Excel vs Enterprise MES vs Voltrus: Which Layer Fits Your Factory
Between a paper traveler and a fully digital factory there are three layers, and most mid-market plants live on one of them. The first layer is the spreadsheet: production plans in Excel, coordination over WhatsApp, one person who knows where everything is. The second is the enterprise MES program: a multi-year implementation led by a system integrator, with a name like Siemens or Rockwell on the purchase order. The third layer is the one in between: a small cloud MES that starts on one line instead of twelve.
This article will not declare a winner. A feature shootout between these three is meaningless because they are not priced, timed, or scoped for the same factory. What matters is fit: how fast each layer pays back, what breaks when your volume doubles, what an auditor puts you through, and what it costs to leave. If you are weighing MES against ERP money systems instead, start with MES vs ERP; this article assumes the floor is the problem.
The Comparison
Seven dimensions decide this choice in practice. None of them appear on a vendor's homepage, and all of them show up in month six.
| Dimension | Excel + WhatsApp | Enterprise MES Program | Voltrus MES |
|---|---|---|---|
| Time to first value | Days. But the value decays as the sheet drifts from the floor | 6 to 24 months to go-live; value arrives after the project ends | 2 to 4 weeks: one line, live machine data |
| Up-front cost shape | Near zero cash, high hidden labor in keying and reconciling | Six to seven figures, mostly integrator fees paid before any benefit | Pilot-first subscription on one line; grows only when you add lines |
| What breaks at 2x volume | Version chaos: two schedules exist and both are wrong somewhere | Nothing breaks. You already paid for the headroom | Add stations to the line config; the data model does not change |
| Audit response | Days of file archaeology across inboxes and personal drives | Strong, if e-records were in scope and configured; expensive if not | Audit exports by lot and work order, generated when asked |
| Traceability depth | Whatever one diligent person keyed in, when they keyed it | Full genealogy, if the integrator built the lot model you needed | Lot genealogy captured at station terminals as work happens |
| Who maintains it | One person, no backup, no version history | Integrator during the project, then your team inherits the configuration | Vendor-managed cloud; configuration, not code, on your side |
| Exit cost | Trivial. That cuts both ways: nothing accumulates value | Contract terms plus a re-implementation to change course | Export your records and go; no multi-year lock-in |
Read the table by column, not by cell. The Excel column is a system that works until it suddenly does not. The enterprise column is a system that works once an enormous project finishes. The Voltrus column is an attempt to keep the first column's speed and the second column's records without either's failure mode.
Stay on Excel If...
Sometimes the spreadsheet is the right answer, and pretending otherwise sells factories software they do not need. Stay on Excel if all of this is true.
Your product mix is stable and your routings rarely change, so the schedule is a template you copy and nudge each week rather than a puzzle you solve. Nobody audits you: no customer second-party audits, no BPOM or ISO pressure, no recall exposure that would force you to trace material lots under time. One person genuinely owns the spreadsheet, keeps it current, and would tell you before going on leave. And margins are fat enough that a 4% scrap rate or two lost hours per shift is an acceptable cost of doing business.
If that describes your plant, the honest advice is to invest the software budget in maintenance and tooling instead. When the scheduling question specifically starts to hurt, the real cost of spreadsheet scheduling breaks the numbers down; until then, Excel plus discipline beats any MES nobody uses.
You Outgrew the Spreadsheet If...
The failure is rarely dramatic. It shows up as four everyday symptoms.
Two schedule versions exist: the planner's master file and the version the floor actually works from, and nobody can say which one is current. Audit preparation takes days instead of an afternoon, because every answer requires pulling files from inboxes and reconstructing what happened. A customer asks which of their orders used material lot X and the answer takes a week of walking the floor, cross-checking logbooks by hand. Month-end is a scramble to reconcile the sheet against what was actually produced, and the reconciliation keeps finding surprises.
One of these is a to-do. Two are a trend. Three or more mean the spreadsheet has become a system nobody questions and nobody trusts, which is the worst of both. That is the point to add a layer, not to reorganize the sheet again.
