How did the digitized asset management system land? Data acquisition and coding system
RFID Knowledge

How did the digitized asset management system land? Data acquisition and coding system

PublishedShanghai Zejie Smart Technology Co., Ltd.

Many companies get stuck at the first step of digital asset management: how to collect data and how to assign codes. The wrong capture method leads to costly rework. If codes are too broad or too detailed, the system soon turns into a register nobody maintains. Two factors guide the decision: asset volume and frequency of change. Use those to decide how much to automate and how detailed the codes should be. For a few hundred assets and a few dozen changes a month, manual entry and simple category codes suffice. For more than 10,000 assets with frequent additions and transfers, use barcodes or RFID and structured codes. Otherwise, high maintenance costs may lead staff to abandon the system within three months. Shanghai Zejie Technology Co., Ltd. has repeatedly validated this approach in its projects.

Data acquisition: manual entry vs automated collection

Data can be sent into the system in both directions, but the human capacity behind them and the cost of follow-up maintenance vary widely.

Real application of manual entry

Manual entry is not necessarily outdated. It's appropriate for up to 500 assets with annual changes of no more than 20%. A small service company with 300–400 items of office equipment and furniture and a few dozen purchases and disposals per year can have administrative staff maintain a spreadsheet. Regular counts can keep physical discrepancies within 2%.

Above 1,000 assets, the hidden costs of manual entry grow quickly. A warehouse services provider first used spreadsheets to manage nearly 3,000 rack and pallet sets. Entering additions, transfers, and disposals took 30% of two warehouse employees' time each month. Barcode scanning reduced entry time per asset from four minutes to 15 seconds, saving nearly 180,000 yuan in annual labor (2023 project review data).

Real input from automated collection

Automation of collection is not just a two-dimensional code. The complete scheme consists of three blocks: identification vehicle (barcode label, RFID electronic tag, BeiDou or GPS module), harvest terminal (handheld PDF, fixed-reader, sensor gateway), data feed link (Wi-Fi, 4G, industrial bus).

Investment varies widely by application. One equipment rental company tagged 8,000 units and installed four fixed read points. The upfront hardware cost was about RMB 350,000, twice that of a barcode-only setup. Counting and locating equipment became more than five times as efficient, and the investment paid back within two years.

The most cost-effective option is a two-dimensional tag plus a general mobile phone scanning, with little hardware input but requires workers to have each active sweep, with low efficiency caps and suitable for a limited budget and an inefficient start-up phase of asset movement.

How?

Coding system: One size vs classification code

The coding determines whether assets can be located, measured and aggregated quickly in the system.

Encoded Two Layers

A workable coding system has two layers. First, a classification code identifies the category and subcategory: for example, 03 for machinery, 0301 for pumps and valves, and 030101 for centrifugal pumps. Second, a sequence or serial number uniquely identifies each item in the category. It typically uses four to six digits and increments under company rules.

How do you get the level of detail?

The classification is too thick to cause statistical confusion – the forklift and trailers are classified as “transport equipment” and the finance and maintenance services cannot do a fine analysis. The classification rules expand the code list to dozens of pages, so that new subcategories will change the rules and the cost of maintenance will increase dramatically.

In our experience, three or four classification levels work best. A manufacturer with annual output worth CNY 500 million typically has 8-12 main asset categories and 40-60 second-level categories. A three-page document is enough to explain the coding rules.

Coding requirements and financial, physical reconciliation

The most problematic is the coding of three sets of codes -- one set of financials, one set of systems, one set of systems and one final ZZZQQZ reconciliation -- that nobody can tell for which asset. This is particularly common when the enterprise history system is replaced, the old code cannot be directly disabled, usually requiring two-track operation over six months to switch.

Book cost vs hidden cost

Digital accounts cannot be counted only for software and hardware.

The books will clear the money.

Software procurement or self-study costs, hardware (labelling, terminal, network) costs, implementation consulting fees, which usually account for more than 70% of the project budget. A medium-sized digitization project (assets 5000 to 10000) with an overall budget between 300,000 and 800,000 is a normal area.

Implicit costs easily omitted

The more expensive are the ones:

What do you think of the return cycle?

Returns come mainly from three areas: less stocktake labor, fewer lost assets, and greater redeployment of idle assets. One project we tracked had an initial investment of about CNY 500,000 and covered more than 6,000 equipment assets. After 18 months, annual stocktaking had fallen from 45 days to 12, asset loss had dropped from 3.2% to 0.8%, and internal transfers had put about CNY 900,000 in idle assets back to use. Overall payback was around 14 months.

Apply scene and advance rhythm

Logic of phased advancement

The digitization of assets is not appropriate to be fully staffed at first sight, and it is recommended that three instalments be organized according to the density of the asset ' s value:

  1. First: high unit cost, high-risk core production equipment covering more than half of the total value of assets, first keeping a high value component under control
  2. Second instalment: support equipment and office assets, large but low unit cost, with classification codes plus simplified process processing
  3. Third batch: Low-value, consumer and tool types, coded as required, with batch management without having to do one size for each object

Key cross-sectoral nodes

The three essential roles that the project can run: operations, which provides asset management rules and inventory alignment; finance, which is responsible for coding and mapping of accounts; or IT, or information, which is responsible for system landings and interfaces. Without either of the three roles, the project will be stuck at the data alignment level.

The pits that are common when you land.

Pipe One: The code rules work and we have to follow.

The system is being updated, the code is being recoded once, all downstream statements, historical inventory records, depreciation data are being re-reconciled at a much higher cost than two weeks more than the one-time design. One enterprise changed its asset classification rules three times in two years, and Treasury is always going to go back to depreciation data for almost a year.

Pipe II: Data collection and process disconnection

Two-dimensional coded but uncoated, which is not a sticker; the system is set up, but nobody's recorded, and it's no different from manual. The drive-through habits require that sweeps be embedded into routine operations -- when they are in the lead, when they are in maintenance, when they are back, when they are back, when they correspond to a change in status in the system, so that the sweep itself has business value.

Pipe Three: No master data manager

Encoding, asset classification, field standards, these master data (i.e., basic data quoted in a corporate context) need to be maintained for long periods of time. Most enterprises fail because they do not have a master data administrator appointed, code rules are added and modified in their use at will, and three years later, the system piles up a large amount of dirty data from history, and nobody moves.

Three specific recommendations

The first step in digitizing assets is not to buy software, but to secure the two foundations of data acquisition and coding. The foundation is solid, the back count, depreciation, transfer, and the end-of-life data are available; the more complex the system is, the less well-established it is. Most projects fail because the software was not wrong, but the data governance for the last three months was not solid.