for work that never leaves Shopify, and assuming the free layer will eventually stretch to cover the work that does. They are stacked, not opposed. The useful question is not which one wins — it is where the boundary of your workflow sits.

The last line of each card is the one that decides your bill, and it is the least-read line in this whole subject. A task, an operation and an execution are three different sizes of thing. Comparing the prices attached to them without comparing the units is like comparing a price per litre with a price per gallon and declaring one vendor cheaper. If you would rather have the arithmetic before the argument, the worked example in the same workflow, three different bills runs one ordinary workflow through every count.

Why Most Stores End Up Running Two

A store that has grown past a couple of dozen orders a day usually has both kinds of work at once. Tagging a repeat buyer, holding a suspicious order, unpublishing a sold-out product, alerting the warehouse — all of that happens entirely inside Shopify. Copying an order into a CRM, pushing a file to a 3PL portal, reconciling with accounting — none of that does.

The mistake is treating this as one decision. It is two, and they are decided on different grounds: the inside work is decided by what Flow can do, and the outside work is decided by what the crossing costs per unit. Flow itself is one capability among many on the platform — our guide to what Shopify includes out of the box puts it in the context of everything else your plan already carries.

Where Shopify Flow Stops

Before you price anything external, it is worth knowing exactly how far the free layer reaches — because on most stores it reaches further than people expect, and stops in places people do not expect. Two questions have to be answered in order: does your plan carry Flow, and does Flow carry the action you need. Clearing the first tells you nothing about the second.

What Flow Does Inside Your Store

Shopify's own description is precise about scope: Flow is an ecommerce automation platform for automating tasks and processes within your store and across your apps. In practice that means event-driven building blocks — an order is created, a customer is tagged, inventory crosses a threshold — wired to actions that Shopify already ships.

The catalogue is larger than most merchants realise. Shopify's trigger reference listed 83 triggers when we counted it on 1 August 2026, matching an independent count four days earlier, and the actions reference carried 109. One of those triggers is worth calling out on its own, because a great deal of published advice about Flow is quietly out of date on it.

Flow is not purely event-driven — it has a clock. The Scheduled time trigger starts a workflow at a date and time you set, with an optional repeat rule and an optional stop rule. Shopify describes the repeat window as running from a longest interval of one year down to a smallest possible interval of 10 minutes, while the same page's setup steps offer Days, Weeks and Months as the period units. Both statements are Shopify's own; we print them as they stand rather than resolving them for you.

The catch is what a scheduled run arrives with, which is nothing. Shopify states that because no store or app event started the workflow, there is no data automatically included — so a scheduled workflow pulls its own data in through one of the 18 Get actions (Get order data, Get product data, Get customer data and so on). That is an extra step, not a barrier, and it puts a whole class of recurring housekeeping inside the free layer.

Shopify's own AI assistant can now generate Flow workflows from a plain-language description, which changes how fast you build one but not what one can do — our guide to Shopify's built-in AI tools covers where that helps. If you want to see the editor itself rather than read about it, the walkthrough below builds a workflow end to end.

The Two Gates Inside Flow

Two gates sit inside Flow: Send HTTP Request, and tasks built by custom partner apps. Before either of them comes a plan check. Shopify publishes the list twice — on its Help Center and on its Flow marketing page — and both name the same four: Basic, Grow, Advanced and Plus. Starter appears in neither. You check this one by reading the plan name on your Shopify bill; if it says Starter, nothing in this section applies to you and the section on external layers does.

Gate one is the one that surprises people. The action that lets Flow talk to anything outside Shopify — Send HTTP Request — is available only on the Shopify Plus, Advanced, or Grow plans. You check this one by opening the Flow editor and looking at whether the action is offered when you add a step. If it is not, your workflow stops at the edge of Shopify, and the practical question becomes which is cheaper: the plan, or a layer that does the crossing from outside.

One boundary worth stating plainly rather than discovering in production: Shopify publishes no payload or body-size limit for Send HTTP Request. We fetched the full action page and found none, which means the limit is untested rather than generous — see the search boundary below for what was checked and when.

