Picture an accounting clerk at a textile company pulling the first inventory report from a newly installed ERP (enterprise resource planning) system. The numbers on the screen do not match what she saw in the warehouse that morning. The same yarn entry appears three times, and one of the records shows a zero balance. The owner walks over and asks: ‘What exactly is this programme doing for us?’ The problem is not the software. The problem is the data that was fed into it. An ERP system processes whatever it receives. Put in wrong data and you get wrong reports. Put in clean data and you get clean reports. This sounds obvious, yet many small and mid-sized companies only discover this truth after the project is finished and the consultants have gone home.
An ERP programme brings together a company’s sales, purchasing, inventory, accounting and production functions into one shared database. Before such a system, each department kept its own records. Accounting worked from its own ledger, the warehouse used a separate notebook, and the sales team tracked orders on its own sheets of paper. ERP eliminates that fragmentation. But the integration only works when every department enters the correct information into the system on time and in full. If one link in the chain is broken, the whole system produces unreliable results.
One of the most critical steps in any ERP project is entering accurate opening balances — the financial figures that reflect where the company stands on the day the system goes live. Cash on hand, bank balances, outstanding receivables and payables must all be entered correctly at that moment. Most companies using a proper accounting programme for the first time assume their existing records are clean enough to transfer directly. However, ledgers maintained by hand over many years tend to carry small errors that accumulate over time. A customer account card (the record that tracks what a client owes or is owed) might still show an invoice as unpaid even though the debt was settled long ago. If those errors are carried into the new system, the programme starts its life on a faulty foundation. Reviewing and reconciling old records before entering opening balances is not optional work — it is the foundation of the entire project.
Inventory counting matters just as much. The programme can only know what is sitting in the warehouse if someone counts the stock and enters those figures correctly. Companies that skip a proper count or rush through it quickly run into phantom stock problems. A product appears in the system but is nowhere to be found on the shelf. Or the reverse: goods are stacked in the warehouse but the system has no record of them at all. When this happens, purchasing decisions, production planning and cost calculations all start producing meaningless results. A thorough, accurate stock count is a precondition for a healthy system launch, and repeating that count on a regular schedule keeps the data reliable over time.
Customer and supplier account cards, along with product cards, are two more places where data quality makes or breaks the project. Every customer, every supplier and every stock item needs to be entered consistently and without duplication. If the same customer appears under two slightly different spellings, the system creates two separate accounts and splits the balance between them, making reconciliation a nightmare. If a product card uses the wrong unit of measurement — say, boxes instead of individual pieces — stock quantities turn into figures that nobody can interpret. Fixing these errors is always possible, but fixing them after the system is already running takes far more time and erodes the trust of the people using it daily. Entering these cards carefully feels like slow, repetitive work, yet the care taken at this stage determines whether the whole project succeeds or struggles.
Data quality does not look after itself once the system is running. The first month everything might be entered correctly. Then a staff member leaves, a replacement joins and brings different habits, and a few entry steps start getting skipped. A warehouse worker dispatches goods but forgets to create the dispatch note (the document recording that the goods have left the building) in the system before the truck pulls away. These missed steps create gaps and inconsistencies in reports. Software support teams responding to complaints about strange report figures routinely ask the same first question: ‘Were all the entries actually made?’ In the majority of cases, the answer reveals a process issue rather than a technical fault. The programme can do a great deal, but it can only work with what it has been given.
Any small business owner considering an ERP installation or in the early stages of one should ask three questions before going live: Have our opening balances been verified and are they ready to enter? When was our last full inventory count, and can we stand behind those figures? Are our customer, supplier and product cards clean and consistent? If the honest answer to any of these is no, the right move is either to delay the go-live date or to set aside the time and the people needed to get those foundations right. A system that starts with poor data does not get better on its own — the errors compound week by week until cleaning them up feels harder than starting from scratch. In the end, it is not the software that determines the value of an ERP report. It is the quality of the data behind it.
This article was originally written in Turkish by Gökhan MERCANOĞLU on July 3, 2000 and has been automatically translated into English and other languages using machine translation.