A machine signal fired by a PLC and a purchase order entered by a planner start in completely different formats, on completely different timescales, and somehow need to end up describing the same production run. Mapping the data flow between SCADA, MES and ERP means tracing that signal's actual path — what protocol carries it, what format it changes into, what middleware or historian sits in between, and what finally lands in each system's database. Without that map, "integration" is just a word; with it, a team can diagnose exactly where a specific number got lost or delayed. You can trace your own plant's data flow with our team.
Smart Manufacturing Data Flow, 2026
Follow One Signal From the Machine to the Balance Sheet
iFactory, an on-premise AI platform for industrial data integration, maps the full path a data point travels from SCADA sensor through MES execution to ERP reporting, so every hop is visible and diagnosable.
PLC / Sensor
SCADA / Historian
MES
ERP
What it is
What Does Mapping SCADA-MES-ERP Data Flow Actually Involve?
Data flow mapping documents the literal path a piece of information travels: which protocol carries a machine signal off the PLC, what format converts it along the way, which system is the first to transform raw data into a business-meaningful record, and what finally reaches ERP for reporting.
This is different from architecture in general — architecture defines which systems should exchange what; data flow mapping traces how a specific piece of data actually gets there today, hop by hop, protocol by protocol.
The protocols
What Typically Carries Data at Each Hop
Machine to SCADA
OPC-UA, Modbus or proprietary PLC protocols carry raw signals and tags.
SCADA to historian
Time-series data logged continuously, often at sub-second resolution.
SCADA/historian to MES
Aggregated events and state changes, not raw tag streams, feed execution records.
MES to ERP
Batch or near-real-time API calls carry completions, scrap and consumption.
Where it stalls
Four Points Where Data Flow Commonly Breaks
| Break point | What happens | What to check |
|---|
| PLC to SCADA | Tag mapping drifts after a firmware or program update | Validate tag lists after any PLC change |
| SCADA to MES | Raw signals never get aggregated into meaningful events | Confirm event logic matches what MES expects |
| MES to ERP | Batch sync runs too infrequently for the business need | Match sync frequency to decision cadence |
| Cross-system IDs | Work order or batch numbers do not match across systems | Enforce one ID scheme across the whole flow |
How it works
Four Steps to Mapping Your Own Data Flow
1
Pick one data point that matters — completion quantity, for instance — and trace it from source to final destination.
2
Document every hop: protocol, format change, and any system that transforms or aggregates it along the way.
3
Note the latency at each hop — how long between the event happening and the data arriving at the next system.
4
Repeat for your highest-value data points, building a complete map rather than relying on a vendor's generic diagram.
Expert review
What a Controls Engineer Would Check First
Holds up
OPC-UA as the dominant interoperability standard between PLCs and higher-level systems is well established industry practice, not a vendor-specific choice.
Watch for
A data flow diagram that looks clean on a slide but was never validated against an actual live signal — tag mapping drifts silently after equipment changes.
Do next
Trace your most business-critical data point end to end first, since that is where a silent break costs the most.
FAQ
What Manufacturing Teams Ask About Data Flow Mapping
What protocol typically connects a PLC to SCADA?
OPC-UA is the most common modern interoperability standard, though Modbus and proprietary PLC protocols remain common on older equipment. Most plants run a mix depending on when each machine was installed.
Check your own protocol mix with our team.
Why does data flow mapping matter if the architecture is already defined?
Architecture defines what should happen; data flow mapping confirms what actually happens today, hop by hop. A plant can have a correct architecture on paper and still have a broken tag mapping or a stalled sync that only a live trace will reveal.
Ask how mapping differs from architecture design.
How do we know if a specific data point is arriving late?
Compare the timestamp of the originating event against the timestamp it was recorded in the downstream system; a consistent gap larger than the decision cadence requires signals a latency problem worth investigating.
Does a historian replace the need for direct SCADA-to-MES integration?
No — a historian is excellent for storing time-series data for analysis, but MES typically needs aggregated events rather than raw tag streams, so a separate integration layer is usually still required between the two.
Discuss your historian's role in the flow.
What is the biggest risk of not mapping the data flow explicitly?
Without an explicit map, a break anywhere in the chain is invisible until someone notices a downstream number looks wrong, and tracing it back through undocumented hops can take far longer than it should.
Get a documented map of your own flow.