What Flow Will Not Do at Any Plan Price

Plan gates are negotiable — you pay Shopify more and they open. The four below are not: no tier above yours removes any of them. One can still be stretched by a different lever — installing an app that registers its own actions with Flow — but not by the plan, and knowing all four before you build is the difference between an automation plan and a rebuild. They are also, collectively, the honest case for the external layer.

Shopify documents that third one itself. Its developer docs define a Flow action as "a task that's executed in a store or in an app when certain conditions are met", and tell partners they can build tasks related to their app for merchants who have both that app and Flow installed. Which extra actions you have therefore depends on what you have installed — and Shopify publishes no readable list of which apps register them, so it is a per-app question rather than a count.

The email limit is the one that catches most merchants, because it looks like a feature that must exist. It does not, and Shopify says so directly on the action's own page — the guidance is to route customer email somewhere built for it.

This is also where a naming collision does real damage. Marketing platforms call their automated email and SMS sequences flows, and a Klaviyo flow has nothing to do with Shopify Flow — different product, different vendor, different job. If the automation you are picturing sends a series of emails to a customer, you are looking for the marketing layer, and our guide to running Klaviyo on a Shopify store is the right page for it.

How Many Workflows Can Flow Run a Month?

This is the most-asked question about Flow that has no published answer. We checked the Flow hub, its getting-started, create, manage and reference pages, the Send HTTP Request page and Shopify's Flow marketing page, and none states a runs-per-month, per-day or per-hour cap. What the documentation does give you is a pointer to a different number entirely.

Source: Shopify — API rate limits. Shopify's table names the buckets Standard, Advanced, Plus and Shopify for enterprise, and does not say which plans sit in each; read against the plan names used in this article, a Basic or a Grow store is in the Standard bucket — our mapping, not Shopify's wording. These are Admin API limits, not a Flow allowance; Flow's documentation defers to them rather than publishing its own figure.

Read that as a ceiling on how hard your store can be worked, not on how many workflows you may own. A Basic store and a Plus store both run Flow; the Plus store's bucket is ten times deeper, which matters when several tools — Flow, an orchestrator, an app — are all working the same store at once. Shopify's own wording ties the two together explicitly.

The Unit Each One Bills You For

Everything downstream — which vendor is cheap for you, why a competitor's price comparison misled you, why your bill jumped after a change that felt trivial — falls out of these four rules.

Zapier and n8n rules verified 28 July 2026 on each vendor's own pricing and documentation pages; Make verified 1 August 2026. Shopify Flow: no billable unit is published on any Flow page.

Zapier's rule is the most generous of the three and the easiest to reason about. Nothing is charged for watching, filtering or failing — you pay when something actually got done, and Zapier notes that some apps or actions may use more than one task for a single step. The consequence is that step count, not run count, is what moves a Zapier bill: doubling the actions in a Zap at least doubles the invoice at identical volume.

Make's rule diverges at exactly one point, and it is the expensive one. A trigger module costs one operation per check, regardless of the number of bundles returned — which includes checks that return nothing at all. Make charges you for looking. Zapier and n8n do not, and on a frequently polled workflow that single difference can outweigh everything else on the invoice.

The Same Workflow, Three Different Bills

Take a workflow most stores would recognise: something happens 1,000 times a month, a filter throws away 40% of those as irrelevant, and the 600 that survive each run three action steps. The tool checks Shopify every 15 minutes, which over a 30-day month is 2,880 checks. That single description produces three completely different unit counts.

Counts derived from each vendor's own published billing rules, verified July–August 2026.

Notice what did the damage. The filter is free everywhere, so it is not the culprit. The three action steps are free on n8n and charged three times over on the other two. And the 2,880 checks — the same checks, on the same schedule — are worth nothing on Zapier and n8n and 2,880 operations on Make. n8n's rule is the one that reads most strangely at first, and it is written down plainly.

Now put prices on those counts. 1,800 tasks lands on Zapier's 2,000-task Professional step — $49.00 a month billed annually, $73.50 billed monthly. 4,680 operations sits under Make's entry price of $12 a month for Core, the price at the lowest position of its usage slider. 1,000 executions sits inside n8n Cloud Starter at $20 a month billed annually or $24 billed monthly. And on Shopify Flow it is nothing — provided the whole workflow stays inside Shopify, which by construction this one does not.

