Nicolas · 18 Oktober 2026
Choosing between a ready-made ERP and a custom-built one is one of the most expensive software decisions a growing business makes, because ERP touches inventory, sales, finance, and operations all at once and is painful to replace later. This guide compares the real 2026 costs of both paths, shows where the break-even point falls, and explains when building custom is the smarter move.
For small teams with standard processes, an off-the-shelf ERP such as Odoo is almost always cheaper over three years. A custom ERP starts to win when your user count grows, when your processes drift far from the standard, or when you look at a five-year horizon, because custom software has no per-user licence fee and you own the code. In a published 2026 scenario by Kolonell, the 36-month crossover lands at roughly 60 users. Under the same assumptions, it moves to about 34 users if the upfront build costs $50,000 instead of $82,000.
These figures come from a New York scenario published by Kolonell in September 2026. They are orders of magnitude for one market and should be read as a benchmark, not a quote.
Off-the-shelf (Odoo Enterprise)
Licence: $27 to $41 per user per month.
Implementation by a partner: $16,000 to $65,000.
Custom modules on top: $3,500 to $22,000.
Time to launch: 6 to 14 weeks.
Code ownership: stays with the vendor.
Custom build
Initial build: $50,000 to $130,000.
Support: $550 to $1,650 per month.
Hosting: $45 to $330 per month.
Time to launch: 8 to 20 weeks.
Code ownership: yours.
Other sources give different ranges, because scope matters more than anything else. A CodeStringers analysis puts a small custom ERP (three to five modules, single site) at $80,000 to $200,000 in the US market, and notes that about 60 to 70 percent of any custom ERP budget is developer time. Both views agree on the structure: off-the-shelf is cheap to start and keeps charging per user, while custom costs more upfront and then stays flat.
A note on “free” options: Odoo Community has no licence fee, but Kolonell still budgets roughly $16,000 to $40,000 for implementation, hosting, and maintenance.
Kolonell’s model assumes Odoo at $34 per user per month with $38,000 implementation and $11,000 in custom modules, against a custom build of $82,000 with $1,000 per month support and $130 per month hosting. Over 36 months:
10 users: Odoo $61,240 vs custom $122,680.
20 users: Odoo $73,480 vs custom $122,680.
40 users: Odoo $97,960 vs custom $122,680.
60 users: Odoo $122,440 vs custom $122,680 (near even).
80 users: Odoo $146,920 vs custom $122,680.
120 users: Odoo $195,880 vs custom $122,680.
The custom cost stays flat because it has no per-seat fee. The Odoo cost climbs with every person you add.
In the model above, the three-year crossover sits near 60 to 62 users. Kolonell notes that two things move it sharply: how many custom modules you pile onto the off-the-shelf product, and how long you plan to run the system. Over five years, they report that the build almost always wins, because the licence keeps running.
Their worked example is a 45-person industrial firm with 42 software users. At 36 months, Odoo comes to $105,384 against $128,680 for custom, so Odoo is cheaper. At 60 months, it flips: $157,944 for Odoo against $155,800 for custom.
The break-even depends heavily on what the custom build costs. Using the same Kolonell assumptions and changing only the upfront build price, our own arithmetic gives the following 36-month crossover:
Build at $82,000: about 60 users.
Build at $50,000: about 34 users.
Build at $30,000: about 18 users.
The calculation is simple. Custom cost over 36 months is the build price plus $40,680 (36 months of support and hosting). Odoo cost is $49,000 (implementation plus modules) plus $1,224 per user. Set them equal and solve for users. This is an illustration, not a quote, but it shows why the cost of the team building your system matters so much.
This is where hiring an Indonesia-based team becomes relevant. Developer rates differ widely by region: CodeStringers reports agency rates of roughly $25 to $50 per hour in Asia against $120 to $220 in North America. For Indonesia specifically, GoodFirms lists a median of about $37 per hour for web development companies. Since labour is the majority of a custom ERP budget, a lower rate can shift the break-even to a much smaller team size. Judge a team on portfolio, process, and contract terms as well, not rate alone.
Code ownership. With custom software, the code is yours to adapt, extend, or resell. With a licensed product, you rent it.
Vendor lock-in. A licence increase of 10 percent per year, Kolonell notes, adds thousands of dollars every year on a large team.
