Digital transformation in manufacturing: first steps with ERP, RFID and maintenance

Monday morning, the production manager has three separate lists on their desk: planning’s work order table, the warehouse’s stock count book and the fault notes the maintenance team pulled from its WhatsApp group. None of them says the same thing as the others. A part is needed for a shipment; planning says it is in stock, the warehouse cannot find it, and maintenance has only just learned that the machine producing that part stopped this week. In factories, digital transformation usually comes up on mornings like this and immediately turns into a software list. Yet the real question is not which software to buy, but which data breaks at which point.

Digital transformation in manufacturing is the move from paper and scattered files to a systematic data flow, in which data from a factory’s processes such as stock, work orders, production, maintenance and shipping is recorded once at the source and seen by everyone involved from a shared flow.

How do you digitalise a factory? First find where the data breaks

Transformation work starts not from a technology list but from business impact, priority, dependencies and the reality of the data. Your first job is to turn symptoms like those in the morning scene above into data break points. The table below offers a starting point for symptoms common in manufacturing plants:

Digital transformation strategies for manufacturing: a step-by-step roadmap

The six-step sequence is built on the idea that every step relies on the data produced by the step before it. Doing it the other way round, buying AI or automation first, usually means a system that rests on incomplete data.

1. Process discovery: Make the current flow, actors, data, decisions and exceptions visible. The real process differs from the process people describe.
2. Bottleneck analysis: Identify waiting, rework, errors and dependence on individuals. This list is the main input that sets the order.
3. ERP backbone: Sales, purchasing, stock, production and finance should meet in the same data flow. If you have a working ERP, completing the missing processes without ripping it out is often the more sensible route.
4. Data at the source: Record counts, goods receipt, shipping and production movements with barcode or RFID where they first happen; reduce manual entry after the fact.
5. Maintenance and field processes: Bring fault records, planned maintenance and technician assignment into one flow, so the causes of production disruptions become visible.
6. Reporting and automation: Once reliable data exists, automate management views and rule-based tasks; AI opportunities come last.

Why is maintenance tracking one of the early steps?

Unplanned downtime can instantly make even the tidiest work order in the ERP meaningless. If fault records are not kept, nobody knows which machine stops how often; periodic maintenance dates depend on people’s memory, and the same fault is rediscovered on different shifts. That is why maintenance data is an input not only for the technical team but also for planning: if planning learns in time that a machine has stopped, it can move the work order to another line. For this flow, KurumsalPortal’s Maintenance & Fault Management module records the fault, classifies it, assigns a technician and updates the equipment history when the intervention is complete.

Which process should a mid-sized factory start its digital transformation with?

There is no single answer for every factory, but five criteria help when choosing the first process. Scoring candidate processes against these criteria takes personal preference out of the discussion:

• Does it repeat every day or every shift?
• Is the cost of an error visible (scrap, downtime, wrong shipments, recounts)?
• Does it have a single owner and a clear group of daily users?
• Does it feed other processes? Stock and work order data are also inputs for maintenance and shipment planning.
• Is it easy for users on the shop floor to accept? A record that adds no extra workload on the floor is adopted faster.

In many factories, stock and counting and work order tracking are the two candidates that meet these criteria best; but your own bottleneck analysis, not a rule, should set the order.

Fictional example: first steps in a metal parts factory

A fictional mid-sized metal parts factory finds three problems in its bottleneck analysis: count differences, work orders that cannot be followed between stations, and faults that stay in WhatsApp. The team focuses on the counting routine first, because stock data affects both the production plan and purchasing. To speed up counting at carton and pallet level, they test bulk reading with RFID in a pilot area. Next, they split work orders in the ERP into station-based steps and consider matching production movements with the same tag. Finally, they bring fault records into a single flow. The management report comes only after these three flows have settled, because without reliable data a report only looks nice.

Five common mistakes in factory transformation

• Choosing the software first and trying to fit the process to it afterwards; the software should support the process, not force the process to fit it.
• Trying to transform the whole factory at once. A pilot on a single line or warehouse area teaches lessons cheaply.
• Involving users on the shop floor too late. The person who attaches the tag, does the count and enters the record is the real owner of the process.
• Not naming a data owner. Who corrects a stock record when it turns out wrong should be written down from the start.
• Not measuring the starting point. If count differences, downtime and scrap are not recorded today, there is no way to show improvement later.

When are Excel, paper and WhatsApp no longer enough to run a factory?

In a small workshop with one shift and a few products, Excel and a WhatsApp group do the job, and they are easy to use. Things start to break down when the number of shifts and the product range grow. The same data is kept in different files and people argue about which one is current. The status of a work order changes depending on who is asked. Fault notes get lost in a chat history and no maintenance history can be produced. When management asks for a report, merging the data takes days, and nobody fully trusts the merged data either.

ERP, RFID and production tracking: how does YazSis approach this sequence?

The sequence starts with digital transformation work: process discovery, bottleneck analysis, an integration and automation roadmap, and data and reporting architecture. The principle is simple: clarify the process before the technology, and the need before the investment. On the backbone side, YazSis ERP brings sales, purchasing, stock and warehouse, production and finance together in the same data flow, according to the business’s real way of working; on the production side it covers bills of materials, recipes, work orders, subcontracting, scrap, yield and cost. Completing missing processes without ripping out an existing, working ERP is also part of the approach.

For data at the source, RFID and IoT systems can come into play: with an RFID tunnel and handheld terminals, tagged products in cartons, on hangers and on pallets are read in bulk, and the production tracking system matches work order, station, operation and time information with RFID movements. How RFID reads will work on site is determined by measuring conditions on location, such as product type and the effect of metal or liquids. On the employee side, KurumsalPortal handles work such as requests, tasks, maintenance and faults; it runs on the customer’s own server and employees use it from their phone browser. If a data flow with the ERP or another system is needed, the scope is agreed with you during the needs analysis.

To work out together which data flows break in your factory and in what order to tackle them, see the process discovery and roadmap steps on the digital transformation page.