Shopify Middleware and iPaaS: When You Need It and When You Do Not

Middleware gets sold to Shopify brands as inevitable. Grow past a certain size, the pitch goes, and you will need an integration platform sitting between your store, your ERP, your warehouse and your finance system. Sometimes that is true. Often it is a monthly platform fee for a job two well-built connectors were already doing.

The opposite failure is just as common. A brand hand-rolls a set of scripts between Shopify and its ERP, they work for eighteen months, and then the developer leaves and nobody knows why orders stopped syncing on a Tuesday. This article is a decision framework for Shopify middleware and iPaaS, not a recommendation. Some of you should buy a platform. Some of you should not, and we would rather say so now than after you have signed a three-year contract.

What Shopify middleware actually does

An integration platform as a service (iPaaS) such as Celigo or Patchworks is a hosted layer that moves data between systems. It listens for events (an order is created in Shopify), transforms the data (map Shopify line items to ERP SKUs, apply tax codes, split bundles), pushes it to the destination, and handles what happens when the destination is down. That last part is the real product. Anyone can call two APIs. Retries, queuing, alerting and an audit log of every record that moved are what you are paying for.

A point-to-point connector does a narrower version of the same thing for one pair of systems. Your 3PL's Shopify app, a NetSuite connector from the ERP vendor, a Xero sync app: each one is built for exactly that route and nothing else. We covered the ERP side of this in our guide to Shopify ERP integration with NetSuite, Business Central and Brightpearl, and the warehouse side in connecting Shopify to a 3PL or WMS.

The honest test: five questions before you buy a Shopify integration platform

Run your own operation through these. Be strict about the answers.

How many systems are you joining? Two systems is a connector. Three with simple flows is still usually a connector or two. Four or more systems that all need to know about the same order is where a platform starts to earn its keep, because you are no longer managing integrations, you are managing a web of them.How much does the data change in transit? If a Shopify order can be posted to the ERP more or less as it arrives, a connector will do. If you need to split kits into components, route orders by market or channel, apply different tax logic per country, or enrich records from a third system before they land, that transformation logic has to live somewhere. A platform gives it a home. A connector usually does not.What is your order volume and throughput? Shopify's GraphQL Admin API gives a standard store 50 cost points per second, 100 on Advanced and 500 on Plus, according to Shopify's rate limit documentation. At a few hundred orders a day nothing touches that. At a few thousand a day during peak, with inventory syncing every few minutes across two warehouses, a badly written integration will start hitting 429 errors and silently dropping records. Platforms are built to queue and back off. Most cheap connectors are not.How much error handling do you actually need? If an order fails to reach your warehouse and someone notices in an hour, what does it cost you? For a brand shipping same-day, a lot. For a made-to-order furniture brand, almost nothing. Retry logic, alerting and a searchable log are worth paying for in the first case and overkill in the second.Who owns it in a year? This is the one people skip. A platform can be run by an operations person after training. A custom build needs a developer who understands it, and if that developer is a freelancer who has moved on, you have a liability, not an asset.

If you answered 'two systems, light transformation, hundreds of orders a day, and we would notice a failure quickly', you do not need middleware. Fix the process, pick good connectors, and put the platform fee into conversion work instead. Most Shopify brands at that stage do not have an integration problem or a traffic problem. They have a conversion problem, and that is where the money is.

Signs you have outgrown a simple connector

The tell is rarely a dramatic outage. It is drift. Watch for these:

Silent failures. Orders that never reach the ERP, with no error anywhere, discovered when a customer chases. Connectors often fail quietly because they have nowhere to report to.Manual reconciliation. Someone exports Shopify orders and the ERP order list every Friday and compares them. If that is a recurring task on anyone's calendar, your integration is not doing its job.The spreadsheet person. One person in the business holds the whole thing together with a workbook of SKU mappings, exceptions and notes. When they are on holiday, things break.Connector sprawl. You have five single-purpose apps installed, each with its own schedule, its own field mappings and its own definition of 'in stock'. Nobody can say which one is right when they disagree.Peak-season firefighting. Research commissioned by Patchworks and run by OnePoll in November 2025 surveyed 200 senior UK retail decision makers. It found that 39% of retail teams spend more time firefighting integration issues than optimising sales, and 48% said they lose more than £50,000 a year to integration failures. It is vendor research, so read it with that in mind, but the pattern matches what we see with brands in Glasgow and across the UK.

