For an Italian SME, the question "do we build a custom ERP or use TeamSystem/Zucchetti/Mexal?" is one of the most prejudice-loaded. Some say custom is always better, others swear off-the-shelf is always safer. Both wrong. The correct answer is: it depends on three variables.
Variable 1: how "standard" your process is
If your order-warehouse-invoice flow is what 70% of companies in your sector do, there is already a package built for you. Adapting it is cheap; maintaining custom would cost triple to do the same thing.
If your process has real specificity — a non-obvious production cycle, complex pricing rules, integrations with proprietary machines, sector-specific regulations — generic packages will force you to bend, lose efficiency, and still pay for invasive customisations.
Variable 2: how much you want to control the data
With a standard ERP, data lives in the vendor's database. Extracting it for advanced analytics is possible but always limited. Automatic updates can shift schemas.
With custom (typically PostgreSQL + Node/Java/.NET app), data is yours: controlled schema, free BI access, no surprises from a vendor release.
For companies that lean on analytics or ML on internal data, this variable often decides.
Variable 3: time horizon
An off-the-shelf ERP has a predictable 5-year cost: licence + maintenance + a few customisations. A custom ERP has a higher upfront cost but a different curve: after 3 years TCO converges, after 5 it can be lower.
If your company grows predictably and the process is stable, custom pays. If you are evolving fast and the business model might change, packages give you "throw it away" flexibility without pain.
The hybrid model that often wins
The pattern we most often suggest is hybrid:
- Standard ERP for masterdata, accounting, warehouse, e-invoicing. Where regulation changes often and you do not want to keep up.
- Custom application for the "core business" piece: pricing, product configurator, sales workflow, customer integrations.
- Sync middleware between the two, which becomes the critical piece to design well.
More expensive to architect but combines the best of both: standard where standard helps, custom where it creates competitive advantage.
The four questions we ask before recommending
- How many ERP users will you have in 3 years?
- Is your order-to-invoice process similar to your competitors', or does it have unique aspects?
- Do you need to analyse internal data for operational decisions?
- Does your in-house IT have the capacity to operate a custom application, or does everything go to outsourcing?
Answers reveal the right path quickly. What does not work is deciding a priori "we want custom" or "we want a package" without doing the math.