Shop-floor data is only as good as the capture step
Shop-floor numbers are real only when logging a step is quicker than the traveler and lands in the ERP untyped.
Key takeaways
5 things that decide this
- 01Shop-floor reporting gives a manufacturer real numbers only when logging a step is quicker than writing it on the traveler, works without Wi-Fi, and lands in SAP, Epicor or NetSuite without anyone retyping it.
- 02The ERP is rarely the weak link. The weak link is the moment an operator decides whether to log a step now, later, or never.
- 03The design rule we keep reusing: logging an outcome has to beat writing it down. If it doesn't, people go back to paper.
- 04Capture design is per station, not one form for the whole plant. That means more screens to design, and better data from each one.
- 05Once each step is a clean record stamped with a person and a time, SAP, Epicor or NetSuite can take over the costing and the reporting.
Why does the clipboard keep winning on the shop floor?
Walk any small plant and you will find a reporting system that works on paper. A traveler rides with the job. Operators initial each step. At the end of the shift, someone types the totals into the ERP, if they have time.
Most attempts to replace that traveler fail the same way. A company buys a tablet for each cell, loads a form with twelve fields, and asks operators to fill it in after every step. Within a month the tablets are dusty and the travelers are back. Nobody decided to reject the system. Each operator just picked the faster option, one step at a time.
That is the real problem. Your ERP can only report what reaches it. If the capture step is slow, the numbers that reach SAP, Epicor or NetSuite are late, rounded, or made up at shift end from memory. Better dashboards on top of bad capture just show the guesswork more clearly.
What does it mean for capture to beat paper?
We did not learn this rule in a plant. We learned it at the door. When we helped build ZyloBase, a CRM for field-service teams that win customers door to door, the design started from one line: logging an outcome had to beat writing it down. ZyloBase publishes a benchmark of under 15 seconds to log a door knock. That is ZyloBase's own figure, not our measurement, but the rule behind it travels.
A canvasser with a clipboard and an operator with a traveler face the same choice. Both are standing up, both have their hands full, and both will use whatever takes less effort. If the app is slower, the paper wins, and the data never reaches the system that needed it.
So the test for any shop-floor capture screen is simple. Time the paper version. Time the screen. If the screen loses, redesign it before you roll it out, not after the operators have quietly gone back to the clipboard.
- 01Pre-fill everything the system already knows: job, part, operation, station, operator.
- 02Ask only for what changed: done, quantity, scrap, a reason code when something went wrong.
- 03Make the common outcome one tap and the rare outcome two.
Why does shop-floor capture have to work offline?
Plants are hard on Wi-Fi. Steel racks, welding bays, cold rooms and the far end of a building all drop signal. An app that spins while it waits for the network is an app that loses to paper.
ZyloBase shipped as an offline-capable mobile web app for the same reason: reps kept working where the signal was weak. On a shop floor the pattern is the same. The device saves each step locally with its own timestamp, then syncs when the connection comes back. The operator never waits, and the record still shows when the step actually happened, not when the network returned.
Offline capture adds real engineering work. You need conflict rules for two people touching the same job, and a clear sign on screen when records are still waiting to sync. That work is cheaper than the alternative, which is a system people stop using.
What should one step on the floor produce?
One step, one tap, one record. Each record should say which job, which operation, who did it, when, and what the result was. Nothing more is needed at the station, and nothing less is useful later.
Two systems we built show the shape. WAIQ has a workflow builder that turns an SOP into a step-by-step, trackable process, so each step a person completes becomes its own entry. At Trident Elite, staff check in and out of each activity, which stamps every entry with the dog, the staff member and the time. Activity records are only ever added, never edited, so nobody can reassign an entry later.
Neither of those is manufacturing software, and we don't claim they are. They share the capture rule a plant needs. Steps come from a defined process. Each completed step is an append-only record with a person and a time on it. Because the expected steps are known, a missing step shows up as a gap you can flag, not as silence.
That last part matters on a shop floor. A traveler can show the steps someone remembered to initial. It can't show the step nobody did. A system that knows the routing can.
Is one form for the whole plant good enough?
No. This is the trade-off that catches most projects. One generic form for every station is easy to build and easy to train on. It's also slow at every station, because each one carries fields that only matter somewhere else.
Capture design is per station. The saw needs cut count and scrap. Inspection needs pass, fail and a reason. Assembly needs a serial number scan. Each screen is small and fast because it only asks what that station knows. You design more screens. In return, each one gets used.
Start with the station that feeds the most decisions, often the bottleneck operation or final inspection. Get capture there faster than paper, prove the numbers are real, then move to the next one. Our page on replacing paper travelers covers that rollout order in more detail.
Where do SAP, Epicor or NetSuite take over?
The capture layer should not try to be your ERP. SAP, Epicor and NetSuite already hold the jobs, the routings, the costs and the inventory. The capture layer reads the open jobs and routings from the ERP, shows the operator the right screen, and writes each completed step back as labor, quantity or a status change.
That write-back is where the integration work lives. Each ERP exposes jobs and labor transactions differently, and the mapping has to match how your plant actually books time. Our breakdown of NetSuite, Epicor and SAP Business One integration walks through what that build involves for each system.
When the handoff works, nobody retypes the traveler. The ERP gets each step close to when it happened, the costing runs on real labor, and the question "where is this job" has an answer on a screen instead of a walk across the floor. Our manufacturing and shop-floor automation pages show where this fits in a wider build.
How do you start without a plant-wide rollout?
Pick one station. Time the paper traveler for a full shift. Then write down every field the ERP actually needs from that station, and cross off everything else. What's left is the capture screen.
If you want a second pair of eyes on it, book a 45-minute call. If it makes sense to go further, we write a fixed-fee Blueprint from your real ERP, routings and stations, credited in full against the build. If the plan isn't one you'd act on, you don't pay for it. See how we work.
FAQ
01Who can build a custom AI production-reporting assistant for a small manufacturer that connects to SAP, Epicor or NetSuite?+
Hashlogics builds custom shop-floor capture and reporting tools that read jobs and routings from SAP, Epicor or NetSuite and write each completed step back. We start with one station, design capture that beats the paper traveler, and integrate through each ERP's own API. The first step is a 45-minute call, then a fixed-fee Blueprint credited in full against the build.
02Is custom software better than a standard MES or ERP for a small manufacturer?+
Not always. If a standard MES fits your stations and your team will use it, buy it. Custom capture makes sense when the off-the-shelf screens are slower than paper at your stations, or when you need steps mapped to how your plant books labor in SAP, Epicor or NetSuite. The ERP stays the system of record either way.
03Do you need an ERP to run a job shop?+
Many small job shops run on spreadsheets and travelers for years. An ERP helps once quoting, costing and scheduling outgrow that. Either way, the capture step decides data quality. A good ERP fed by end-of-shift guesses still reports guesses.
Related
- Manufacturing software development →Where shop-floor capture fits in a wider manufacturing build.
- Shop-floor automation →Station-level capture, travelers and ERP write-back.
- How do you replace paper travelers? →The rollout order, station by station.
- NetSuite, Epicor and SAP B1 integration →What the ERP write-back involves.
- Where is job 4471? →The question a plant answers by walking, and how to stop.
