Scan reporting is live but data is still wrong—put capture points where status actually changes
The most common mistake of the online code scanner is to think of it as "moving a paper ticket to a mobile phone". The real design problem is to collect data on which action, whether it is enough to collect it once, and when it must not be recorded.
What you'll take away
- Inaccurate reporting usually comes from the wrong capture point, not employee resistance; once the point is right, compliance issues tend to shrink.
- Each of the six candidate collection points has a clear purpose, and it is not necessary to collect all of them, but the step of acquisition determines the depth of traceability, OEE and dispatch feedback.
- Three design principles - no duplicate recording, no cross-job recording, no postmortem recording; no postmortem recording is the bottom line of traceability and the credibility of efficiency indicators
- Four typical anti-models: unified recording at the end of the shift, only reporting that the total amount is not defective, sticking the code to the equipment instead of the workpiece, abnormal verbal transmission
- Reporting workers must rewrite dispatch progress and equipment operation events, otherwise MES is just an electronic notepad
Start with the conclusion
The failure rate of workshop scan-and-report work projects is relatively high, but reviews often blame 'employees unwilling to scan'—this attribution is mostly wrong.
The real reason is usually:The collection point was chosen outside the correct action.Scanning the entire day's workload at the end of the shift is not data collection, it's backfilling; scanning codes is more troublesome than writing by hand, so employees naturally avoid it. Whether the data is accurate is primarily a design issue, and only secondly a management issue.
Reporting work in MES is not an isolated module: it needs to write backWork Assignment Progress, supplyOEE Output and Operating Time, supportProcess Quality and Traceability. If the wrong location is sampled, the next three links will break together. Before going online, answer three questions first—up to which layer to track, whether the plan needs real-time feedback, and where the numerator of OEE comes from—then draw the collection points, rather than sending the terminal first and then looking for its use. Barcodes/RFID are just carriers; without action binding, no matter how fast you scan, it is still an empty scan.
Controversial Scene: Why Is It Difficult
| Scene | superficial reason | A deeper layer |
|---|---|---|
| The planner does not trust the system schedule | "The workers don't sweep" | The timestamp is squeezed at the end of the class, essentially a make-up recording |
| The output looks good, but the quality doesn't match. | “There are separate statistics for delinquency” | Only reporting the total amount, the defects are still on the paper form |
| The review needs human-machine verification, and the system should respond to the team. | "The feature hasn't been enabled yet." | Started work without binding the human-machine interface, or stuck the code label on the wrong object |
| Scanning the code is slower than writing | Employee resistance | Re-enter multiple times, each system scan once |
| OEE numerator jumping around | Unstable collection | The completion definition is not aligned with the Ideal cadence |
Six candidate collection points
| Collection point | Collected Content | Main value | Is it necessary |
|---|---|---|---|
| Material requisition and feeding | Material batch, quantity, feeding station | Trace and back-check materials; complete actual performance | Essential for Traceability/Customer Complaint Scenarios |
| Start work | Work order, process, equipment, personnel, shift, start time | Dispatch feedback, working hours, OEE starting point | Recommended essentials |
| First article confirmation | First article inspection result | Prevent mass errors; quality gate | Essential for processes with high quality risks |
| Completion Report | Completed, Qualified, Defective, Defect Code | Output and quality master data; performance rate numerator | Essential |
| Process inspection | Spot Check/Inspection Results | Process Control and Traceability | According to the importance of the process |
| Warehousing / Transfer Sequence | Quantity, target location, or next step | Visual control of work-in-progress, inventory accuracy | Essential when coordinating with warehousing/next process |
Design trade-offs: Do not pursue complete data collection. When listing quality disputes, delivery delays, and efficiency analysis, determine the 'three questions that must be answered,' and then work backward to identify the data collection points; others can be added later. Each additional point increases the rhythm tax—before levying the tax, first confirm that someone is using this piece of data.
Three Design Principles
Do not record repeatedly. A single scan should simultaneously update progress, record work hours, trigger inspection tasks, and deduct or transfer work-in-progress—not record each on multiple pages. Repeated recording is the number one reason why 'scanning is slower than writing.'
