Most businesses should start by looking at packaged ERP. Some should build. The deciding factor is rarely company size. It is how much of your process customers pay you for, and how much of it a package would have to bend to fit.
Default to buying, then look for reasons not to
Packaged families such as NetSuite, Microsoft Dynamics 365 Business Central, Odoo and Acumatica have spent years on the standard cases: ledger, payables, receivables, purchasing, basic inventory. If your operation looks like others in your industry, that maturity is worth more than a tailored screen. Rebuilding the general ledger, payroll or a tax engine is almost never justified. Buy those, and integrate.
Run a fit-gap before you decide
- List your ten to fifteen most important processes: quote-to-order, receiving, pick and ship, invoicing, returns, replenishment, month-end close and so on.
- Describe each one as it runs, including exceptions: split shipments, customer-specific pricing, unit conversions, rush orders.
- For each package you consider, mark every process as fit (works as shipped), configure (settings and workflow rules), customize (code or a paid add-on) or gap (cannot be done).
- Count the customize and gap rows, and note which ones touch the work customers pay you for.
If nearly everything lands in fit or configure, buy. If your core operations sit in customize and gap, price a custom build or a package plus custom extensions before you sign anything.
Signals that push toward custom
- Your quoting, routing or fulfillment logic is a reason customers choose you, and a package would flatten it.
- You pay for modules you never open, while the parts you need live in spreadsheets bolted onto the package.
- Several systems cannot exchange data, so people copy between them all day.
- Exceptions (special pricing, split shipments, unusual units of measure) take more effort than standard orders.
- Per-user license fees would become a large permanent cost as the team grows.
A three-year cost worksheet
Compare over three years, not at signing. Fill every line for every option, including the free one you are tempted to skip.
| Cost line | Packaged ERP | Custom build |
|---|---|---|
| Licenses or subscriptions | Per user, per module, often per year | None for your own code; licenses for tools you choose |
| Implementation partner | Usually required | Included in the build |
| Customization and add-ons | Priced separately, and may break on upgrades | Part of the scope |
| Data migration | Yours or the partner's | Included, rehearsed |
| Training | Vendor courses plus internal time | Built around your screens |
| Hosting and support | In the subscription or extra | Hosting plus a support agreement you choose |
| Upgrades | Vendor schedule, sometimes forced | When you decide |
Risks on each side
- Packaged: customizations can complicate upgrades, and vendor roadmaps set your priorities.
- Custom: you need an owner on your side and a plan for maintenance after launch, whether you contract it or hire for it.
- Both: a project without a named decision-maker slows down, whatever the technology.
The hybrid most companies overlook
Keep a packaged system for the commodity parts (accounting, payroll) and build custom software for the parts that differentiate you (order management, pricing, warehouse operations), joined by integrations. You buy reliability where the problem is standard and get a fit where it matters.