Theme, custom theme, or custom app: an honest decision guide

Every Shopify build sits somewhere on a ladder: a theme as it ships, a theme customized or built for you, or custom software around the store. Each rung up buys capability and costs money and maintenance. Since we sell the upper rungs, our incentive is to talk you upward — so this guide is deliberately written to talk you down to the cheapest rung that actually solves your problem. That's the same filter we apply to our own enquiries.

What each rung really is — and really costs

A theme, as shipped. One-time purchase, working storefront in days, and — the under-priced part — the theme's developer maintains it: platform changes, bug fixes, new features arrive as updates. In money, it's the cheapest rung by an order of magnitude; in maintenance, it's nearly free. What you trade away: anything the theme's settings can't express.

A customized or custom theme. This rung has a trap in it that matters more than any price: the moment a theme's code is edited, its update path usually breaks — you now own a fork, and the maintenance that was the theme developer's problem becomes yours. Light customization (settings, sections, an app block or Custom Liquid drop-in) avoids the trap entirely; real code changes should be treated as adopting a codebase, not buying a tweak. Budget-wise, think in multiples of a theme, plus ongoing care.

Apps — off-the-shelf capability, rented. Subscriptions shift cost from one-time to monthly, forever. A good app is usually far cheaper than building the same thing — the discipline is the audit: each app's monthly price, and its performance price, against what it actually does for revenue.

Custom software. The expensive rung in money and in relationship: real engineering is quoted per project (anyone quoting a flat price before understanding your workflow is guessing), and it implies maintenance for as long as the system matters. What it buys is the thing no template sells: software shaped like your operation — dealer pricing that lives in a system instead of a spreadsheet, labels printed while the show runs, a catalog managed through pipelines instead of heroics.

The decision, by signals

You don't need custom work if:

• Your problem is presentation — how the store looks and converts. That's the theme rung, possibly with better merchandising, and it's fine to stop there.
• One well-reviewed app already solves it. Rent it. Building a worse version of an existing app is the most expensive form of pride.
• The workflow behind the problem is still changing weekly. Custom software freezes assumptions; freeze them after the workflow settles, not before.
• The budget only stretches to starting. Half-built custom systems are worth less than zero — they consume attention and deliver nothing.

You've outgrown the lower rungs if:

• People are the integration layer: the same data typed into multiple systems, spreadsheets bridging tools, pricing applied from memory.
• Your selling model itself doesn't fit the platform's shape — account-specific catalogs and terms, live-selling operations, approval workflows.
• The "app stack" solving your problem is five apps deep, none quite fits, and the monthly total has quietly become a salary.
• The risky part of your operation is invisible: no audit trail on money-touching actions, stock levels everyone estimates, orders lost in inboxes.

Notice the pattern: the lower rungs solve storefront problems; custom work earns its cost when the problem is operational — and operational pain shows up in your team's hours, not your conversion rate.

The order of operations we actually recommend

1. Fix the data and workflow first (free, and every later rung benefits).
2. Exhaust theme settings and app blocks before touching theme code.
3. Rent capability where a good app exists; audit the stack yearly.
4. Build custom when the remaining problem is genuinely yours — shaped by how your business differs — and the manual cost of living with it is a number you can name.

If you can't tell which side of the line you're on, that's normal — most real situations touch several rungs. Describe the problem, not the solution: what the work looks like today, where it hurts, what you've tried. An engineer reads it and tells you within one business day which rung we'd honestly recommend — including, regularly, the rungs we don't sell.