Isometric illustration: a block of identical prefabricated cubes with two custom-made white modules fitted into carved-out slots, joined by electric-blue connectors.
Software Development
ERP
Odoo
CRM
Custom software
Comparison

Custom ERP or Odoo: when each one makes sense

When Odoo is enough, when a custom ERP comes out cheaper, and when the hybrid path makes sense, with a real integration case.

Mario VelázquezNovember 18, 202610 min

The question almost always arrives framed as a choice: do we set up Odoo, or do we have our own system built?

That is the wrong framing. And more often than you would expect, the right answer is "both, in this order".

This article tries to give you the criteria to decide it without a sales meeting in the middle.

What Odoo solves well out of the box

Odoo is a modular ERP with twenty years behind it and a huge community. That is not decoration: it means problems that are well understood already come solved.

Where it gets there on its own, and gets there well:

  • Accounting and invoicing. Chart of accounts, journal entries, bank reconciliation, Mexican e-invoice stamping (CFDI) through the Mexico localization. Nobody should build that from scratch.
  • Inventory and warehouses. Receipts, deliveries, lots, serial numbers, multiple locations, cost valuation.
  • Standard purchasing and sales. Quote, order, receipt, invoice. The classic cycle works on day one.
  • Point of sale. Register, sessions, end-of-day closings, cash counts.
  • Payroll, projects, maintenance, manufacturing. Mature modules that cover 80% of what a midsize company needs.

There is something else people underestimate: Odoo ships a standard, well-documented API over XML-RPC and JSON-RPC. That makes it a good citizen in an architecture, not a black box.

Buying what already exists is almost never the mistake.

What it does not solve

Odoo is built on one assumption: your process looks like the average process of thousands of companies. When that is true, you win. When it is not, the fighting starts.

The places where we see it get stuck:

  • Business rules that are yours and nobody else's. A commission formula with seven conditions, a pricing scheme based on volume and customer tenure, an approval flow that depends on the order's margin.
  • Shop-floor operations that do not fit in a form. Service orders with photos, signatures, downtime, and spare parts that get decided halfway through the repair.
  • An experience for people who do not work in an office. A field technician, an end customer, a driver. An ERP interface is designed for a data-entry clerk sitting at a desk.
  • Customer portals that look like your brand. It can be done, but only by fighting the product's theme.

None of those is a flaw in Odoo. They are the edges of a packaged product.

When the off-the-shelf ERP gets expensive

This is where the math flips, and it almost never shows up in the initial proposal.

Customizations that fight the product

There are two ways to extend an ERP. One is to add on top of what the product already anticipates: fields, reports, declared workflows. That kind is cheap, and it survives.

The other is to rewrite the base behavior because your process does not match. Overriding core methods, changing how inventory is valued, altering the logic of a standard module.

That second kind gets quoted as if it were "just one more customization". It is not. Every line written there is debt you pay on every upgrade, forever.

The practical rule: if the consultant has to explain why they need to "touch the core", that part is no longer Odoo. It is custom software with an ERP layer on top, and none of the advantages of either.

Upgrades send the bill

A packaged ERP ships new versions, with security patches, tax compliance updates and new functionality. Staying on an old version is not free; it is risk piling up quietly.

But every major upgrade stress-tests everything written on top of it. What leaned on the product survives with minor adjustments. What fought the product breaks.

We have seen companies stop upgrading because the cost of the migration grew larger than the benefit. They stay frozen on a version from four years ago, with a consultant who is long gone and a module nobody dares to touch.

That is the truly expensive scenario. And it arrived by accumulation, not through one bad decision.

The license grows with your team

Odoo is distributed in two editions. Community is open source and charges no license fee; you pay for hosting, implementation, and whoever keeps it running. Enterprise charges a license per user, per period, and modules and services are added on top depending on the plan.

We are not going to put Odoo figures here, because they change by country, by version and by negotiation. What we can describe is the shape of the cost, which is what decides the three-year result.

  • Per-user license: variable, grows as your headcount grows.
  • Hosting and infrastructure: recurring, grows with your data.
  • Initial implementation: one-time payment, almost always the largest of the first three.
  • Customization and support: recurring, for as long as the system stays alive.

That first line is the one you need to project. A system that costs X with ten users costs double with twenty, and you do not control the unit price.

Ask for the three-year quote with your projected headcount, not today's.

When Odoo is clearly the answer

This section has cost us projects, and we are keeping it anyway.

