MES gets used as a catch-all term for “software that watches the shop floor,” which makes it nearly impossible to answer the question “do I need one” without knowing more, because the answer depends entirely on what specific job you actually need done. A machine monitoring system, a Manufacturing Execution System, and an ERP system solve three different problems, and a lot of shops buy the wrong one because the sales conversation blurred the line between them.
The three layers, and what each one owns
ERP (Enterprise Resource Planning) owns the business side: orders, purchasing, inventory, accounting, and typically high-level production planning, what needs to be made and by when, at the level of a work order or a sales order. It generally doesn’t know what happened at a specific machine during a specific minute, and it was never built to.
MES (Manufacturing Execution System) sits between ERP and the floor. It owns the execution layer: dispatching a specific job to a specific machine or operator, tracking that job’s genealogy and routing step by step, managing work instructions, and often material and labor transactions tied to that job. A full MES answers “where is this specific work order right now, and what exactly happened to it at every station.”
Machine monitoring owns the data layer underneath both: what a machine is actually doing right now, run or down, at what speed, making how many good and bad parts. It reads the equipment directly and turns that into OEE, downtime reasons, and reliability metrics. It generally doesn’t manage work orders, routings, or genealogy on its own.
Why the confusion is expensive
The cost of buying the wrong layer isn’t abstract. A shop that buys a full MES expecting it to replace machine-level monitoring often finds it has excellent job tracking and genealogy, and a shallow, manually-entered picture of actual machine downtime, because a lot of MES data entry still depends on someone keying in what happened rather than a system reading it off the equipment directly. A shop that buys monitoring expecting it to manage routings, genealogy, and work instructions finds excellent real-time machine data and nothing that actually dispatches or tracks a work order through its full lifecycle. Neither shop is wrong to want what it bought, it’s just not what it thought it was buying.
The checklist: do you need a full MES
A full MES is a real, often expensive, often lengthy implementation. It earns that cost when a shop needs several of the following at once:
- Regulatory genealogy, a legal or contractual requirement to trace a specific finished part back to the specific material lot, machine, and operator that made it, common in aerospace, medical device, and automotive supply chains
- Complex multi-step routings that need to be enforced, not merely recorded, where a part physically cannot move to the next station until the system confirms the prior step passed
- Paperless work instructions tied to a specific revision of a specific part, delivered to the operator at the station, with no room for an outdated printed sheet to get used by accident
- Material and labor transactions that need to post automatically as work happens, feeding cost accounting or inventory in near real time
- A genuine multi-system integration need, ERP, quality, and the floor all needing to stay synchronized automatically, not reconciled by a person at the end of the week
If a shop’s actual pain is “we don’t know how our machines are really running, downtime, speed, scrap,” none of those five items are the problem, and a full MES implementation, often a year-plus project with significant change management, is solving a problem the shop doesn’t have yet, at a cost calibrated for the problem it doesn’t have.
What monitoring actually gives you without the MES cost
For a shop whose real question is closer to “where is our capacity going” than “can we prove exact part genealogy for an audit,” machine monitoring answers that directly, and faster to stand up. Real OEE, a downtime Pareto ranked by dollars, reliability metrics like MTTR and MTBF, and a light job or product designation on the floor, what’s running right now, without the full weight of routings, genealogy, and enforced work instructions.
That light designation layer isn’t trying to be a full MES, and shouldn’t pretend to be one. Spall’s What’s-running flow, for example, lets an operator tag a station with a product and a start time in a few taps, the same work order ceremony question that shows up any time a shop weighs a full job record against a lighter one, which is enough to attribute production data to a name, but it isn’t dispatching a work order through an enforced routing or tracking material lot genealogy, and it doesn’t claim to. That’s a clear scope boundary, not a missing feature, the same way a torque wrench being an excellent torque wrench doesn’t make it a bad thing that it isn’t also a drill.
What to buy first if you’re not sure
Start with monitoring. It’s the layer that answers the most common first question a shop actually has, where is the capacity going, and it does it without months of process re-engineering. It also produces the exact evidence needed to justify a later MES investment: real downtime data, a real dollar figure attached to a real problem, rather than a guess that an MES will probably help.
The sequencing matters because an MES implementation without good underlying floor data tends to inherit the same blind spots a clipboard had, manually entered downtime, optimistic self-reported cycle times, just wrapped in a more expensive interface. Getting the machine-level truth first, then layering execution management on top of it once a specific, named need shows up, is a more reliable path than buying the biggest system available because it promises to eventually cover everything.
Quick recap
- ERP owns the business, MES owns work execution and genealogy, monitoring owns real-time machine truth, three different jobs
- Buying the wrong layer is expensive in a specific way: good data in one area, a blind spot exactly where you needed the truth
- A full MES earns its cost when several specific needs stack up: regulatory genealogy, enforced routings, paperless work instructions, automatic transactions, or real multi-system integration
- If the actual question is “where is our capacity going,” that’s a monitoring problem, not an MES problem
- Monitoring is the faster, cheaper layer to stand up, and it produces the evidence that justifies an MES investment later if you need one
- A lightweight product or job designation on a monitoring system is not a substitute for full MES genealogy, and shouldn’t be marketed as one
MES FAQ
What does MES stand for? Manufacturing Execution System. It manages the execution of work on the floor: dispatching jobs, tracking routings step by step, work instructions, and often material and labor transactions.
What is the difference between MES and machine monitoring? MES manages work, jobs, routings, genealogy, often through manual or semi-automated data entry. Machine monitoring reads equipment directly for real-time state, OEE, downtime reasons, and reliability data. Many shops need monitoring first and a full MES later, if at all.
Do I need an MES for a small job shop? Usually not right away. A full MES earns its cost when you have a genuine need for regulatory genealogy, enforced multi-step routings, paperless work instructions, or automatic material and labor transactions. If your real question is where capacity is going, machine monitoring answers that directly and faster.
Is MES the same as ERP? No. ERP owns the business layer, orders, purchasing, inventory, accounting, and often high-level planning. MES owns execution on the floor. Some vendors bundle both, but they answer different questions.