Almost nobody outgrows Shopify’s published limits. The pages telling you otherwise are counting products and variants, and those ceilings sit far above where real stores get into trouble. What you outgrow is your back office — and that happens at any catalogue size, on any plan, without a single limit being hit.
We implement Odoo, so we have an obvious interest in you concluding that you need an ERP. Which is why this piece is mostly about how to tell that you don’t.
Key takeaways
- The “limits” framing is the wrong test. Shopify’s real published ceilings are high, and the ones most often quoted online are either stale or were never Shopify’s numbers.
- Above 25,000 items, Shopify stops giving exact counts. That is a documented API behaviour, and it is a better early signal than any product-count ceiling.
- The variant throttle starts at 500,000 variants — and it does not apply on Shopify Plus at all.
- What breaks first is the question you cannot answer, not the record you cannot create.
- Adding apps is not evidence of outgrowing anything. Four apps that agree are fine. Two that disagree are not.
- The one Shopify constraint that bites ordinary stores is configurable products, and it is a genuine architectural fork rather than a ceiling.
What Shopify actually publishes
Search “Shopify limitations” and you get numbers: 100 variants, 100,000 products, 250,000 SKUs, 1,000 uploads a day. They disagree with each other, several trace to forum posts from years ago, and some were never Shopify’s figures.
Here is what Shopify’s own developer documentation states, retrieved August 2026:
| Constraint | What Shopify’s API documentation says |
|---|---|
| Pagination | Arrays of objects are limited to 25,000 objects |
| Counts | Accurate up to 25,000; beyond that the API returns 25,001 |
| Input arrays | Maximum 250 items |
| Variant creation throttle | Begins at 500,000 product variants, then 10,000 new variants per day |
| Plus exemption | The variant throttle does not apply to Shopify Plus |
| Admin API rate | 100 points/second standard, 200 Advanced, 1,000 Plus, 2,000 enterprise |
| Single query cost | May not exceed 1,000 points regardless of plan |
| Bulk operations | Exempt from the max-cost and rate limits that apply to single queries |
Read that variant line again. The throttle starts at half a million variants. If you are anywhere near that number you have a catalogue problem that no ERP will solve either.
We have deliberately not repeated the product-count and variant-per-product figures that circulate widely, because we could not verify them at source. Shopify’s own variants documentation states no numeric cap at all. If a specific ceiling matters to your decision, check it against your own store’s admin rather than against a blog post — including this one.
Why ceilings are the wrong test anyway
A ceiling tells you when a record cannot be created. It tells you nothing about whether your business can answer its own questions.
The count limit is the interesting one. Past 25,000 items, Shopify will tell you there are “more than 25,000” rather than how many. That is a sensible engineering decision and it is also a small preview of the actual problem: at a certain size, your storefront stops being a system of record and starts being a sales channel that happens to hold some data.
Nothing broke. You just stopped being able to ask.
What actually breaks first
The generic warning signs — spreadsheet reconciliation, overselling, accounting that no longer scales — are well covered, including in our own piece on when a store needs an ERP. Assume you know those. These four are the ones specific to running a back office behind Shopify, and each is answerable by looking at how your own week actually goes.
You cannot say what an order cost you. Not the price you paid the supplier — the landed cost, with freight in, duty, and the returns against it. If answering that requires exporting to a spreadsheet, your cost data lives outside the system that holds your sales data. Every margin decision you make is downstream of a join you are performing by hand.
Purchasing happens in email. Supplier lead times, partial deliveries, backorders and what is on the water sit in somebody’s inbox and somebody’s head. This one is invisible until the person holding it is on leave, which is also when it is most expensive.
Two systems disagree and your team knows which one to trust. This is the clearest symptom on the list. Not that stock is wrong — stock is always a bit wrong — but that your staff have developed folklore about which number lies. That knowledge is unwritten, unowned, and walks out of the building at five o’clock.
New staff have to be taught the workarounds. When onboarding includes explaining which report is wrong and why, the workaround has become the process.
What looks like outgrowing Shopify but isn’t
A lot of apps. App count is not a signal. Four apps that agree with each other are a functioning stack. The problem is not quantity, it is when two of them hold the same fact and differ. One app reading a single stock count can carry you a long way.
Hitting a Shopify feature gap. A missing feature that one app fixes is a missing feature, not a structural limit. Buy the app.
Revenue milestones. Revenue is the weakest possible trigger. A high-revenue single-channel store with simple fulfilment may need nothing; a modest multi-channel B2B-and-retail operation may need an ERP badly.
