You're evaluating SAP, Odoo, and a custom build. The off-the-shelf options don't quite fit — there are two or three workflows that are core to how you operate, and neither platform handles them the way you need. The custom build sounds expensive and slow. So you're stuck, evaluating proposals that all feel like the wrong answer.
The reason you're stuck is that the question is framed wrong. It's not SAP vs. Odoo vs. custom. The real question is: is your process the differentiation, and if so, what does it cost to build that differentiation vs. what does it cost to abandon it?
The stakes
The wrong choice has a specific cost. Off-the-shelf ERP at enterprise scale runs $200k–$500k per year in licensing, plus $150k–$400k in customization and integration work upfront, plus ongoing consultant fees every time you need the platform to do something it wasn't designed to do. You pay those costs continuously, and you never own the result.
A misscoped custom build runs 18–24 months to deliver what you needed in 6, costs 2–3x the initial estimate, and sometimes doesn't survive to production. The failure mode here is real — it's not hypothetical. Custom builds fail when the scope drifts, the technical partner changes, or the requirements turn out to be more complex than anyone admitted.
The specific cost of the wrong call: being locked into either path for 3–5 years. ERP migrations are expensive and disruptive. Custom rebuilds are expensive and disruptive. The decision you make now is the decision you live with until the next operations cycle.
One place we've seen this resolved correctly: Cadrivit, a manufacturing SaaS engagement where purpose-built tooling delivered 70–80% operational cost savings and came in 30% ahead of schedule. The off-the-shelf options existed. They weren't the right answer for a workflow that was genuinely differentiated.
When off-the-shelf genuinely wins
Off-the-shelf wins in three specific situations, and it wins convincingly. If you're in one of these situations, stop evaluating custom and pick the right platform.
Your process is industry-standard. Accounts payable, accounts receivable, basic inventory, standard procurement — these are solved problems. SAP, Oracle, and mid-market alternatives like Odoo and NetSuite have spent decades optimizing for them. A custom build for standard processes is a decision to maintain something that doesn't need to exist.
The workflow gap is small and tolerably worked-around. If the off-the-shelf option covers 90% of your operations and the remaining 10% can be handled with a documented manual process or a simple integration, the math almost always favors the platform. Custom software has ongoing maintenance costs — every version upgrade, every new hire who needs to learn the system, every bug that only you can fix. If the workflow gap isn't significant enough to justify those costs, it isn't significant enough.
You need it running in three months. Custom builds take time. If you need full ERP functionality in a short window and you have the budget for the platform costs, off-the-shelf with customization is the faster path. You can always migrate later. You can't manufacture time.
The test: could your competitors buy the same platform and run the same workflows? If yes, the platform is not your differentiation. Get the platform.
When custom wins
Custom wins under three conditions, and all three need to be true.
The process is the differentiation. The way you do the work is why customers choose you, pay you more, or can't easily switch. Your workflow isn't industry-standard — it's specific to your business model, your customer commitments, or your operational constraints. The off-the-shelf platforms handle the general case; your case is specifically not general.
This is a genuine test. Many businesses believe their processes are differentiated when they're actually just unfamiliar with what the platforms can do. The way to test it: bring in someone who knows the platforms well, map your workflows, and ask them to show you how the platform handles each one. The gaps that remain after that exercise are the real gaps.
The core workflow isn't covered. Not "the platform doesn't do it exactly the way we do it" — that's almost always configurable. The test is: does the platform's conceptual model of the problem match yours? If the platform models inventory as a count and your business models inventory as a set of tracked serialized items with provenance records, that's a conceptual mismatch. Configuration won't fix it. You're either bending your process to the platform's model, or you're building something purpose-built.
The build budget is right. Custom software has a cost floor below which it doesn't get built well. That floor, for a serious operational system, is in the range of $150k–$400k for the initial build, with ongoing maintenance budget of 15–20% annually. If that budget isn't available, the custom build conversation is premature. The question becomes: can you get to the budget, or should you accept the platform with its limitations?
What "custom" actually means in this context
Custom doesn't mean starting from zero. It means purpose-built on modern tooling, scoped to the specific workflows that need custom handling.