Do not read those three prices as a ranking, because only two of them are one. The Zapier step and the n8n tier are rungs on published ladders you can hold either vendor to. Make's is a floor: it buys the lowest position of a slider with nothing published above it. Make's help centre states that credits replaced operations as its billing unit and that for non-AI apps one operation equals one credit, so 4,680 operations sits inside the 10,000 credits/mo that entry price covers — but "4,680 operations for $12" is an entry price sitting next to a count, not a quote.

The counts are not a ranking either, and this is the part worth carrying out of the section. n8n produces the smallest number and is not the cheapest line. Make produces the largest and has the lowest published entry. The unit does not tell you who is cheaper today — it tells you which way your bill moves when your workflow changes shape. That is why the next paragraph changes exactly one thing.

Then change one thing and watch what moves. Poll every 5 minutes instead of every 15, and the checks go from 2,880 to 8,640 a month. Zapier still bills 1,800 tasks. n8n still bills 1,000 executions. Make goes to 10,440 operations — past the only allowance its published price covers. Same work, same result for the store, and one of the three vendors has just charged you for being impatient.

What Your Own Workflow Shape Costs

The example above is one shape. Yours will differ in the four things that matter — how often it fires, how much a filter removes, how many action steps follow, and how the tool finds out something happened. Put your own numbers in below and read the unit counts first; the prices underneath them are a consequence, not the point.

Three behaviours are worth watching as you move the inputs. The action-step slider moves Zapier and Make together and leaves n8n flat, because n8n does not count steps. The polling choice moves Make alone — switching to a webhook collapses its check count to the number of real events. And the self-hosted line stays empty until you supply all three of its inputs — a server cost, the upkeep hours and your hourly rate — because those three numbers are yours: no page we read prices the server, and a self-hosting line that renders as zero would be the exact myth this article exists to correct.

One workflow most people forget to model at all: the nightly sweep that re-reads yesterday's orders to catch whatever a webhook missed. Shopify's own advice makes that sweep mandatory rather than optional for anything moving money or stock (the reasoning is in webhook or poll), and it is a second workflow with its own unit count on whichever layer runs it. Model it as its own shape rather than assuming it rounds to nothing.

What We Looked For and Did Not Find

Every gap below is a verified absence rather than an assumption — we went looking, and this is what was and was not there. Where a later section hedges a claim about something not being published, this block is the record of which pages were read and when.

The Three External Layers, Priced

All three sit outside Shopify and reach into it, and all three charge monthly. That is where the similarity ends. What each one publishes differs so much that presenting them in one grid would force us to invent the missing cells — so each gets the shape its own disclosure allows.

Zapier: the Ladder You Actually Climb

Zapier is the most transparent of the three on price: an 18-step ladder from 750 tasks up to two million, with both billing cycles published at every step. The steps below are the band this article prices — the calculator above carries all eighteen, so it can price a volume this table does not show.

Source: Zapier's own pricing page — Free covers 100 tasks a month, verified 28 July 2026. The highlighted row is the step the worked example above lands on.

Two things to take from that table. First, the billing-cycle gap is 50%, not the 15–20% most SaaS charges — any Zapier figure quoted without naming the cycle is potentially half the real number. Second, the free tier is smaller than it sounds: 100 tasks is about 33 runs of the three-action workflow above, so it is a proving ground rather than a plan.

Above Professional sits Team, which starts at $69 a month billed annually for 2,000 tasks and adds shared seats — it is a seats plan rather than a different volume ladder, so it does not change the arithmetic here. Enterprise carries no self-serve price at all.

Make: the Cheapest Entry and the Slider Behind It

Make has the lowest entry price of the three: $12 a month for Core, $21 for Pro and $38 for Teams, verified on Make's pricing page on 1 August 2026.

