The First Step of Digital Transformation Is Not Buying Software

Last month I visited a hardware factory in Wenzhou with 50 million RMB annual output. The owner's first question: “Sir Leng, friends keep telling me to get an MES — which brand is best?”

I said: hold on, show me the shopfloor first.

What I saw: production schedules living in a three-person WeChat group, BOMs that don't match actual material usage, inventory records off by more than 30%. And the owner was still asking which system to buy.

My answer: deploying MES on this foundation is not digitization — it's paying to suffer.

1. Software Is an Amplifier, Not a Savior

Many owners treat digital transformation as “buying software” — the biggest misconception. Software is not a savior; it is an amplifier: the worse your management, the faster a system kills you. Chaotic processes just get amplified and run faster.

This is not scare talk. CAICT data: the success rate of manufacturing digital transformation in China is only 18.6%, and over 65% of failures stem from overreach and losing touch with business fundamentals.

Painful cases abound: a transmission-housing plant spent 4.2 million RMB on ERP+MES and rebuilt 28 approval nodes; on trial-run day the veteran technicians revolted — parameters they used to tune by feel in 5 minutes now required 12 data fields across 3 approval levels, 40 minutes of entry alone. Unplanned downtime multiplied. The owner hired back a retired master who walked around the machine twice and fixed it in 5 minutes without ever opening the system.

A multi-million system losing to a retired craftsman. The problem is not the system — it's that the fundamentals were never aligned.

2. Check Three Numbers Before Any System

I tell owners to check three numbers, a hundred times more important than reading software quotes:

First, where is the production schedule? In a system, in Excel, in WeChat, or in the workshop supervisor's head? Without a single source of truth for scheduling, system-generated plans are a joke. A Foshan lighting factory spent 220k on cloud MES but kept the habit of verbal order changes — order versions turned to mush in the system, the shop built from wrong drawings, and rework cost an extra 300k a year.

Second, is the BOM accurate? Large BOM-to-actual deviations make every MRP purchase requirement wrong — wrong buys, overbuys, or line-stoppage shortages. A Dongguan hardware factory deployed ERP without cleaning data; ledger-vs-actual gaps over 30% made purchasing, issuing and counting all chaotic, forcing a system pause to re-sort everything.

Third, how large is the inventory ledger gap? Records saying stock exists when it doesn't (and vice versa) make system data garbage-in-garbage-out. A Chengdu custom-furniture maker spent 150k on ERP+MES and ended up using only the inventory module — 80% of functions idle, because nobody maintained master data and material codes were chaotic; cost accounting actually got slower.

3. The Real First Step: Three Things

• 1. Standardize processes first. Turn WeChat agreements, verbal arrangements and rule-of-thumb practices into written processes: who does it, when, how, and to what standard. Without this, the system merely cements chaos.
• 2. Clean the data first. One code per item, BOM calibration, inventory counts. Inaccurate data makes bigger systems more dangerous.
• 3. Adapt the organization first. Define who enters, who maintains, who is assessed. Go-live is 30% of the work; people catching up is the other 70%.

One blunt closing remark: if your annual output is under 3 million RMB, products are simple, and Excel plus WeChat manages things fine — don't rush to buy a system. That's not digitization; that's donating money.

The first step of digital transformation is not opening a software procurement list — it's aligning those three management numbers. Software is the last step, not the first.

Reposted from the Toutiao account 「精益智造践行者」 (Lean & Smart Manufacturing Practitioner).

Original article: https://www.toutiao.com/article/7678289638678741538/