Zapier vs Make vs n8n vs custom code: which fits an ecommerce brand?

Most comparisons of Zapier, Make and n8n are written for marketing teams. An ecommerce brand has different questions: what happens to an order when a step fails, how the bill grows with order volume, and whether the tool can reach Amazon, a 3PL or the accounting system at all. This guide compares Shopify Flow, Zapier, Make, n8n and custom code on those questions, with a worked example at 120 orders a day. Every vendor fact links to the vendor's own page, checked October 2, 2026. No vendor paid for or reviewed this guide; we build with whichever option fits the workflow, including custom code.

Which tool fits which ecommerce workflow?

Work through the options in order and stop at the first one that does the job reliably. Your platform's own automation comes first because it costs nothing extra. A workflow tool comes next, when the job spans a few apps. Custom code comes last, when volume, exceptions or the number of systems outgrow the tools. The first table covers how each option counts usage, where it runs and how it recovers from errors; the second covers connectors, upkeep and when each one is the better choice. Every row was checked October 2, 2026.

Amazon is the weak spot of every workflow tool. Zapier and Make both offer an Amazon Seller Central app, and n8n reaches Amazon through its generic HTTP Request node. That is enough for order alerts and reports. Anything that writes to Amazon, or pulls a lot of data, meets the Selling Partner API's rate limits, covered below. If Amazon is in the workflow, check that limit before you compare prices.

The n8n license matters too: n8n calls its Sustainable Use License fair-code, not open source, and it allows use only for your own internal business purposes, for non-commercial or personal use, or for free non-commercial distribution. Running your own brand's workflows fits; n8n's license FAQ rules out hosting n8n as a service where clients build their own workflows (both checked October 2, 2026).

How much usage does a 120-order-a-day workflow use on each tool?

Take an illustrative brand shipping 120 orders a day from Shopify through a 3PL, and one common five-step workflow. It is a made-up example to show how each vendor counts, not a client result:

1. Trigger: a paid order arrives in Shopify.
2. Create the order in the 3PL's system through its API.
3. Tag the Shopify order with the 3PL's order number.
4. Add a row to an order log in Google Sheets.
5. Post the order number to the operations channel in Slack.

Assume a 30-day month, every step succeeds, each order starts its own run from an instant trigger, and no step loops over line items. Each vendor's published counting rule then gives these numbers:

Orders a month = 120 orders a day × 30 days = 3,600

Zapier tasks = 3,600 orders × 4 actions = 14,400 (the trigger uses no tasks)

Make credits = 3,600 runs × (1 trigger operation + 4 module actions) = 18,000

n8n executions = 3,600 runs × 1 execution = 3,600, however many steps each run has

The last column shows each vendor's lowest paid tier exactly as its pricing page listed it on October 2, 2026, and how far that tier stretches for this workflow.

The shape of the workflow changes the math, and each tool rewards a different shape. On Zapier, Filter, Paths and Formatter steps cost nothing, so logic is free and actions are what you pay for. On Make, every module action counts, and a polling trigger costs a credit on every check: polling Shopify every 15 minutes instead would be 2,880 trigger credits plus 14,400 action credits, or 17,280. On n8n, a Schedule Trigger counts every time it fires, but one run can process every new order: a run every 30 minutes is 1,440 executions a month, inside the Starter tier.

Failures and retries count differently too. Zapier does not count actions that error or halt, but replaying an entire Zap run counts every successful step again. Its pricing page says that past your tier's limit you switch to pay-per-task billing, at a higher per-task rate, unless you turn it off or move up a tier. n8n does not count manual test runs or error workflow runs. Budget for the second and third workflows every brand adds next, such as tracking back from the 3PL, cancellations and returns, each with its own count.

When do you need no workflow tool at all?

