Production Tracking with Reader Devices: Automatic Consumption and Shipment in ERP

When reader devices go live on a factory floor, the first feeling is one of control. An operator scans, the system records, the entry is done. Then the first week passes, and data inconsistencies begin to surface: incomplete consumption postings, inventory mismatches, shipment discrepancies. A mid-sized machine manufacturer in Kayseri with 312 employees experienced exactly this at the end of 2015. Significant investment had gone into barcode infrastructure. But the ERP workflows were not mature enough to interpret the data the readers were producing. The result was a second data layer sitting alongside the old manual process — not replacing it, but contradicting it. The thesis of this article is direct: purchasing a reader device does not automate production tracking. In a manufacturing environment without process maturity, that device produces only a digital version of the same underlying chaos.The core function expected from reader devices in production tracking is the reflection of material consumption — linked to a production order — and finished or semi-finished goods movements into ERP without human intervention. The mechanism enabling this is commonly called ‘backflush’ or retroactive consumption: when an operation is confirmed, the bill-of-materials components are automatically deducted from inventory. The reader device is the trigger source. The operator scans the product or operation, ERP recognizes the completion signal, and the consumption posting runs. For this to work reliably, the preconditions are strict and non-negotiable: bills of materials must be accurate and current, routing steps must be defined, raw material batch tracking must be running consistently, and the physical location of reader points must match the actual workflow. If any one of these conditions is absent, every scan feeds incorrect data into both current inventory and, very likely, customer shipment records.Returning to the Kayseri case. The factory manager’s objective was to eliminate the manual data entry burden carried by three operators at line stations — roughly two hours of hand-keyed input per shift. The theoretical calculation was sound: near-real-time consumption postings, reduced clerical effort, faster order visibility. Six weeks after go-live, a review comparing inventory variance to the prior period showed the opposite of what was expected. The variance had not narrowed — it had widened. The root cause was in the bills of materials: roughly 52 percent of the active BOMs had not been updated to reflect the current production configuration. The backflush mechanism consumed against old quantities. Operators noticed the discrepancies and quietly reverted to paper forms. The reader infrastructure was technically functioning. The business process beneath it was not ready.The lesson from this case goes beyond ‘fix your BOMs first.’ The structural problem is more persistent: in reader device projects, technical integration is typically completed successfully by the ERP consultant, while process ownership remains undefined. Who is responsible for keeping BOMs current? Who assigns batch codes on incoming raw materials? Which department intervenes when a batch mismatch is detected? These ambiguities are survivable in a fully manual environment — they surface only when they slow someone down. But when data is being processed in near-real time, every ambiguity becomes a posting error. This is why the questions a manufacturer must answer before a reader project begins are not technical but operational: Are production orders being opened and closed consistently? Is batch code assignment during goods receipt complete and accurate? Is there a defined process for BOM maintenance? If the answer to any of these is ‘not reliably,’ the hardware investment should wait. Stabilizing those processes first is not a delay — it is the investment that makes the reader project worth its cost.The shipment side carries its own layer of complexity. When reader infrastructure is extended to the dispatch point for automated shipment verification, ERP must match the sales order against the warehouse goods issue. The most common failure mode here is a mismatch between the packaging unit defined in ERP and the physical unit being scanned on the dock. In machine manufacturing this appears frequently: a production order closes in ‘piece’ units while the shipment leaves in ‘pallet’ units, and if ERP cannot resolve that conversion cleanly, the barcode scan reports incorrect order fulfillment rates. In the same Kayseri facility, this issue did not surface through a customer complaint. It emerged during a period-end stock count, when physical inventory matched ERP stock balance at only 63 percent. The remaining 37-percent gap had multiple contributing causes, but shipment unit mismatches accounted for a disproportionate share of the divergence. The reader had recorded every scan correctly. The unit-of-measure logic had not been configured to match reality.How should a reader project be structured to avoid these outcomes? The field-tested framework has four steps. First, before any hardware is installed, map every ‘trigger point’ on the production line to the specific ERP transaction it should generate. This ‘trigger map’ is a document — not a verbal agreement — and it must be reviewed and signed off by both the production manager and the ERP project lead before go-live. Second, activate backflush only for production orders whose BOM accuracy can be confirmed above a defined threshold. For orders that fall below that bar, require manual confirmation during the transition period. This hybrid approach is less elegant, but it protects data integrity while the underlying process is brought up to standard. Third, run a ‘dry operation’ period of two to three weeks before reader points go live: operators scan, the system records, but ERP does not post actual transactions — only simulated output is monitored. Discrepancies identified during this window are corrected before they become real inventory errors. Fourth, assign a named owner. Without someone monitoring backflush-sourced variance daily, clearing the error queue, and approving BOM updates, the process cannot self-correct.A reader device is a tool for simplifying production tracking within ERP. It is not a system that matures your processes on its own behalf. That distinction sounds minor, but confusing the two in a live manufacturing project generates measurable cost. In 2016, many mid-sized Turkish manufacturers are directing investment toward mobile readers and barcode infrastructure — and that is the right direction. The problem is that most project proposals reaching a management board lead with hardware budget, not with a process maturity assessment. The practical recommendation for decision-makers is this: before approving a reader project, ask your ERP consultant a single question: ‘How many of our BOMs will pass a validation review before backflush goes live?’ A team that can answer that question with a plan can handle the others. A team that cannot is not yet ready to install the hardware.

This article was originally published in Turkish by Gökhan MERCANOĞLU on January 23, 2016. The English edition has been reviewed and edited by the author.


ERP governance is not merely a technical choice; it reflects how the organization makes decisions. When process, data, and ownership are unclear, investment creates speed in the short term and complexity in the long term. Real value begins when technology is connected to a business outcome.


Gökhan Mercanoğlu
ERP ve Kurumsal Yazılım