Being told you have. The pages ranking for these searches are largely published by companies selling the alternative. Including, in the interest of consistency, this one.
The one constraint that does bite
Configurable products. This is where Shopify’s model gets genuinely rigid for ordinary-sized stores, and unlike the ceilings it does not require scale to hit.
Shopify’s own bundle documentation sets out the constraints:
- Bundles cannot contain other bundles.
- They are incompatible with subscriptions and pre-orders.
- They do not support import, export or bulk editing.
- If a component product’s SKU changes, you delete and recreate the entire bundle.
- Mix-and-match bundles need either a third-party app or the Shopify Plus plan and the Bundles API.
None of that is a criticism. Shopify is a storefront, and those are reasonable storefront constraints. But if your product is assembled, configured, or made to order, the configurator belongs in the system that knows your bill of materials — and that is an architectural decision, not a capacity one.
So what do you do about it
Answer three questions honestly.
- Can you state the landed cost of your best-selling SKU right now, without opening a spreadsheet?
- If two systems disagree about stock today, does anyone have to think about which one is right?
- Is reconciling between systems part of somebody’s actual job description?
If all three are comfortable, you have not outgrown anything, and buying an ERP will cost you money and goodwill to solve a problem you do not have.
If the third one is a yes, that is the honest trigger. Not revenue, not app count, not a product ceiling — the moment reconciliation became a role.
And if you do act on it, keeping Shopify is usually the right call. The storefront is rarely the thing that hurts. Running an ERP behind Shopify as the operational backbone is a smaller, cheaper, less risky project than replatforming, and it leaves the part that converts customers alone. Which system to put back there is a separate question — we work through the options, including where our own answer loses, in the best ERP for Shopify.
One practical note if Odoo is on your list: it has no Shopify connector built in, so you buy one, and that purchase probably drags your hosting with it. Odoo’s documentation states that Odoo Online cannot run third-party modules containing Python code, which every Odoo Shopify connector is — and VentorTech’s listing says outright that Odoo Online is not supported. Confusingly, the App Store’s own availability field on those same listings says otherwise, so this is worth getting in writing rather than assuming. It matters because hosting sets your Odoo plan tier, which sets your per-user rate — and Odoo prices by region, so check odoo.com/pricing from your own country rather than trusting a figure in an article. We compare the connectors in Odoo Shopify connectors compared.
Frequently asked questions
How do I know if I’ve outgrown Shopify?
Test whether you can answer questions, not whether you have hit a limit. If you cannot state the landed cost of an order without a spreadsheet, if purchasing lives in somebody’s inbox, if your team knows which system to distrust, or if reconciling between systems is part of somebody’s job, you have outgrown your back office. Shopify’s published ceilings — such as the variant throttle that begins at 500,000 variants — sit far above where most stores get into difficulty.
What are Shopify’s actual limits?
Shopify’s developer documentation states that pagination is limited to 25,000 objects, counts are accurate only up to 25,000, input arrays max out at 250 items, a single query may not exceed 1,000 points, and a variant-creation throttle begins at 500,000 product variants and does not apply on Shopify Plus. Admin API rate limits run from 100 points per second on standard plans to 2,000 for enterprise. Many of the product and variant figures quoted elsewhere online are stale or unverifiable.
Does a lot of apps mean I’ve outgrown Shopify?
No. App count on its own is not a signal. Several apps that agree with one another are a working stack. The problem begins when two systems hold the same fact and disagree about it, because then someone has to arbitrate, and that job never goes away on its own.
Should I replatform off Shopify or add an ERP behind it?
For most businesses the storefront is not the problem, so keeping Shopify and running an ERP behind it is the smaller and lower-risk project. Replatforming makes more sense when the storefront itself is the constraint — particularly with configurable or assembled products, where Shopify’s bundle model is restrictive and the configurator needs to sit with your bill of materials.
Is it possible I need nothing at all?
Yes, and it is common. A single-channel store with straightforward fulfilment can run on Shopify’s native multi-location inventory plus a purchase-order app for a long time, on a single stock count with no sync to fail. Revenue alone is a poor reason to add an ERP.
Working out where you actually are
The three questions above will tell you more than any threshold list. If the answers are uncomfortable, the next question is which system fits — and that depends far more on your entity structure and who will own it internally than on your order volume.
Talk to us about whether you actually need this →
About the author
Nguyen Tran is the founder of Ministers.io. He has worked on Odoo ERP implementations across ecommerce, retail, manufacturing, and maintenance, repair and operations (MRO) — five delivered directly and more than twenty advised on.