Process fit. Packaged ERP fits the average company. The further your processes drift from the standard, the more you pay in workarounds. Kolonell points to a rough threshold: when processes drift more than 30 percent from the standard, a custom build catches up faster.
Speed. Off-the-shelf wins on time to launch and instant functional depth (accounting, CRM, inventory out of the box).
CodeStringers lists concrete signals that you have outgrown a packaged system:
Your process is a competitive advantage, such as proprietary pricing or a non-standard fulfillment flow.
The numbers everyone trusts live in a spreadsheet that one person rebuilds by hand.
You pay for packaged software that your team routes around.
Your data is split across systems that do not talk to each other.
You need multi-entity, multi-site, or multi-currency support that packaged tiers charge heavily for.
If none of these apply, a well-chosen off-the-shelf system will serve you better and cost less.
You do not have to choose one extreme. Kolonell calls the mix often optimal: keep standard Odoo for accounting and CRM, and add an API-connected custom module for the process that makes your business different. CodeStringers makes a similar argument, advising companies to buy the generic modules (general ledger, payroll, standard HR) and build only what encodes how they actually make money. In their comparison, integrating best-of-breed tools costs roughly $30,000 to $150,000, with 6 to 16 weeks per integration.
Gartner predicts that by 2027, more than 70 percent of recently implemented ERP initiatives will fail to fully meet their original business goals. It names strong alignment with corporate strategy as one of the top predictors of success. In practice, you reduce risk by:
Mapping your real processes before choosing an approach.
Treating data migration as its own phase with its own budget.
Shipping in increments, one module at a time, instead of one large go-live.
Building only the differentiating modules and buying or integrating the rest.
Budgeting for the second year: CodeStringers reports maintenance commonly runs 15 to 25 percent of the build cost annually.
If you consider outsourcing an ERP build, ask these questions before signing:
Time zone. Indonesia (WIB, UTC+7) overlaps fully with Australia and Singapore, partially with European mornings, and mostly asynchronously with the US, according to Second Talent. Ask how the team handles documentation and weekly demos.
Ownership. Require a contract that transfers source code and design files to you.
Scope and change process. Fix the scope at the process level and agree on how changes are priced.
Support after launch. Agree on a maintenance plan before the build begins.
Genvori Tech is a software house in Surabaya, Indonesia. We build custom ERP, CRM, and web applications with Next.js, React, Laravel, Supabase, and PostgreSQL, and automate workflows with n8n. Every project starts with business analysis and architecture planning, moves into iterative development with regular demos, and ends with launch and ongoing support. Our work includes a custom ERP system for CV Andicha Pratama, built with Next.js, Tailwind CSS, and TypeScript.
We recommend custom only when it fits. If your needs are standard and your team is small, we will tell you to start with an off-the-shelf product.
Not at the start. In Kolonell’s New York scenario, Odoo is cheaper over 36 months until roughly 60 users. Custom becomes cheaper as the team grows, as customization needs increase, or over a five-year horizon, because it has no per-seat licence.
In Kolonell’s standard 36-month scenario, around 60 users. That drops to about 34 users if the build costs $50,000 instead of $82,000, using our own arithmetic on the same assumptions. It can happen even earlier if your processes differ heavily from the standard.
Kolonell’s 2026 scenario puts support at $550 to $1,650 per month. CodeStringers suggests budgeting 15 to 25 percent of the build cost per year.
Kolonell estimates 8 to 20 weeks for a targeted build, against 6 to 14 weeks for an Odoo implementation. A build focused on one critical process can ship in 8 to 10 weeks. Full multi-department systems take longer.
Yes, and it is often the best option. Use a standard product for accounting and CRM, and build a custom module for the process that differentiates your business.
You should. With a custom build, make sure the contract transfers the source code and design assets to you after final payment.
Kolonell, ERP Build vs Buy in New York: The Honest 3-Year TCO for 2026, September 15, 2026.
CodeStringers, Custom ERP Development: When It’s Worth It, What It Costs, and the Smarter Middle Path, July 10, 2026.
Gartner, What IT Leaders Must Do to Avoid Disappointing ERP Initiatives.
GoodFirms, Top web development companies in Indonesia.
Second Talent, IT staffing in Indonesia.