No cross-position recruitment. Whoever does it records it. If the team leader records for the whole team, it is equivalent to shifting the data collection point backward, causing both timeliness and accuracy to drop. The team leader looks at the data but does not record the data.
Do not record afterwards. Tracing back to the bottom line of OEE. The results recorded afterward are clumped timestamps and estimated quantities. Supplementary recordings must be tagged and isolated in statistics; completely prohibiting supplementary recordings often forces people to take notes on paper or record after work, which is even more distorted.
Judgment criteria: randomly select a day to report work and observe the distribution of timestamps. If they are highly concentrated at the hour marks or the end of shifts, it indicates that data collection has failed.
How does the system connect (subsequent production reporting driven)
In comparison with JEDP-MES 'Barcode/RFID scanning for reporting work, collecting start and finish times, personnel, output, and process parameters':
| Downstream | How to drive report work |
|---|---|
| Dispatch work | Starting work occupies resources, completion releases them; pauses require reasons, for rescheduling and delivery risk notifications |
| Quality | Defective code enters the defect library; the first piece/IPQC not completed can intercept the transfer sequence |
| Trace back | Feeding batch Human-machine binding Timestamp, forming the skeleton of the batch number to the workstation |
| Equipment/OEE | The operating and shutdown states are cross-verified with reporting events; the production volume is based on the completion standard. |
| Andon | Create work orders at the point where material shortages, malfunctions, or quality abnormalities occur, and do not rely solely on verbally calling people. |
If reporting work does not trigger any action, what is collected is just an electronic record—this is the common outcome of 'scanning in as if not scanned in'.
The collection point determines the capability boundary
| Ability | Questions that can be answered | Required collection points |
|---|---|---|
| Batch traceability | Which batch of material was used for the finished product? | Material requisition and feeding |
| Process progress | Which step has been reached and when it will be completed | Start Work Finish Work |
| Workstation-level traceability | Who, which device, which shift | Start Work Human-Machine Binding |
| Plan Feedback | Is the dispatch carried out according to the window? | Start/Completion/Pause Timestamp |
| OEE Performance/Quality | Output and First-Pass Yield | Completed Quantity Defects; Status Collection |
From the batch level to the workstation level, each additional layer increases the collection density by one step.First decide which level the business needs, then acquire up to that level.。
Go-live acceptance: a stricter standard than 'can scan'
| Acceptance item | Pass the standard | Don't rush to expand the scope when it doesn't pass |
|---|---|---|
| Timestamp distribution | The start/completion of key processes is distributed over working periods, and the end-of-shift peak can be explained | First make supplementary entries, then discuss retroactive measures. |
| Inappropriate content on the same screen | Completion must fill in qualified/defective split, and the completeness rate of defect codes meets the standard | Output and quality will still be divided |
| Human-machine consistency | You can see people, equipment, and shifts on the random inspection work order | Workstation-level traceability is not available |
| Multipurpose | Progress in the same scan drive is being tracked, with no secondary data entry. | Employees will bypass |
| Consistent with dispatch | The dispatch status and the reporting status are consistent in spot checks | The plan feedback is still unreliable |
Pragmatic compromise for processes with extremely short cycles: start each batch at once, let the equipment automatically count output, and have manual intervention only at changeovers, exceptions, or sequence transition points. Strictly requiring 'scan each item once' leads to a slowdown of the production line, and the project is likely to be abandoned on-site within three months—after abandonment, it often leaves behind fake data that is even harder to clean.
Include the make-up recording strategy in the system: who can make up, within how long, whether it affects piece rate and OEE numerator, and whether it goes into quality audit sampling. Without a system, 'make-up recording is prohibited' in reality equals 'everyone making up secretly'.