The structural detail matters more than the numbers. Those three prices all sit at the same place: the lowest position of a single usage slider that moves quota and price together, identically across all three paid tiers. The page labels that position 10,000 credits/mo. The one-credit-per-operation rate quoted earlier is Make's rate for non-AI apps; some AI modules consume more than one credit per operation.

Make does have a Shopify App Store listing, under the handle make. It carried a 1.0 rating from three reviews when checked in July 2026 — worth knowing before you install from there, and worth noting that the listing's own displayed prices do not reconcile with Make's pricing page, which is the source used throughout this article.

n8n Cloud: Executions, Billed Annually by Default

n8n's pricing page defaults to annual billing with a "Save 17%" banner, which makes the month-to-month price easy to miss. Both are published, and the gap is much smaller than Zapier's.

Source: n8n's pricing page — Starter covers 2,500 executions at $20/mo billed annually, verified 28 July 2026.

A third Cloud tier exists above Pro, at 40,000 executions a month. We are not printing its price: the only capture that carried it rendered in euros while Starter and Pro rendered in dollars, and a mixed-currency figure inside a dollar comparison would be worse than no figure. It also sits well outside the volume band this article prices.

Is Self-Hosted n8n Really Free?

This is the claim that sends most people down this road, and it is half true in a way that costs money. What is free is the licence. What is not free is everything that turns a licence into a service your business depends on.

• The software: nothing. A standard self-hosted version of n8n is available on GitHub.
• The server: your number, not ours. n8n's docs name an instance class rather than a price — the Hetzner CPX11 (2 shared vCPU, 2 GB RAM, 40 GB SSD) is described as enough for most usage levels. Its monthly price is not published anywhere a tool can read it, so this figure has to come from your own provider.
• Upkeep: hours × your rate. Updates, backups you have actually restored from once, and monitoring that tells you before your merchants do.
• The 2 a.m. question. Who is on the hook when a workflow stops silently during a sale. On a hosted plan that is the vendor; here it is a person you employ or are.

One question belongs in a vendor's inbox before you connect any hosted layer, and it applies equally to all three: where are the Shopify credentials it stores held, and what happens to them when you cancel? We are not answering that here — we have no sourced answer for any of the three — but it is a reasonable thing to ask before handing over access to your store.

Fair-Code Is Not Open Source

Plenty of write-ups call n8n open source. n8n does not, and the distinction has practical consequences if you are an agency or planning to resell what you build. The licence permits use or modification for your own internal business purposes, or for non-commercial and personal use, and permits distribution only free of charge for non-commercial purposes.

For a merchant running n8n to automate their own store, none of this bites — that is squarely internal business use. For anyone thinking of hosting it for clients, it does: n8n names white-labelling and selling it on as an example of what the licence does not allow. The word n8n uses for this model is fair-code, a term it coined for source-available software with restricted commercial use.

How Each Layer Reaches Your Store

A tool that connects to hundreds of apps is not automatically the one that connects best to Shopify. And a tool with a rich Shopify catalogue still has to hear about your orders somehow, which is a separate mechanism with separate costs. Both halves are below.

How Much of Shopify Each One Can Touch

Counts taken from each vendor's own reference pages — Shopify and Make in July 2026, Zapier on 4 August 2026: 20 triggers and 39 actions on Zapier, 27 modules on Make, and a Shopify app node with 10 operations on n8n, which n8n's node documentation lists alongside a separate Shopify Trigger node.

The bottom row is the one to read twice. n8n splits Shopify across two nodes — the Shopify app node for the ten order and product operations, and a separate Shopify Trigger node, listed among n8n's built-in trigger nodes. n8n's docs page for the trigger node links usage templates rather than enumerating the events it supports, so confirm yours is covered before you plan around it.

Webhook or Poll: What Your Bill and Your Latency Hang On

There are exactly two ways for anything outside Shopify to learn that an order arrived. Shopify pushes it — a webhook — or the tool asks repeatedly. Both have a cost, and they are different kinds of cost: one is reliability, the other is money and delay.

Webhooks are instant, and Shopify is unusually direct about their limits. Its developer documentation tells app builders not to rely on receiving webhook data, and to run reconciliation jobs that periodically fetch from Shopify so the two sides stay consistent. That advice applies just as squarely to a Zap or a scenario as it does to an app.