If your case is on this list, commissioning a custom ERP is an expensive mistake:

  • You need standard accounting, invoicing and inventory. That problem is solved. Rebuilding it is paying to reinvent twenty years of product.
  • You need to be operating in weeks, not months. A packaged ERP starts with configuration; a custom one starts with definition.
  • Your process is still changing. If the workflow gets rearranged every quarter, a system of your own freezes it too early.
  • You do not have the means to sustain a system of your own. Custom software is not delivered, it is operated. Without a maintenance budget, it dies precisely when you already depend on it.
  • Tax compliance is the heart of the project. The SAT (Mexico's tax authority) changes the rules, and someone has to keep up with them. Let that be the product vendor.

In those cases the argument is not about price. It is about who should carry the maintenance of a problem that is not yours.

When custom wins

The reverse case exists too, and it has clear signals.

  • Your competitive edge is in the process. If the way you operate is why customers buy from you, forcing it into a standard mold makes you the same as everyone else.
  • The customization list already touches the core. When half the implementation is rewriting base behavior, you are already paying for custom development, with a surcharge.
  • The main user is not in the office. Technicians, delivery drivers, end customers. They need a screen designed for their context, not an ERP form.
  • You need one module, not an ERP. Many companies do not need an ERP; they need to sort out service orders, or quotes, or shop-floor control.

At Avanzia, custom development starts at $4,000 USD, and a first working version usually lands between $4,000 and $15,000 USD. Maintenance afterwards starts at $1,000 USD per month and varies with the scope.

We publish the range on purpose. If it is far from what you had in mind, we just saved you three meetings. We break it down in full in how much custom software costs.

That number buys ownership of the code, the data and the roadmap.

The hybrid path

The option almost nobody proposes in the first meeting is usually the right one: Odoo as the base, custom development on top.

The idea is simple to state and takes discipline to sustain. Odoo keeps the standard work and gets touched as little as possible. What is yours gets built separately and talks to Odoo through its API.

That way the ERP stays upgradable, because nobody rewrote its core. And your own process lives in code you actually control, without fighting the product.

That is what we did with DAGOS and with Welding Republic.

What it looks like in practice

DAGOS is a Mexican retailer and distributor that runs on Odoo. Its historical volume was past 11,000 point-of-sale orders and 7,800 invoices, across a catalog of 835 products with more than 2,000 variants.

The ERP stored everything correctly. The problem was a different one: reporting took hours, and inconsistencies between the catalog and sales were caught weeks later.

We did not migrate anything. We connected Odoo to Supabase and an analysis layer built on Claude, and the system was in production in three days. Today it syncs every hour, checks the deltas, and alerts the director when something does not add up: products selling below cost, inventory that does not reconcile, SKUs with abnormal turnover.

The most quotable result is that it prevented stockouts in 22% of its key products. The director stopped asking for reports and started receiving them.

Welding Republic already had a WhatsApp bot answering customers, but the bot could not see the ERP. When someone asked about availability or a delivery date, a human rep answered by hand.

We connected that bot to Odoo using the same pattern. Stock questions were about 60% of message volume, and they stopped landing on the team.

Neither of them changed platforms. Both gained what the product did not give them.

How to decide it in one afternoon

Three questions to run against your own operation, with no vendor in the room.

How much of your process is really yours?

Write down the processes that hold the company up. Mark the ones any competitor would run the same way, and the ones only you run like that.

If almost everything falls in the first column, it is Odoo. If there are two or three processes no product anticipates, that is where custom development goes.

How many of the customizations touch the core?

Ask the implementer to split the list in two: what gets configured, and what gets programmed over the base behavior.

That second list is the thermometer. If it is long, the packaged ERP is already too small for you before you start.

What does it cost over three years?

Add up licenses, hosting, implementation, customization and support. Project your users; do not count today's.

Compare that against the upfront investment plus monthly maintenance on the custom side. An honest comparison needs both columns filled in.

Frequently asked questions

Can we start on Odoo and migrate later?

Yes, and it is usually the right order. You start fast with the standard pieces, learn what is missing by using it, and build custom only that part.

Odoo Community or Enterprise?

It depends on which modules you need and whether you want support from Odoo itself. Community charges no license fee, but it shifts the upkeep to you or your provider.

Do you implement Odoo?

We work on top of Odoo, integrating and extending it, which is different from selling licenses. If what you need is a pure implementation, we will tell you in the discovery call.

What happens to the data I already have?

It gets migrated in both scenarios, and it always costs more than people expect. What moves the schedule is not the programming; it is how clean that data is.

Is the custom code mine?

Yes. Code, data and repository belong to the company paying for the development. No arrangements where you are tied to the vendor to touch your own system.

The uncomfortable conclusion

The useful question is not which of the two is better. It is which part of your operation deserves software written for it.

For almost everything else, there is already a product that has been refined for twenty years and costs a fraction. Fighting that product is the most expensive way to end up with a custom system.

If you want the math done with your own numbers, discovery is free. And if it turns out Odoo alone is enough for you, we will tell you on that same call.

Related articles