If a person and a spreadsheet are the integration, you do not have an integration. You have a single point of failure with a salary.

Connector vs iPaaS vs custom build: a realistic three-year cost

First-year sticker prices are misleading because they hide the implementation and the ongoing ownership. Here is a fairer comparison for a mid-sized UK brand, based on what we see in the market rather than any vendor's price list. Treat the ranges as indicative. Both Celigo and Patchworks price on quote.

Off-the-shelf connectors

Typically £0 to £300 a month per connector. Setup is days, not weeks. Over three years, two connectors might cost £5,000 to £20,000 including the occasional developer fix. The hidden cost is the reconciliation time and the failures you did not catch. If those are small, this is the right answer and there is no shame in it.

iPaaS platforms such as Celigo and Patchworks

Neither publishes list prices. Celigo tiers by endpoints and flows, with the entry tier covering three endpoints and 20 flows. Third-party estimates put entry contracts at roughly $1,000 to $1,500 a month and larger ones at $2,500 to $5,000 a month or more, before implementation. Patchworks tiers by connectors, flows and monthly operations, with its entry tier around two connectors, ten flows and 250,000 operations a month. Add an implementation project (commonly £10,000 to £40,000 depending on how much transformation logic you need) and internal time to own it. A realistic three-year figure is £50,000 to £150,000. What you get for that is monitoring, retries, an audit trail, and the ability for a non-developer to change a mapping without raising a ticket.

Custom build

A senior team can build a well-engineered integration on Shopify's APIs, webhooks and a queue for roughly £15,000 to £50,000, depending on scope. The three-year cost is dominated by maintenance: Admin API versions that Shopify retires on a fixed schedule, ERP upgrades, and the developer time to keep it healthy. Budget a further 15-25% of the build cost per year. The advantage is that it does exactly what you need and nothing else, with no per-operation fees. The risk is question five above: who owns it. If the answer is 'us, with a partner on retainer', it can be the cheapest option over three years. If the answer is a single freelancer, it is the most expensive, you just have not paid for it yet. We have written more on that trade-off in custom Shopify app development: when to build and when to buy.

The option most brands overlook: Shopify Flow and a native app

Before any of the above, check whether Shopify can do it natively. Shopify Flow is free on every plan and handles a surprising amount of light integration: tagging, routing, notifications and, on Grow, Advanced and Plus plans, a Send HTTP request action that can post order data straight to another system's endpoint. It has real limits (it waits a maximum of 30 seconds for a response, loops cap at 100 items, and there is no order-updated trigger), so it is not middleware. But for a brand with one downstream system and simple rules, Flow plus a good connector plus a tidy process beats a platform fee every time.

Where Flow runs out and a full platform is still too much, a small custom Shopify app that owns one specific job (say, pushing fulfilled orders to a finance system with retries and a log) is often the right middle ground. That is the kind of work we do under Shopify app development: narrow, well-tested, and built by a senior team with no juniors and no outsourcing, so the person who built it is the person who maintains it.

A decision rule you can actually use

Two systems, simple data, you would notice failures: connectors and process. No middleware.Two or three systems, but transformation is complex or volume is spiky: a custom app or a single well-scoped build, with a partner who will still be around.Four or more systems, high volume, real cost to failures, and an operations team to own it: iPaaS. Get quotes from both Celigo and Patchworks and price three years, not one.

Whichever you choose, do not let it become the project that eats the year. Integration is plumbing. It should be reliable, boring and invisible, so that the effort in the business can go back to the thing that actually grows revenue: getting more of your existing visitors to buy. That is what our Conversion Growth Retainer is for, and it is usually a better use of the same budget than a platform nobody asked for.

If you are not sure which bucket you are in, talk to us. We have spent 12 years and 100+ projects looking at Shopify stacks from Glasgow, and we are happy to tell you that you do not need middleware if that is the honest answer.