Seven signs your store has outgrown installed apps

Nobody decides to build a custom Shopify app. What happens is that a store accumulates workarounds until somebody adds them up.

The signals below are the ones we see repeatedly in stores that turn out to need custom software. None of them is "the app we installed is missing a feature," because that one usually resolves by finding a better app. They are all operational: things the business is already doing, at a cost nobody has written down.

1. A spreadsheet has become production infrastructure

There is a sheet somewhere that someone updates every morning, and if they stopped, orders would go out wrong. It might map SKUs to a supplier's codes, or hold the rules for which products can ship together, or track which wholesale customer gets which price.

It is the clearest signal there is, because the sheet is a specification. Somebody has already worked out the logic and written it down. The only thing missing is that it runs on a person.

2. A rule lives in one person's head

Ask why an order was routed to a particular warehouse and the answer is a name. That person knows the region rules, the exceptions for oversized items, and which customer always gets expedited shipping regardless.

This is a business continuity problem before it is a software problem. The cost is invisible until they take a holiday.

3. You are paying several subscriptions to approximate one behaviour

A bundling app, a discount app, a shipping-rules app, and a middleware tool to make them agree — none of which quite does the thing, so the combination does most of it, and the seams show at checkout.

Worth doing the arithmetic. Add the monthly cost, then add the time spent every month reconciling their disagreements. Stacks like this frequently cost more per year than building the specific thing once, and the stack keeps charging.

4. Your data lives in two places and they disagree

Inventory in Shopify and inventory in the ERP, drifting apart. Somebody reconciles them, usually manually, usually after a customer notices.

This one often does not need an app at all — it needs an integration built properly, with idempotent syncs and retries. But it belongs on the list because it is one of the most expensive symptoms to leave alone: every hour of drift is a potential oversell.

5. An app's roadmap has become your roadmap

You have requested the same feature more than once. Support is friendly and non-committal. Your launch depends on something that is on somebody else's backlog, behind priorities you cannot see.

Note that this is not a criticism of the app. A public app has to serve every store that installs it, so a request specific to your operation is correctly deprioritised. It is just that "correctly deprioritised" and "we still need it in March" are both true.

6. Onboarding a new person takes weeks because the process is not written anywhere

The process exists — it is just distributed across four tools, two spreadsheets, and a Slack channel. New hires learn it by asking.

Software is a way of writing a process down that cannot be forgotten. That is often the real deliverable, and it is worth more than the feature list that gets scoped.

7. You have started avoiding growth because of the operational cost

The strongest signal, and the least often said out loud. A new sales channel would mean more manual reconciliation, so it is postponed. A wholesale line would need pricing rules nobody wants to maintain, so it stays a conversation.

When the operational cost of a change is what stops it, the operation is the constraint.

The counter-test

Three signals is not a mandate to build. Before scoping anything, ask the opposite question honestly.

Has anyone searched the App Store recently? Not two years ago. The category moves, and building something that now exists as a well-maintained app for a monthly fee is a bad trade.

Would a Shopify Function do it? If the requirement is really about pricing, discounts, checkout validation or delivery options, a Function is a much smaller thing to own than an app.

Is the process worth automating as it stands? Automating a bad process makes it permanent. Sometimes the right answer is to simplify the rule, not to encode it.

If those come back negative and the symptoms are still there, the case is real — and it is worth writing down before anyone estimates anything. What the software must do, what it must never do, who owns it afterwards, and what happens when Shopify's API version changes next quarter. That is the conversation we start with.