The case for off-the-shelf, made sincerely
ERP platforms encode decades of operational best practice. Odoo covers accounting, inventory, CRM, HR, and manufacturing from roughly $25–$60/user/month with a massive module ecosystem; NetSuite handles serious multi-entity accounting at $1k–$3k+/month base plus per-user fees; Dynamics 365 fits Microsoft-centric enterprises.
Buying gets you battle-tested compliance logic (tax, depreciation, audit trails), a hireable talent pool that already knows the system, vendor-maintained security and updates, and community answers to nearly every question.
If your operations are conventional, you buy things, hold inventory, invoice customers, pay staff, a custom build recreates solved problems at your expense. That's not a gray area. Chalk Labs turns down custom ERP requests that describe a standard Odoo deployment, because building one would waste the client's money and everyone's year.
Where off-the-shelf quietly becomes custom-priced
The honest complication: 'off-the-shelf' pricing assumes off-the-shelf operations. The moment your processes deviate, you enter implementation-consultant territory, NetSuite implementations routinely run $25k–$100k+, Odoo customization projects $10k–$50k, and the customizations create upgrade friction that recurs with every version.
This is the pattern to watch: an ERP where every meaningful workflow lives in custom modules, maintained by consultants at $150–$250/hour, on a platform whose license you're also paying. You've bought a custom system's costs with an off-the-shelf system's constraints.
The 40% heuristic follows from this. If scoping reveals you'd customize less than 40% of the system, buy and adapt your processes where reasonable. Past 40%, the platform is scaffolding you're fighting, and a purpose-built system becomes cheaper and cleaner.
The operations that genuinely need custom
Some operating models don't map onto ERP assumptions at any customization budget. Crypto-native companies are the canonical case: treasuries holding fiat and tokens across multiple entities and chains, revenue arriving in stablecoins, token vesting schedules as a core liability class, and compliance reporting no NetSuite module contemplates. QuickBooks and standard ERPs structurally cannot represent this.
Other qualifying patterns: marketplaces with exotic settlement logic, businesses whose core workflow is their competitive advantage (encoding it in a vendor's system leaks the moat), and companies needing deep integration with proprietary internal tools.
A custom ERP scoped to actual operations, not a NetSuite clone, but the 30% of ERP surface you use, built around your real flows, runs $20k–$60k at AI-assisted development rates. Chalk Labs builds these primarily for crypto and AI-native startups whose back office outgrew spreadsheets but never fit the templates.
- Multi-entity, multi-currency (fiat + token) treasury operations
- Core workflows that are the competitive advantage
- Settlement or vesting logic no vendor module models
- Deep integration with proprietary internal systems
A decision path you can run this week
Step one: write down your ten most important operational workflows, honestly, as they actually happen. Step two: demo Odoo (it's the cheapest lens) against that list, and mark each workflow as native-fit, configurable, or requires-custom-development.
If seven or more are native or configurable, deploy off-the-shelf and adapt, revisit in two years. If four or more need custom development, get a custom scope priced before signing anything; the comparison might surprise you. If the blocker is crypto treasury or token operations specifically, skip the demo, the answer is already custom or purpose-built crypto accounting tools, and which one depends on how much beyond accounting you need.
Either way, refuse both vendors' worst pitch: the ERP consultant's 'we can customize anything' and the dev shop's 'we can rebuild NetSuite.' Both sentences are technically true and financially catastrophic.
Questions we hear about this
Yes, for conventional operations (invoicing, inventory, HR, standard accounting), Odoo at $25–$60/user/month with light configuration is hard to beat, and building custom would recreate solved problems. Custom only wins when operations structurally don't fit.
A scoped build covering your actual workflows, not a full NetSuite clone, runs $20k–$60k at current AI-assisted development economics, typically delivered in 4–8 weeks, plus 15–20% annual maintenance. Compare against implementation-plus-license TCO, not license price alone.
If scoping shows you'd need custom development for more than 40% of an off-the-shelf system's workflows, a purpose-built system is usually cheaper and cleaner than fighting the platform. Below 40%, buy and adapt your processes.
Their data models assume fiat, single-ledger accounting. Crypto operations involve multi-chain token treasuries, stablecoin revenue, vesting-schedule liabilities, and multi-entity flows that standard modules can't represent, making crypto startups the most common legitimate custom-ERP case.