We implement all three, which means we have no particular reason to push one. What follows is the reasoning we use with clients, including the cases where we recommend against the platform someone arrived wanting.
They are not the same shape of product
Zoho is a broad suite of applications that integrate well and are quick to stand up. Odoo is a modular ERP with genuine depth in inventory and manufacturing. ERPNext is a complete open-source ERP with no per-user licence and a strong customisation model.
| Zoho | Odoo | ERPNext | |
|---|---|---|---|
| Best fit | Sales-led businesses, services, multi-app needs | Manufacturing, inventory-heavy distribution | Businesses wanting control and no licence lock-in |
| Licence model | Per user, per app or suite | Per user, per module | No per-user licence |
| Depth in production | Light | Strong | Strong |
| Customisation | Deluge scripts, low-code | Python modules | Custom doctypes and server scripts |
| Who should own it | Business admin | Technical partner | In-house or partner with dev capability |
Process fit beats feature lists
Every platform demo looks good, because demos are built around processes the software already handles well. The useful exercise is the opposite: take the three processes that cause you the most trouble today and ask to see those, with your own document types.
If a vendor cannot show your awkward process in the demo, you have not tested the thing that will decide the project.
The size of your finance team matters more than you expect
A two-person finance team needs the system to do more of the thinking: automated matching, defaults that hold, fewer places to make a decision. A larger team can absorb more configuration complexity and often wants more control.
This single factor changes our recommendation more often than industry does.
What drives cost over three years
Licences are visible and easy to compare, which is exactly why they get too much weight. Over three years the larger numbers are usually implementation effort, customisation you did not plan for, and internal time spent on data.
The cost people forget
Whoever owns the system internally will spend real hours on it every month. Budgeting zero for that is the most common planning error we see, on every platform.
When we advise against a platform
- Heavy discrete manufacturing on a suite with light production depth.
- No internal technical owner, on a platform that assumes one.
- A requirement to keep data on infrastructure the platform does not support.
- A team that wants zero customisation, on a product that needs some to fit.
None of these are faults in the software. They are mismatches, and they are much cheaper to spot before signing than after.