Advancement Sequence / 90 Days
| Order | Do it first | Sign of effectiveness |
|---|---|---|
| 1 | Key process start Completion (including defects) | Merging the accounts of progress and output quality |
| 2 | Prohibit batch make-up recordings at the end of the class without marking; make-up recordings are counted separately | Timestamp recovery is trustworthy |
| 3 | Feeding batch collection (according to traceability requirements) | Customer complaint reverse check available |
| 4 | Human-Machine Binding and Abnormal Online Activation | Workstation-level traceability and Andon have data |
| 5 | Joint adjustment with dispatch gate and OEE definition | Reporting work changes from 'recording' to 'execution system' |
| Stage | Action | Completion mark |
|---|---|---|
| Month 1 | Sample Process Dual Collection Points Timestamp Weekly Report | End-of-class peak decline |
| Month 2 | Feeding batch and defect code enforcement | Customer complaint reverse check can be demonstrated |
| Month 3 | Joint Debugging Dispatch Occupancy and OEE Production Definition | The planner began to rely on the system's progress. |
Four Typical Anti-Patterns
- Unified make-up recording at the end of the class: Give up real-time performance, both auditing and analysis are unreliable
- Only report the total amount, not the defects: Output and quality split, OEE quality rate and traceability both fail
- The code is only attached to the equipment, not the workpiece/flow card: Multiple workpieces serial data
- Abnormally relies on verbal: The most valuable downtime and quality events are instead missing
Accompanying symptoms: The same output does not match in reporting, warehousing, and shipping — often due to missing routing/warehousing points or allowing goods to be warehoused without a work order. Another hidden anti-pattern: to boost the reported rate, the team leader scans for the entire group — the indicator meets the standard, but the data is dead.
When not to apply
- Single-piece small batch, primarily manual operations: According to the output of orders and key quality points, do not forcibly scan every process
- The workstation has no network: First resolve the terminal and offline retransmission, otherwise the collection point is essentially meaningless
- The process cycle time is extremely short: Change batch/automatic collection, hard scanning will slow down the rhythm
- No data users: Collection that is not monitored and does not trigger any actions should be stopped first or merged
- The process route still changes on a weekly basisFirst stabilize the route and process definitions, then solidify the data collection points
Layered governance of codes by objects: traceability of artifacts/transfer card code service products; OEE and maintenance of equipment code services; qualifications and piece-rate counting of personnel code services. The three can be implemented in stages, butYou cannot use a device code to impersonate a product code——Mixing usage is the root cause of data corruption; any traceability reports afterward cannot recover it.
What to do next
First, confirm the questions that need to be answered for traceability, plan feedback, and OEE, and then map the list of collection points. In the first week after going live, check whether the timestamps are still clustered at the end of the shift—if they are, don’t discuss traceability depth yet. In the second week, perform spot checks: whether defects appear on the same screen as completion, and whether the dispatch status is written back after reporting. If both weeks fail, expanding the scope will only replicate ineffective collection points. Whether the terminal is user-friendly is a matter of experience; clustered timestamps are a design failure, so adjust the points first.
- MES Manufacturing Solution: Scan Code for Work Reporting, Process Quality and Equipment Monitoring
- Trace from batch number to workstation
- APS Dispatch Closed Loop
- OEE three rates
- Workshop Reporting and Process Traceability Self-Test (No Data Required)
Want to apply this in your plant?Resources has matching self-assessments and templates—or book a live demo.
FAQ
Quick answers on this topic—still stuck?Contact
Is it better to have more collection points?
No. Each collection point has an operating time cost, and the benefits are retrospective accuracy, planned feedback, and real-time. First question: What are the most important questions to answer when it comes to quality disputes, delivery disputes, and efficiency analysis? The answer determines which points must be taken.
Is it allowed to record work after the fact?
Limited supplementary recording channels can be left, but they must be marked with supplementary recording marks, supplementary recording people, and supplementary recording reasons, and be statistically distinguished from normal acquisition. Completely prohibiting making up records often leads to note-taking and off-duty recording, which is more distorted.
Should the QR code be attached to the device or the workpiece?
Tracing equipment codes and product artifacts or turnover card codes, both of which are usually required. When only the equipment code is pasted, the multiple workpieces share a code, and the data will be strung.
Can we use barcode scanning to report work without MES?
Technically yes, but of limited value. The value of the report is to immediately drive the dispatch progress, process inspection, inventory deduction and traceability; the follow-up is still manual, and the collection is only one more record.
How to measure a process with a very short cycle time?
Avoid each manual sweep. You can automatically collect batch reports, equipment counts, or only collect key events at the start/batch switch/transition sequence point, and the details are automatically collected and supplemented.