Shop Floor Data Collection: Connecting Machines and Operators to Your ERP
Shop floor data collection is the automated capture of production events - labor, machine cycles, counts, scrap, and downtime - and their posting into ERP transactions, replacing end-of-shift paper travelers and manual keying. In most discrete plants, manual reporting means WIP positions are 8-24 hours stale and 3-8 percent of transactions contain errors, which corrupts scheduling, inventory, and actual cost. Modern collection combines barcode and RFID scanning for operator transactions with direct machine connectivity via MTConnect and OPC UA for cycle and count data. This guide covers the technology stack, ERP integration points for Infor SyteLine, LN, and M3, and where AI agents eliminate the remaining manual work.
Why Paper Travelers Cost More Than They Appear To
The visible cost of manual data collection is clerical: an operator spending 20-30 minutes per shift writing and a clerk keying it, roughly $60,000-100,000 per year in a 50-operator plant. The invisible cost is larger. Stale WIP means planners expedite jobs that are actually complete and miss jobs that are stuck; a plant running 500 open jobs typically carries 10-15 percent phantom WIP on paper reporting. Keying errors post labor to wrong jobs, distorting actual costs and job profitability reports that quoting then relies on - a compounding error loop. Backflush-only plants avoid keying but lose scrap visibility and operation-level status entirely. When plants move to real-time collection, the measured effects cluster consistently: inventory record accuracy rises from the high 80s to 97-99 percent, and expedite-driven overtime drops 15-25 percent within two quarters.
The Collection Stack: Barcode, Machine Data, and IIoT
A complete stack has three tiers. Operator transactions - job start/stop, material issue, move, scrap - come from barcode or RFID scanning on rugged handhelds or fixed kiosks. Machine data - cycle counts, run/idle/fault state, feeds and speeds - comes from controller protocols. Environmental and process data - temperature, vibration, torque - comes from IIoT sensors when quality or maintenance use cases justify it.
- Barcode layer: Infor Factory Track for SyteLine/LN/M3, or RF-SMART; label every job packet, location, and employee badge with Code 128 or 2D
- Machine layer: MTConnect for CNC (Haas, Mazak, DMG Mori publish native agents), OPC UA for PLCs (Siemens S7, Allen-Bradley Logix), Fanuc FOCAS for legacy Fanuc controls
- Edge gateway: Kepware (PTC), Ignition by Inductive Automation, or HighByte to normalize protocols before they touch the ERP
- IIoT sensors: retrofit current transformers and vibration pucks ($150-500 per machine) give run/idle state on machines with no digital controller
ERP Integration Points That Actually Post Transactions
Collected events are worthless until they post as ERP transactions with validation. In SyteLine, labor and machine events post through job transaction IDOs (SLJobTrans) or Factory Track workflows, which enforce the same validation as manual entry - valid job, operation, employee, and quantity-remaining checks. In LN, hours accounting and production reporting sessions (ticst, tisfc) accept transactions through ION BODs or LN APIs. In M3, MI transactions (PMS060MI ReportOperation, MMS850MI for inventory moves) are the wire-in points.
- Post at event granularity, not shift summaries - operation start/stop timestamps enable true queue and run time analytics
- Validate at capture: reject a scan against a closed job at the scanner, not in a nightly error report nobody reads
- Map machine fault codes to a curated downtime reason tree (10-20 reasons), because 400 raw PLC codes produce unusable Pareto charts
- Keep an error queue with same-day resolution SLA; unresolved collection errors quietly rebuild the phantom WIP problem
Implementation Economics and a Realistic Rollout Plan
For a 40-machine discrete plant, expect $150,000-400,000 total: handhelds and labels ($30,000-60,000), machine connectivity hardware and edge software ($60,000-150,000 depending on controller ages), and integration services. Phase it: start with barcode labor and material transactions on one value stream (8-12 weeks to live), add machine-state monitoring on constraint work centers next, and defer full IIoT sensoring until specific quality or maintenance use cases demand it. The single best predictor of success is supervisor engagement - plants where supervisors run their morning meetings from the collected data sustain accuracy; plants where the data goes only to management see scan compliance decay below 80 percent within six months. Payback typically lands in 9-18 months from labor savings, expedite reduction, and inventory accuracy alone, before any OEE improvement is counted.