Why Enterprise MES Projects Fail in Mid-Market Factories
The software is not the problem. Siemens-class and Rockwell-class platforms run some of the most advanced plants in the world. The failure pattern in a 50 to 500 person factory is structural, and it repeats for four reasons.
Deal size. The integrator's business model needs large projects to be profitable. A mid-market deployment is big enough to be complex and small enough to get the B-team: senior people sell and architect the project, junior people configure it, and the factory cannot tell the difference until month nine.
Time to value. Six to twenty-four months of implementation means the costs are certain and weekly while the benefits are projected and distant. The plant manager who sponsored the project may have moved on before go-live. The CFO who approved it has by then seen only invoices, and the second phase gets cut.
Integrator dependency. During the project, every configuration change is a purchase order. After go-live, the factory does not own the system in any practical sense: nobody internal dares touch the configuration, so the MES calcifies exactly where it was on the day the integrator left, while the factory keeps changing underneath it.
Feature weight without a spreadsheet story. Enterprise suites are built to replace paper plus SAP-class complexity in plants that already have both. They arrive with a thousand functions and no answer to the actual daily pain, which in a mid-market plant is usually "replace the scheduling spreadsheet on line 3 and stop losing material lots." If the first month of a MES does not remove a spreadsheet somebody maintains every day, the floor concludes the software is for head office, and adoption dies politely.
What Migrating Off Excel Actually Involves
The honest version, for a pilot that works: pick one line and one pain. Not the whole plant, and not a wish list. The pain is usually scheduling, downtime that nobody attributes, or traceability that takes a week. Everything else waits until the pilot pays.
Then comes the part nobody markets: work order data cleanup. The spreadsheet has product codes with three spellings, routings that reflect a machine you scrapped in 2022, and standard times nobody remembers setting. Before any import, someone who knows the floor names every product, every routing step, and every work center properly. For a single line this is typically a day or two of hard thinking, and it is permanent value: the factory has never had this list written down.
Next, station terminals. A tablet or small PC at each station takes the operator's confirmations: order started, order done, good count, scrap with a reason. Machine data comes from the machines where possible: Voltrus ingests machine states over keyed SCADA connections, so a PLC or a counter with a digital output becomes a live data source without touching the controls program. Where no signal exists, keyed entry at the terminal covers it; SCADA vs MES explains where the line sits between the two.
From first workshop to first useful output, a disciplined one-line pilot takes two to four weeks. The spreadsheet does not die on day one. It runs in parallel for the first weeks while trust is built, then it shrinks: reporting moves first, scheduling next, and eventually the master file becomes a view nobody maintains instead of a system nobody questions. That is the migration working, not failing.
Frequently Asked Questions
Can Voltrus import our Excel BOMs and schedules?
Yes. Bills of material and work orders import from the spreadsheets you already run, and the cleanup step described above is mostly about agreeing what each column means before it becomes a record. Expect to fix product codes, units, and standard times during import. Expect to keep planning in Excel for a short overlap: the import gives Voltrus the truth, and planning moves over once the floor trusts the new schedule view.
We already budgeted for an enterprise MES. Should we still pilot small?
Yes, and the pilot protects the big program rather than competing with it. If your plant genuinely needs enterprise scope, a four-week pilot on one line tells you what your real data looks like: which machines can emit signals, where the routings lie, what the operators will and will not key in. Walking into a multi-year integrator project with that knowledge changes the requirements you sign. Walking in without it is how a six-figure project discovers the standard times are fiction in month nine.
Is one line representative of the whole plant?
Not fully, and pretending otherwise would be dishonest. A pilot proves the data model, the adoption pattern, and the reporting on the line it covers. What it also does, at a fraction of program cost, is surface the plant-wide questions: which stations have machines that can talk, how scrap actually gets counted, who owns the schedule. Those answers make the expansion decision, in either direction, on evidence instead of a vendor slide.
Start the Pilot on One Line
Voltrus MES is in an early-access pilot cohort: one line, keyed SCADA ingest, lot genealogy, audit exports, live in two to four weeks. Bring the spreadsheet you have. We will tell you honestly whether you need us yet.
See Voltrus MES