Why this matters. The Bitkom study positioning ERP as the backbone for AI and ecommerce data flows is clear: a shop without ERP truth sells stock that no longer exists. The architecture choice — plugin, middleware or native — decides whether growth carries or sync becomes debt.
When this view holds — and when it does not
It holds when:
-
inventory drifts between shop and marketplaces and oversells appear;
-
online list price and negotiated B2B price collide;
-
returns are reconciled manually to the ledger.
It holds less when:
-
the shop stays pure D2C single-channel, one warehouse;
-
no meaningful revenue runs via marketplaces;
-
tax logic stays domestic and EU-OSS-free.
Three architectural decisions you cannot defer
The methodology question comes before the product question: which of these decisions does today's stack carry — and from when does it stop?
1. Plugin, middleware or native ERP integration. A plugin typically carries to about 5,000 orders per month on a single channel. Once multi-channel, multi-warehouse or EU-OSS tax logic enter, the translation layer becomes its own discipline — the moment for middleware (iPaaS). When inventory truth or consolidation exceed an iPaaS, native ERP integration is the only defensible option.
2. Channel conflict and B2B price discipline. Online list price and negotiated sales price must not contradict each other. The ERP must derive the price condition from customer master data — otherwise margin leakage appears that only surfaces in reporting months later. The discipline lives in pricing logic, not the front end.
3. Finance integration. Datev or SAP-FI mapping from the shop data model is the most common hidden show-stopper. EU-OSS tax logic, returns posting and refund workflows must be derived cleanly — otherwise the cost surfaces only at audit.
A common mistake
A D2C merchant in Saarland with 12,000 orders per month sold across three marketplaces on a plugin stack. Inventory drifted; two per cent of orders were lost in sync. By the mid-2026 cutover, oversell damage compounded into six figures, plus 1.5 FTE tied up in manual reconciliation. The middleware move came fourteen months too late.
Across more than 1,200 selection projects the pattern is the same: the inflection point shows up in volume, channel count and oversell rate — not in a feature matrix.
What to do next
If you want to test whether your shop-to-ERP architecture can carry the next growth phase:
-
Inventory availability: measure oversell rate between shop and marketplaces for three months.
-
Price condition: document where online list and negotiated B2B price collide today.
-
Returns reconciliation: count manual finance postings per week.
We assess whether your shop-to-ERP architecture can carry the next growth phase — and name the inflection point against your own numbers.
FAQ
A plugin connector typically carries to about 5,000 orders per month on a single channel. Above that, sync errors and oversell risk grow faster than the plugin model can scale — the inflection point sits between multi-channel entry and EU-OSS obligation.
Plugin fits single-channel D2C with one warehouse and domestic tax logic. Middleware becomes useful once multi-channel, multi-warehouse or EU-OSS tax logic enter the picture. The switch is the methodology question, not the product choice.
The ledger tie-in is the most common hidden show-stopper. Datev or SAP-FI mapping must be derived cleanly from the shop data model — including EU-OSS tax logic, returns posting and refund workflow. Otherwise the cost only surfaces at audit time.