The typical pattern: the core platform capabilities (financial reporting, user management, audit trails, integrations with standard accounting software or ERPs) are built on established frameworks and libraries. The differentiated workflows — the ones the off-the-shelf platforms can't handle — are built custom and integrated with the platform capabilities.
This is meaningfully different from "we built it in Rails and everything is custom." It's also meaningfully different from "we bought SAP and customized it until it breaks." It's a purpose-built system where the standard parts use standard solutions and the differentiated parts are actually differentiated.
The result: a system that looks like your business, not a platform you've bent into the shape of your business.
Total cost of ownership comparison
The comparison that matters isn't license cost vs. build cost. It's 5-year total cost of ownership.
Off-the-shelf (enterprise tier, manufacturing/operations context):
- Licensing: $200k–$500k per year
- Initial implementation and customization: $150k–$400k
- Ongoing consultant fees for customizations and upgrades: $50k–$150k per year
- 5-year cost: $1.4M–$3.5M
Off-the-shelf (mid-market, Odoo/NetSuite range):
- Licensing: $20k–$80k per year
- Initial implementation: $30k–$100k
- Ongoing maintenance and customization: $20k–$50k per year
- 5-year cost: $250k–$750k
Custom (purpose-built, serious operational system):
- Initial build: $150k–$400k
- Hosting and infrastructure: $10k–$30k per year
- Ongoing maintenance and evolution: $30k–$80k per year
- 5-year cost: $400k–$900k
The custom build isn't always cheaper. At mid-market scale with standard processes, off-the-shelf frequently wins on total cost. At enterprise scale with differentiated workflows, custom typically wins — and the licensing cost savings often pay for the build within 18–24 months.
The Cadrivit case: purpose-built manufacturing operations tooling delivered 70–80% operational cost savings against the previous combination of off-the-shelf platforms and manual processes, and came in 30% ahead of schedule. The specific savings came from eliminating the integration friction between platforms that couldn't talk to each other and the manual reconciliation work that friction created.
How to structure the decision
If you're in the evaluation phase, the framework is:
-
Map every core workflow. Not high-level — the actual steps, the data that moves, the decisions that happen, the exceptions that occur.
-
Map each workflow against your top two platform candidates. For each workflow: does the platform's conceptual model match? Is the gap configurable or structural?
-
Classify the gaps: tolerable workaround, configurable with effort, or structural mismatch.
-
Count the structural mismatches. If there are zero, buy the platform. If there are one or two and they're not on critical paths, probably buy the platform. If there are three or more and they're on critical paths, you have a custom build case — but only if the budget is right.
-
Get real build estimates for the custom scope. Not ballpark — scoped estimates from a team that's built similar systems.
The decision is usually clear after step 4. The teams that stay stuck in evaluation are usually stuck on step 3 — they haven't done the detailed workflow mapping that would let them actually count the structural mismatches.
What this looks like when it's resolved
You're running the system you built for your business, not the business your system was designed for.
Concretely: the core workflows run without workarounds. The data model reflects your domain — your inventory, your production process, your cost structure — not a generalized manufacturing domain that you've adapted. Reporting answers the questions you actually ask, not the questions the platform assumed you'd ask.
New workflows get added by your team or your technical partner, not by waiting for the platform's next release cycle or paying a platform consultant to configure something.
Your operational data is yours, in a schema you understand, exportable and portable. You're not locked to a platform whose pricing or support model changes in year three.
This is for you if
You're running an operations-heavy business — manufacturing, logistics, professional services, construction — and you're at the evaluation stage for an ERP or operations platform replacement. The off-the-shelf options have left you with specific, identifiable gaps that you haven't been able to get answered satisfactorily.
Typical engagement scope: $50k–$200k for purpose-built operational tooling, depending on the number of workflows and integrations. The engagement starts with the workflow mapping exercise — if it surfaces that off-the-shelf is the right answer, we'll tell you so.
This is not for businesses with standard processes that just want a cheaper or faster platform. It's for operations where the process genuinely is the differentiation, and where building to that process is the financially and operationally correct call.
The decision deserves a careful analysis. Get it wrong and you're re-platforming in two years.