Related Case Studies
Manufacturing
Implementing IoT Digital Twins for a Manufacturing Plant
Manufacturing
Manufacturing visibility was the one thing this group could not buy its way out of. Its plants ran across several European countries, and nobody could say where a given order stood without phoning three departments and waiting on the answer. The ERP held the purchase order, the floor system knew what was running, and the warehouse system logged what had shipped. None of them agreed. Late orders surfaced at the Monday meeting, days after anyone could act on them, and planners kept private spreadsheets because they did not trust the reports they were handed.
For this group, manufacturing visibility meant something specific: knowing where every order, part, and shipment stood without calling three people to find out. The data already existed. It just lived in separate systems that did not talk to each other. The ERP held purchase orders, the floor system tracked what was running, and the warehouse system knew what had shipped. None of them shared a single view.
We started by connecting those systems into one pipeline so a single record followed a part from supplier to delivery. That is the part most teams underestimate. Manufacturing visibility is not a dashboard you buy off a shelf. It is the work of agreeing on what a "late order" means across plants, then making every system report it the same way.
Once the data lined up, we layered exception alerting on top. Instead of staff scanning reports for problems, the system flagged the orders that were slipping and routed each one to whoever could act on it. Supplier delays that used to surface at the weekly meeting now showed up the same day.
The last piece was trust. People act on a number only when they believe it. We ran the new figures next to the old manual counts for a full quarter, until the floor managers stopped double-checking by hand. Manufacturing visibility is reached when the report becomes the number people plan from, not the one they quietly ignore.
The change is easiest to read side by side. Here is how the same four moments played out for this group before the systems were connected, and once manufacturing visibility was in place.
| Moment | Without connected visibility | With manufacturing visibility |
|---|---|---|
| Order status | Call three departments to find where an order actually stands | One shared record tracks every order, part, and shipment as it moves |
| Exceptions | Delays surface at the weekly meeting, often too late to fix | Slipping orders get flagged the same day and routed to whoever can act |
| Reconciliation | Staff rekey figures between ERP, floor, and warehouse systems by hand | Every system reports against one agreed definition, so counts already match |
| Decisions | Managers double-check numbers before they trust them | The report becomes the number people plan from, not the one they quietly ignore |
Getting to the right-hand column is what freed up the hours of manual reporting and caught supplier delays early, the changes behind the €343K saved on this engagement.
Related guides from our knowledge hub:
It is the ability to see the real status of orders, inventory, and shipments across every system and site at once, from one shared record instead of separate spreadsheets and reports. Real manufacturing visibility means a part can be traced from supplier to delivery without phoning three departments to piece the story together.
For a multi-plant group, plan for several months to about a year. Connecting the systems is the quick part. The slower work is agreeing on shared definitions and getting teams to trust a single set of numbers. This engagement ran 13 months from first audit to full rollout.
Off-the-shelf platforms handle standard reporting well, but most manufacturers still need custom integration work. The value comes from how your specific ERP, floor, and warehouse systems connect, and no vendor knows that layout without building to it.
The gains come from fewer surprises: supplier delays caught early, less manual reconciliation, and faster decisions on the floor. In this case the automation built on top of the new data saved €343K and removed hours of weekly reporting.
Every rollout on this page followed roughly the same order. If you are starting from disconnected systems, this is the sequence that worked for us, and it shows where most of the time actually goes.
The bill for manufacturing visibility is mostly people time, not software. Almost nothing here was a licence purchase. The spend went into the audit work, the arguments about definitions, and the parallel running that earns trust. That catches most teams out. They budget for a platform, then find the platform was the cheap part. Below is roughly how the 13 months of this engagement broke down and what each stretch of work paid back. The order matters as much as the effort. Skipping the definitions work to reach a dashboard sooner is the most common way these projects stall, because a shared view nobody agrees on is just a new place to argue.
| Workstream | Share of the 13 months | What it returns |
|---|---|---|
| Mapping the systems and auditing the data | About 2 months | A written list of every place an order, part, or shipment is recorded, and which copy each team actually trusts today |
| Agreeing shared definitions | About 3 months | One meaning for "late", "in stock", and "shipped" across every plant, which is what makes a shared number defensible |
| Building the connected record | About 3 months | One record that follows a part from supplier to delivery instead of being rekeyed by hand at every handoff |
| Exception alerting | About 2 months | Slipping orders routed to a named person the same day, rather than raised at the weekly meeting |
| Parallel running and adoption | About 3 months | A full quarter of new figures running beside the old manual counts, until floor managers stopped double-checking by hand |
Payback did not arrive in one moment. The hours of manual reporting went first, once the connected record removed the rekeying. Supplier delays came next, caught early enough to reroute rather than absorb. The €343K saved on this engagement came from the automation built on top of the new data, and that automation was only possible because the data underneath it had finally stopped contradicting itself. Visibility on its own saves nothing. It makes the things that do save money buildable.
Thirteen months in, almost nothing that went wrong here was technical. The connectors held. The failures were slower and quieter than that, and the same four keep turning up on the visibility rollouts we have looked at since. Each one has a tell you can spot early.
If floor managers keep their own counts a month after go-live, the numbers were connected before they were agreed. Rebuilding the pipeline will not touch that. Go back to the definitions work, publish one figure at a time, and retire the specific spreadsheet each figure replaces. Trust comes back one number at a time.
ERP says the order shipped Tuesday, the warehouse system says Thursday, and the argument lands in the weekly meeting instead of with a person. Name one owning system per field and record it beside the field: order dates from ERP, stock counts from the warehouse, delivery confirmation from the carrier feed. Once the owner is settled in advance, a contradiction turns straight into a question about why the two records drifted.
Exception alerting is the quickest part of a visibility build and the easiest to overbuild. Within weeks a channel filling with amber warnings gets muted, and the genuine ones are muted with it. We cut the alert set back before expanding it. Every alert routes to a named person with one action attached, and anything nobody acted on for a month gets switched off. Roughly half the original set never survived that test.
Internal systems get connected, supplier updates still arrive as email and PDF attachments, and the delays you most need warning about are the ones you still hear about last. Chasing structured feeds from every supplier stalls for years. Start with the handful who account for most of the late deliveries, usually 5 to 10 of them, get a real feed from those, and leave the tail on manual updates until the case for more is obvious.
Tell us what you are working on. We will tell you honestly if we can help.
let's connect