In the industrial parks around Long Thanh and Nhon Trach, "factory digitalisation" is a phrase heard constantly, usually attached to a long list of technologies. This article skips the list and answers the more practical question: if you run a plant and want to start next quarter, what do you do first?
The principle: start where you are losing money, not where the technology is most interesting
This is what separates projects that survive from projects that stall. Three questions locate the starting point:
What information do you routinely wait for? If a supervisor has to wait until the end of the shift to know how much the line produced, that is your starting point.
Which errors repeat every month? Wrong material issued, late deliveries because a bottleneck was not visible early, stock that disagrees with the books.
What data is being entered twice? Every figure copied by hand from paper to a spreadsheet and then into accounting software is an opportunity for error, and the clearest sign of a gap that can be automated.
Whatever shows up in all three answers is where to begin.
Step 1 — Map how information actually flows
Before buying anything, follow one order from receipt to delivery and write down who records what, where they record it, and who they pass it to. Do it on paper, walk the floor, ask the people actually doing the work.
The result almost always surprises management, because the real process differs from the one in the ISO manual. Those differences are exactly what needs addressing — and this map is the most valuable document you can bring to any conversation with a software vendor.
Step 2 — Standardise your codes before digitising
This is the most commonly skipped step, and skipping it is the most common reason later projects get stuck. If the same material carries three different names in three departments, software will only help you produce three wrong numbers faster.
Agree one shared coding scheme for materials, finished goods, customers and suppliers. This needs no software — only someone with the authority to decide, and a few weeks of careful work. It is the foundation for everything after it.
Step 3 — Digitise exactly one process, and run it for months
Pick one process from step 1 and finish it. For most plants that is work-order progress tracking or warehouse management.
Run it live for two to three months before discussing the next phase. That period does three things: it corrects where reality differs from the design, it lets operators get used to working on a screen, and it produces real numbers proving the next investment is worth making.
Three mistakes that sink these projects
Rolling out plant-wide in one go. These look best on paper and fail most often in practice, because nobody has time to change how they work when everything changes at once.
Designing the system in a meeting room. If the people writing the software have not stood at the line and watched the step happen, the software will be wrong in the details only the floor reveals — gloves that do not work on a touchscreen, no power outlet where the scanner needs to go, no one on the night shift who knows what to do when a terminal freezes.
Skipping training and an internal owner. Every plant needs at least one person inside who understands the system well enough to answer colleagues' questions. Without that person, the system gets worked around within weeks.
On cost
Steps 1 and 2 cost almost nothing in software, only internal time — and they determine most of the outcome. Step 3 should be quoted as its own line item, accepted independently, so you see a real result before committing a larger budget. What custom software actually costs breaks down how the number is built.
We survey on site before making any proposal, and that visit is free. If after walking your floor we think you should do step 2 first and come back in a few months, we will say so.