When a gypsum manufacturer launches an ERP project, most plant managers come to the table with two requests: fix the inventory count and get invoices out on time. Neither expectation is wrong, but together they address only about thirty-six percent of the real operational problem. The core issue is not invoice processing — it is consistently capturing the kiln’s quality parameters, fuel consumption rates and grinding band efficiency in a single, trustworthy data layer. Without that layer, an ERP system is not a production management information system; it is a back-office accounting tool wearing a manufacturing label. Projects that miss this distinction routinely deliver expensive screen replacements rather than operational change.Turkey’s gypsum production is heavily concentrated in Afyonkarahisar province, home to deposits that supply a large share of the national market. A typical mid-sized producer in the region employs between two and four hundred people, serves both domestic construction distributors and export channels, and runs continuous-process kilns that demand close monitoring. During advisory work at one such facility — 312 employees, two kiln lines — I found production shift records still maintained on paper forms. Daily burn volumes, energy consumption figures and waste ratios took two to three days to reach the finance team. Management learned the previous week’s production cost on Wednesday of the following week. Installing ERP did not close this gap, because the data collection points were never connected to the system. The reporting screen changed; the operational delay did not.The underlying structure matters here. Gypsum manufacturing generates process data on two distinct layers. The first is the financial-logistics layer: purchase orders, shipments, invoices, stock movements. Standard ERP modules handle this layer reliably. The second is the operational-process layer: kiln temperature curves, calcination duration, raw stone moisture content, band throughput rates. This layer lives at the production line, captured either by sensors or by manually completed operator forms. A genuine MIS derives its value from bridging these two layers — when operational data and financial data share a platform, unit cost becomes visible in near real time. Projects that omit this bridge deliver accounting reports to the general manager and leave the plant manager with his own notebook.The most consistent damage from skipping the data infrastructure step is loss of system credibility. At the Afyonkarahisar facility, the production manager put it plainly: ‘The system shows 400 tonnes in stock. I walk the warehouse and count 340. Which one do I trust?’ That question does not indicate a failed project — it indicates that data-entry discipline was never established. The variance traced back to waste records entered two days late, broken-bag losses tracked on a separate paper form, and incomplete delivery notes processed after shipment. The fix was not a new software module; it was a process map defining which user enters which data at which physical point. No ERP produces reliable stock figures without that map. It is worth adding a caveat here directly: even when roles are mapped and responsibilities assigned, without consistent management oversight the discipline typically unravels within four to six weeks in shift-based environments. The map is necessary but not sufficient.On the production planning side, ERP’s most tangible contribution is linking procurement to the production schedule. Raw stone inventory, fuel supply and packaging materials all need to be planned against forward production commitments. When this coordination runs on spreadsheets or verbal agreements, purchasing and operations managers sometimes make independent decisions. One week raw stone arrives in excess; the following week the line stops because packaging bags ran out. Once ERP’s production module is live and correctly parameterized — with bill of materials, machine capacity and shift structure entered — demand plans can feed procurement calendars automatically. In Turkish gypsum plants, parameterization is almost always the longest phase of the project, because the ‘recipe’ concept is frequently stored in the production foreman’s experience rather than any written document. Extracting that knowledge and encoding it into the system is a process design task, not a software task.On the integration side, two connection points are operationally significant as of 2014: e-Invoice and e-Ledger compliance with the Revenue Administration (GİB), and bank statement reconciliation. Gypsum producers selling in bulk to large construction contractors are already subject to e-Invoice obligations. An ERP system that handles GİB integration cleanly reduces document management overhead on high-volume tender accounts considerably. Bank reconciliation, however, still runs on periodic file imports at most installations — automated real-time matching is not yet standard practice, and presenting it to a client as a default feature sets expectations the current integration landscape cannot meet. Mobile ERP access is technically available, but 3G coverage on the plant floor in a provincial industrial zone limits practical tablet use far more than the software brochure suggests. Acknowledging this to a client before go-live prevents the expensive surprise of discovering it six months in.Before committing to an ERP investment in gypsum manufacturing, three questions deserve direct answers. First: who currently collects production data, at which physical point, in which format, and how long does it take to reach a decision-maker? Second: does a written bill of materials exist for each product line, or does unit cost estimation depend on the production manager’s experience? Third: what is the measurable success criterion at month six — if nothing specific changes by then, will the project be considered worthwhile? Projects that begin without clear answers to all three tend to repair the invoicing workflow while leaving production management untouched. The plant manager at the Afyonkarahisar facility said something that stays with me: ‘The system is running, but it is not telling me what I need to know.’ A consultant who hears that sentence has one job — not to replace the system, but to teach it what to say. That is process design work, and it should be scoped and budgeted before the first server is ordered.
This article was originally published in Turkish by Gökhan MERCANOĞLU on January 2, 2014. The English edition has been reviewed and edited by the author.