More often than tool comparisons suggest. Before you build a workflow, check whether the systems already talk to each other, starting with whether your 3PL or shipping platform imports Shopify orders on its own. ShipStation, for example, says its Shopify integration imports orders with physical products and that tracking can flow back to Shopify when you create a label (checked October 2, 2026). If your 3PL works that way, the order-to-3PL workflow above already exists, and a workflow tool would only add a second path for the same order.

Shopify Flow covers much of the rest for Shopify-only work: tagging risky orders, holding fulfillment, alerting a channel, adding a row to a sheet. On Amazon, the platform's own tools come first as well; our guide to automating Amazon operations goes task by task. A workflow tool earns its place when data must move between systems that have no native link, and our ShipStation integration page shows where that native link stops for a 3PL.

When should you move from a workflow tool to custom code?

Move when a workflow costs more in fixes, fees or risk than it saves in hours. These are the signals we look for when a brand shows us its existing Zaps, scenarios and workflows:

The last signal matters most for Amazon sellers. Amazon's Selling Partner API rate-limits each operation with a token bucket: tokens refill at a set rate up to a burst limit, a request with no token left is throttled, and Amazon asks for a back-off strategy when throttling repeats (checked October 2, 2026). In our SP-API development work, code paces each operation against its own bucket and uses reports instead of polling for bulk data. Our post on building seller tools with AI works through what those limits mean for an order backfill.

EzBookPrep made a related move. The client, an experienced book seller, ran prep services for other book sellers through Quickbase and needed the same functions with a better user experience, without the manual follow-ups that had made him the bottleneck. We built custom software with the Quickbase functions he relied on, customizable dashboards and automated alerts for the next action on each book. Cheddy shows the exceptions signal: kits, multipacks and single-pack ASINs had to be mapped to warehouse components, and ready-made software was missing many of the features that mapping needed.

How do you migrate from Zapier, Make or n8n without breaking orders?

Move one workflow at a time, and never let two systems create the same order. Order workflows fail quietly: a duplicate shipment or a missing one shows up days later as a complaint or a reconciliation gap. These are the steps we follow:

1. List what runs today. Every Zap, scenario and workflow that touches orders, with its trigger, steps, owner and the apps it writes to. Switch off the ones nobody can explain only after you know what they do.
2. Make every write safe to repeat. Use the Shopify order ID as the key for every write, and check whether the 3PL already has that order before creating it. Zapier's full replay reruns every step, including the trigger, so a replayed run can send the same order twice unless the receiving side refuses duplicates.
3. Run old and new in parallel. The new code reads the same orders and logs what it would send, without sending. Compare its log with what the old workflow did for at least a full week, including a weekend and, if you can, a promotion.
4. Log every step. One line per order per step: what was sent, what came back and when. Keep it as long as your finance team needs the history, not as long as a plan's log retention.
5. Cut over at a set time. Switch the old trigger off and the new one on for orders after that time, then check the orders that arrived around the switch by hand.
6. Keep a way back. Leave the old workflow switched off but intact for a few weeks, with a written note of how to turn it back on.

At higher volume, put a queue between the trigger and the work, so a slow or throttled API delays orders instead of dropping them. For Beezboard, a profit dashboard for Amazon sellers, we used Amazon SQS, a message queue, for real-time updates on order activity such as fee changes and sales.

Who can help you choose and build it?

Ecomsellertool is a tech agency that grows ecommerce brands by closing the operations gaps that leak revenue, with custom tech. Because we sell none of these tools, the plan names whatever fits each step, including Shopify Flow or no new tool. Send us one workflow and a person replies within 2 business days with a written plan, timeline and fixed price.

When a workflow tool is enough, we build it in your own Zapier, Make, n8n or Shopify account. When code is the answer, it can run on Ecomsellertool Growth OS, our base system for listings, stock, orders, warehousing, advertising and reporting, deployed on your own accounts. You keep your accounts, your data and the custom code we build; the Growth OS base is licensed to you. Shopify brands adding Amazon or Walmart can start with our page for Shopify brands, and the workflow automation page lists the workflows we are asked about most.