The most dangerous moment in an ERP project is when the consultant wraps up the business analysis and tells the technical team ‘you can start building now.’ I paused at exactly that point in a built-in appliance factory in Manisa. The facility employed 378 people and its injection department ran an average of four color variants and two panel thicknesses per day. The planning manager showed me a hand-written notebook — inside it, which mold went to which machine, how much masterbatch of each color remained. That notebook was, truthfully, the operating system keeping the factory alive. The problem was not the absence of ERP. The problem was that ERP was about to be installed without anyone having understood the variant logic captured in that notebook.The most consistently misunderstood point in ERP projects for built-in appliance manufacturers is this: plastic raw material management cannot be solved with a standard inventory module. In injection molding, the concept of ‘product’ exists on at least three layers. The first layer is the physical item — a built-in oven front panel. The second is the bill of materials: which mold, which polymer granule, which additive ratio. The third is the operational parameter: mold heating time, injection pressure, cooling cycle duration. All three layers influence production planning decisions. Yet in standard ERP deployments, only the first layer enters the system. The other two remain in the heads of operators or inside spreadsheets. In the Manisa factory, the ERP consultant had built the bill of materials but had left mold-to-machine assignment logic and color-change scrap rates entirely out of scope. The result: ERP generates production orders, but the planning manager still reaches for the old notebook.The raw material picture grows more complex still. Polymer granule purchases happen by the ton; masterbatch — the color concentrate — is ordered by the kilogram and far more frequently. When these two purchasing rhythms are forced into a single inventory module, the system’s replenishment suggestions become genuinely unreliable. The reorder point calculated for granules projects three weeks of consumption forward; when the same formula runs for masterbatch, the scrap generated during color changeovers is ignored entirely. Based on measurements taken at the Manisa plant, color-change scrap accounted for roughly one third of total masterbatch consumption — yet standard ERP order calculations routed this loss to an ‘unexplained variance’ line item rather than to production cost. In budget reviews, that unexplained variance attracted hard questions; nobody could read from the system that the source was color-change waste. This is not a configuration detail. It is a structural gap between how ERP models inventory and how injection molding actually consumes materials.On the production planning side, variant management becomes the central challenge. In built-in appliances, combinations of color, surface texture, and dimension multiply the product tree rapidly. A single built-in oven model offered in 12 colors and 3 panel sizes produces 36 theoretical variants. Registering each variant as a separate item code floods the system with thousands of codes and makes ongoing maintenance unworkable. The correct approach is a configurable bill of materials structure: one master template, variable parameters. This capability was available in the majority of ERP systems deployed in Turkey by 2015 — but it was routinely skipped during implementation. Why? Because in the requirements workshop, the planning manager says ‘we have product variety,’ and the consultant translates this as ‘we will open multiple item codes.’ The gap between those two interpretations becomes the project’s most expensive mistake. Configurable BOM is not a luxury feature for large manufacturers; it is the minimum viable structure for any variant-driven production environment.Closing the distance between management expectations and what is operationally achievable is the consultant’s primary obligation in these projects. In Manisa, the general manager wanted two things from ERP: real-time visibility into raw material costs, and production orders generated automatically without requiring the planning manager’s manual approval. The second request required at least six months of process maturation. In the first month of go-live, 68 percent of production orders required manual correction. Meaningful automation only became possible after mold-to-machine assignment logic and scrap parameters were fully loaded into the system and users had built the habit of entering data consistently. The honest sentence to deliver to the general manager is: ‘You have purchased the system, but you have not yet purchased the process that feeds the system.’ Most consultants avoid saying this because it creates discomfort. But when it goes unsaid, the system is abandoned eighteen months later — and the notebook returns permanently to the desk.Three actionable steps apply to any similar project. First: begin requirements analysis with a factory floor visit, not a meeting room. Photograph the planning notebook, the whiteboard, the Excel file on the shift supervisor’s screen. The information the system does not yet hold is written there in plain sight. Second: model raw material inventory with two separate replenishment logics — minimum-maximum stock level for granules, and color-specific consumption forecast with a built-in scrap allowance for masterbatch. Do not attempt to force both into a single reorder formula. Third: do not go live with the production orders module before the configurable BOM structure is in place. Production orders issued without settled variant logic generate erroneous data, and erroneous data drives users away from the system. Winning back a user who has lost trust in ERP data is considerably harder than selecting the software in the first place.Eight months after go-live, the Manisa factory showed measurable improvement: material waste rates had declined, and procurement decisions for raw materials had shifted from the purchasing manager’s individual judgment to system-generated suggestions. Color-change scrap was no longer an unexplained variance — it was a traceable cost line, visible by mold. The planning manager still occasionally opened the old notebook; habit, not necessity. ERP in built-in appliance manufacturing does not require a miracle. It requires that plastic raw material logic and variant structure be modeled completely before the first production order is issued. The time and patience needed to do that properly should be built into the ERP budget from day one — not discovered as an overrun at month four.
This article was originally published in Turkish by Gökhan MERCANOĞLU on January 17, 2015. The English edition has been reviewed and edited by the author.