The multi-site data problem
When a manufacturer opens a second plant, there is a fork in the road that has long-term consequences. Option one: deploy the same ERP at the new site and build an integration to consolidate reporting. Option two: run the new site on the same instance of the ERP as the first site. Most manufacturers take option one because it feels lower risk. In practice, it creates a permanently fragmented data infrastructure that becomes more expensive and less useful with every additional site.
A separate instance per site means separate item masters that drift apart over time, separate chart of accounts that require mapping for consolidation, separate vendor records that create duplicate supplier relationships, and consolidation reports that are always one sync cycle behind reality. The operational cost of maintaining this fragmentation compounds annually.
What one database actually enables
- Consolidated inventory view — real stock position across all sites, searchable in one place
- Inter-site transfers — move stock between plants as a native transaction, not an integration event
- Shared item and vendor master — one item code means one item everywhere, no mapping required
- Consolidated financial reporting — P&L and balance sheet across all sites in real time, not at month-end
- Central purchasing, local receipts — negotiate at group level, receive at site level
- Cross-site production planning — allocate orders to plants based on capacity and inventory across all sites
Site-level autonomy within a shared platform
A common objection to shared-instance ERP is loss of site autonomy. In practice, well-configured multi-company ERP gives each site its own accounts, its own warehouse structure, its own cost centres, and its own approval workflows, while sharing item masters, vendor masters, and reporting infrastructure. Site managers see their site. Regional managers see their region. Group management sees everything.
The key is a platform designed for multi-company from the ground up, not one where multi-company is an add-on. Company, branch, cost centre, and warehouse hierarchies should be native data structures, not custom fields.
Every site on its own system is not independence. It is isolation. Independence means each site can operate autonomously while contributing to a shared picture.
Scaling without per-site or per-user costs
Per-seat licensing creates a perverse incentive in multi-site environments: to control costs, access is restricted. Warehouse staff at the new site do not get logins. Plant managers use shared credentials. The data quality degrades because the people closest to the work cannot enter data directly. A platform with unlimited users across all sites eliminates this compromise. Every person who needs access has access, which means every transaction is recorded by the person who performed it.
Adding a new site should require a configuration exercise — define the company, the warehouses, the chart of accounts extension, the approval hierarchy — not a new licence negotiation, a new implementation project, and a new integration build.