The headquarters template breaks the local plant. That is a blunt statement, but it is exactly what I have seen at manufacturing sites across Turkey over the past two years. A European or North American parent company documents exactly how its enterprise resource planning (ERP) software works at home — which screens to use, which reports to pull, which approval steps to follow. This documentation arrives in Turkey as a package. A consultant shows up, the installation begins, training is delivered, and three months later the plant manager is back to the same Excel spreadsheets. The software runs. The plant does not.Where does the problem start? Not in the software itself, but in what the template cannot see. A headquarters template is built for the parent company’s home environment. In Germany or the United States, supplier payments settle in sixty days, the local currency does not swing week to week, tax classification follows a stable rulebook, and warehouse layouts are standardised. At a plant in Gebze or Bursa, the picture is different. The wounds from the 2001 economic crisis are still fresh. Suppliers cannot wait thirty days; some insist on payment at delivery. The Turkish lira fluctuates, so a single foreign-currency raw material purchase creates a new accounting entry every time the rate moves. None of that is in the headquarters template.I worked through exactly this situation at a metal components manufacturer near Gebze, with 312 employees. The European partner sent a standard ERP package — a reputable product, used widely. The installation team arrived, loaded the system, and completed the setup. The first week looked fine. By the second week, the purchasing team was in crisis. The supplier payment screen in the system offered only bank wire transfer as an option. But at that time, a significant share of smaller suppliers still expected promissory notes or post-dated cheques. Sitting down with the accountant to figure out how to record those payments manually took three full days. Three days spent on payment method alone, before any real production data had even been entered.The first lesson I take from that situation is straightforward: before any ERP software is installed, the local team must write down, step by step, how purchasing, payment, and accounting actually work today. Without that written list, nobody can see what the software will not handle. In my experience, plants that prepared this list encountered far fewer surprises during installation. Plants that skipped it spent months running a patched system alongside the new one.The second major problem is stock valuation and cost tracking. In Turkish manufacturing, production cost changes every month. Raw material prices follow the exchange rate. Energy costs shift. Labour costs are renegotiated each year through collective agreements. A headquarters ERP template typically assumes a fixed monthly cost baseline. With that assumption, calculating the true unit cost at a Turkish plant becomes nearly impossible. At an automotive supplier plant I worked with, the cost figure the system produced differed from the accountant’s manual calculation by roughly fifty-two percent every month on average. The software was not wrong — its assumptions simply did not match Turkish conditions. Reconfiguring the costing module to reflect local dynamics required additional time and budget that the headquarters had not included in the original project plan.The third issue is user habits, and I think this is the most consistently overlooked one. From warehouse staff to production supervisors, most employees at a Turkish manufacturing plant in this period have either never worked with a computer at all or have only used a basic accounting programme. An ERP system introduces the concept of modules for the first time: a finance module, a stock module, a purchasing module — each with its own screens and its own logic. One week of training cannot bridge that gap. It simply cannot. At a plant in Bursa where I oversaw an installation, it took employees four months to begin using the system consistently and correctly. During those four months, workers were learning the new programme at the same time as they continued filling in paper forms, because nobody trusted the new system yet. Running two parallel record-keeping systems simultaneously guarantees data errors, and cleaning up those errors afterwards costs more than the installation itself.Headquarters management almost always asks the wrong question. They ask: ‘Is the software installed?’ The right question is: ‘Is the plant actually operating through this software?’ The distance between those two questions is where most ERP project failures live. Installation can be complete while the plant continues to struggle with the same problems as before. To close that gap, the parent company needs to verify data accuracy on the ground — asking whether reports are arriving is not enough.For a plant manager or purchasing director reading this, three concrete steps make the difference. First, before any software decision is made, map every department’s current workflow on paper. Ask each team ‘how do we actually do this today?’ and write it down. That list will show you immediately where the template fits and where it does not. Second, negotiate the localisation requirements with headquarters in writing before the project starts. Do not discover at the implementation table that the system cannot handle a cheque payment or a foreign-currency cost adjustment. Resolve those gaps in advance. Third, split the user training into two phases. Run the first phase before go-live covering basic navigation. Run the second phase six weeks after the system is live, once employees have worked through real situations and have actual questions. The second session will be far more productive than if you tried to compress everything into day one.I return to that metal components plant near Gebze. We did get the system working — but eleven months later than the schedule the headquarters had signed off on. The parent company struggled to understand the delay. ‘We installed the same software at our Polish plant in three months,’ they said. Yes. Poland is a different situation. Turkey is a different situation. It took three on-site meetings and pages of supporting data to explain that. Their final conclusion was: ‘Next time, we will budget for localisation from the start.’ I hope they do. Because a headquarters ERP template, deployed without confronting local reality, is not a finished product. It is an unfinished draft — and the Turkish plant pays the cost of completing it.
This article was originally published in Turkish by Gökhan MERCANOĞLU on January 26, 2002. The English edition has been reviewed and edited by the author.