How Netray AI Agents Turn Collected Data into Action
Netray connects shop floors to Infor ERPs and then puts AI agents on top of the data stream. Our integration work covers Factory Track deployment, MTConnect and OPC UA gateway configuration, and validated transaction posting into SyteLine, LN, or M3. The agents are the differentiator: an exception agent watches the transaction stream and flags jobs with abnormal cycle times, missing operations, or scrap spikes within minutes, with a plain-language probable cause; a data-quality agent reconciles machine counts against operator-reported quantities and auto-corrects systematic scanner misses; a downtime agent classifies raw fault codes into your reason tree using historical patterns, keeping Pareto data clean without operator burden. Clients typically hold scan compliance above 97 percent and cut transaction error backlogs to near zero, with all processing available on-prem for ITAR and CMMC 2.0 environments.
Frequently Asked Questions
What is shop floor data collection in ERP?
Shop floor data collection is the capture of production events - labor start and stop, quantities complete, scrap, material issues, moves, and machine cycles - at the point they occur, and their automated posting as ERP transactions. It replaces paper travelers and end-of-shift keying with barcode scanning, kiosks, and direct machine connectivity, keeping ERP WIP, inventory, and actual cost data accurate within minutes rather than a day behind.
What is the difference between MTConnect and OPC UA?
MTConnect is a read-only, XML/HTTP protocol designed specifically for machine tools; most modern CNC builders (Haas, Mazak, Okuma, DMG Mori) ship native MTConnect agents, making it the fastest path to cycle and state data from CNCs. OPC UA is a broader industrial protocol supporting read/write, security policies, and complex data models, and is the standard for PLC-based equipment like presses, ovens, and assembly lines. Most plants use both, normalized through an edge gateway such as Kepware or Ignition.
How much does it cost to connect machines to an ERP system?
Budget roughly $2,000-8,000 per machine all-in. Modern CNCs with native MTConnect need only network drops and edge software licensing at the low end. Legacy machines need hardware retrofits - protocol adapters, or sensor kits at $150-500 for basic run/idle state - plus gateway configuration. For a 40-machine plant, total programs including barcode collection, edge software (Kepware or Ignition), and ERP integration services typically run $150,000-400,000, with payback in 9-18 months.
Key Takeaways
- 1Why Paper Travelers Cost More Than They Appear To: The visible cost of manual data collection is clerical: an operator spending 20-30 minutes per shift writing and a clerk keying it, roughly $60,000-100,000 per year in a 50-operator plant. The invisible cost is larger.
- 2The Collection Stack: Barcode, Machine Data, and IIoT: A complete stack has three tiers. Operator transactions - job start/stop, material issue, move, scrap - come from barcode or RFID scanning on rugged handhelds or fixed kiosks.
- 3ERP Integration Points That Actually Post Transactions: Collected events are worthless until they post as ERP transactions with validation. In SyteLine, labor and machine events post through job transaction IDOs (SLJobTrans) or Factory Track workflows, which enforce the same validation as manual entry - valid job, operation, employee, and quantity-remaining checks.
Put this into numbers
Free interactive tools for exactly this problem. No signup to use them.
Shop Floor Digitization Scorecard
A 10-question scorecard measuring how much of your shop floor still runs on paper - and which digitization gaps are costing you the most.
Free ToolManufacturing Downtime Cost Calculator
Combine lost contribution margin, idle labor, and absorbed overhead into a defensible monthly and annual downtime cost - with recovery effects modeled.
Free ToolCycle Time & Capacity Calculator
Translate cycle time, parallel stations, shift pattern, and realistic utilization into hourly, daily, and annual production capacity.
Terms used in this article
Have Netray assess your shop floor data collection gaps - we will map every machine and transaction point and hand you a phased connectivity plan.
Related Resources
MES-ERP Integration for Discrete Manufacturing
MES-ERP integration guide for discrete manufacturers: architecture patterns, ISA-95 mapping, SyteLine and LN connectors, costs, timelines, and AI-driven sync.
ERPPreventive Maintenance and ERP: Closing the Loop
Preventive maintenance ERP integration guide: CMMS-ERP data flows, meter-based PM triggers, MRO inventory, scheduling impact, and AI-driven predictive maintenance.
ERPSPC Integration with ERP Quality Modules
SPC integration with ERP quality modules: control charts, Cpk data flow, AS9100D compliance, SyteLine and LN quality setup, and AI-driven quality automation.