In practice that means any workflow that moves money, stock or fulfilment needs a second, slower workflow that checks the first one's work — a nightly sweep that re-reads yesterday's orders and fills in whatever the webhook missed. That sweep has a unit cost too, on whichever layer you run it.

Polling is the alternative, and here the three vendors diverge in a way that ties this section straight back to the arithmetic above. A poll always lags by its own interval: at a 15-minute schedule a customer waits up to 15 minutes for whatever the workflow does. The only lever that shortens the lag is polling more often — and that lever has a different price on each layer. On Make it costs operations, because every check bills: going from 15 minutes to 5 took the worked example from 4,680 to 10,440 operations. On Zapier and n8n the same change costs nothing at all. Speed is a paid upgrade on exactly one of the three.

There is a floor to it, too. Make's own comparison table puts its minimum scenario interval at 15 minutes on the free plan and 1 minute on Core, Pro and Teams — so on Make, latency is partly a tier question. Shopify Flow's own clock, by comparison, is documented down to a smallest repeat interval of 10 minutes. None of these is a guarantee of speed; they are the boundaries below which the option does not exist.

Which Job Belongs on Which Layer

Run your own backlog of "we should automate that" through this before you price anything — on most stores it moves several items off the paid layer entirely.

Layer assignments follow each vendor's documented capabilities as cited throughout this article. The highlighted row is the one merchants most often pay for unnecessarily.

The last row deserves a note, because it is where this article's scope genuinely ends. Once an ERP and Shopify both own the same orders and inventory, you are no longer choosing an automation layer — you are choosing between a purpose-built connector, an integration platform and a custom build, with reconciliation and cutover to plan. Our guide to ERP integration picks the story up at exactly that point.

Which Layer Is Yours?

The table above answers by job. This answers by situation — your plan, your volume, your team, and the thing that made you start looking. It is a different question from the calculator, which prices the units your workflow produces; this one routes the person reading it. Five questions, about a minute.

When None of the Four Fits

There is a whole category of request that arrives disguised as an automation problem. "Can we automatically apply a different price for these customers?" "Can we show a custom form before checkout?" "Can we stop this product being bought with that one?" None of those is a workflow — each is a change to what the store itself does, and workflows sit on top of that, not inside it.

When a Purpose-Built App Is the Answer

Two examples make the class concrete rather than abstract. Arigato Automation listed plans from $15 to $100 a month when checked in July 2026, metered in actions per month. MESA listed plans from $12 to $299 a month on the same date, metered in automations.

We are not ranking them, and no neutral source exists that does — every comparison we found was published by one of the vendors. What matters structurally is that this class exists: it runs inside the Shopify admin like Flow, but adds code steps and its own meter, which sometimes lands exactly between "Flow can't" and "we need a developer".

When It Has to Be a Custom App

When the answer is a build, the automation comparison stops being relevant and a different set of questions starts: what the build costs, who maintains it, and how it survives a theme change or a platform update. That is a different discipline with a different budget shape, and treating it as an extension of a $12-a-month floor price is how projects go wrong.

There is one bridge between the two worlds worth knowing. Custom partner apps can publish their own tasks into Flow, and Shopify limits that to stores on Plus — meaning bespoke logic can appear as a block your team drags into a Flow workflow rather than a black box nobody understands. If a build is where you are heading, our guide to custom Shopify development covers what that path involves.

How Locked In Are You?

Automation platforms rarely advertise this, and it is the single most underweighted factor in the choice. n8n's documentation describes saving workflows as JSON and importing JSON files into your n8n library; nothing on that page or its siblings mentions another platform's format. The picture is the same in every direction across the three.

The Bottom Line

The decision is smaller than it looks once the two halves are separated. Almost every store ends up with two layers, and the only genuinely difficult part is choosing the second one — which is a question about units, not vendors. A workflow with many action steps and few runs punishes Zapier and Make. A workflow that polls frequently punishes Make specifically. A workflow with enormous volume and someone to run a server argues for n8n on your own hardware, provided you count the hours honestly.

Frequently Asked Questions