A typical multi-line plant runs on five systems that do not talk to each other. An MES records what was made. A CMMS records what was fixed. An OEE dashboard reports a ratio. A planner keeps the real schedule in a spreadsheet. Finance values output in a model nobody on the floor has seen. Five systems, five versions of the truth, and the question that decides whether the plant buys a new line, “where is capacity being lost and what is it worth”, takes a week to answer and is usually answered wrong.
A manufacturing performance platform (MES, OEE, CMMS & AI) is software that connects machine data, OEE and loss analysis, production scheduling, SKU-level output value and maintenance in one system, so a plant can find where capacity is lost, quantify what it is worth, decide what to do about it, and sustain the gain. In short, it turns hidden losses into recovered profit.
This article explains why the category exists, what it replaces and what it does not, the six stages a platform has to cover, who it is for, and how to evaluate one. It is also the hub for a twenty-article series on the losses, the planning decisions and the economics the platform exists to address.
Each of the legacy categories was designed before one of three things was true.
Machine-verified data became affordable per line. Industrial sensors, PLC connectivity and a camera over the line now cost a fraction of what plant-wide instrumentation cost a decade ago. A plant can measure one line for a few thousand dollars and know, rather than estimate, where its time goes. The OEE dashboard category was built when most of its data came from operators’ shift sheets, and much of it still does.
Large language models can turn loss data into prioritized actions. Reading stop records across stations, SKUs, shifts and work orders at a scale no engineer has time for, recognizing the patterns, and writing the action in plain language with the evidence attached: that was not possible when the MES and the CMMS categories were defined, and both were built to record rather than to recommend.
Capital became expensive enough that “recover before you build” is a board question. A new packing line costs seven figures and a year, and boards now ask whether the existing lines are producing what they could before approving one. Answering that question needs OEE, scheduling, SKU economics and maintenance in the same calculation, which none of the single-purpose categories can supply.
The category exists because those three conditions arrived together, and because the question they make answerable, how much capacity does the plant already own, is worth more than any of the five systems’ individual reports.
| Category | Built to do | What it cannot tell you | Where the platform overlaps |
|---|---|---|---|
| MES | Record production: orders, quantities, genealogy, compliance | Why output was lost, or what the loss cost | The platform includes the MES core most plants use day to day (live production tracking, stops, waste, changeovers) and adds the loss analysis and the money a classic MES was never built for |
| OEE / downtime software | Report availability, performance and quality as a ratio | What to do about it, in what order, worth how much | The platform’s Measure stage, with machine-verified data and station-level attribution |
| CMMS | Manage maintenance work: work orders, PMs, parts, technicians | What any of that work did to output | The platform’s Maintain stage, on the same asset model as OEE |
| APS / scheduler | Plan and sequence orders against capacity | What the lines actually do, unless someone feeds it | The platform’s Plan stage, fed from measured rates and changeovers |
| IIoT / data platform | Collect and store machine signals | What the signals mean for the plant’s economics | The platform’s Connect stage; the platform adds the loss model and the money on top |
| BI dashboards | Display data from any of the above | Anything the analyst did not already know to ask | Replaced by actions with owners rather than charts |
What it does not replace. A manufacturing performance platform does not replace an ERP, which remains the system of record for orders, materials and finance and the source of the schedule’s orders. It does not replace a compliance-focused MES where genealogy, electronic batch records or regulatory traceability are required; it runs alongside one. It does not replace a plant historian. For most multi-line plants it covers the MES functions they actually use, and it fills the gap between the systems, which is where the capacity decision has always lived and where no system has owned it. The finite-capacity article draws the boundary with APS in detail.
A platform has to cover six stages, and the point is that they form a loop rather than a pipeline. A dashboard tells you capacity was lost. The loop recovers it and keeps it.
Connect. Machine-verified data from PLC signals, industrial sensors and cameras over the line, not from operator logs. Every stop, count and state recorded at station level with a timestamp. Where a sensor cannot see what happened, the camera stores the footage and suggests the stop reason for an operator to confirm. Without this stage nothing downstream is trustworthy, which is why the series begins with why most OEE numbers are wrong.
Measure. OEE on a stated time basis, with changeovers in the denominator. The loss Pareto per shift, SKU, crew and station, with stops attributed to the machine that caused them rather than the one that showed them. This is where the series spends its second cluster: micro-stops, bad-actor machines, changeovers, starvation and blocking.
Plan. A finite-capacity schedule built on the measured rates and the measured changeover matrix, so Monday’s plan is one the lines can make, and replanned from live line state when they diverge. The third cluster: why schedules break, sequencing by changeover cost, planning at demonstrated rates, finite capacity without an APS.
Value. Every lost and recovered hour priced at the contribution margin of the SKU that was running, with production value, contribution margin and achievable margin kept apart. The plant enters the selling price and margin once per SKU; from then on OEE is reported in dollars as well as points. The fourth cluster: valuing output at SKU level, margin per line-hour, the capex-avoidance case.
Act. Insights that are actions: a specific loss, a value, an owner, a trigger and a measured result, sorted into recoverable capacity, data hygiene and systemic, and ranked by achievable margin. The AI proposes; a named person decides and executes; the system measures. The insights article and the governance article, both coming soon, cover what that means and what it must not mean.
Maintain. The fix becomes a work order against the same asset the OEE module measured, PM compliance is weighted by lost output, and the rate at which fixes hold is measured, so maintenance is expressed in the currency of output. From OEE diagnosis to CMMS cure is the series’ treatment.
Then back to Connect: the stop rate on the fixed asset is measured, the demonstrated rate rises, the plan tightens, the value is booked. The loop is the product. Any one stage alone is a category that already exists.
The word is overused, so the test should be concrete. An output is actionable when it names a specific loss at a specific station on a specific SKU, states what removing it is worth in the plant’s own margin, assigns an owner, sets a trigger or a date, and is re-measured afterwards. Anything missing one of those five is a finding, and a plant has enough findings. “Micro-stops on line 3 are up 12%” is a finding. “Adjust the labeler guide for SKU E on line 3: 14 stops an hour, $31,000 a year, owner the line engineer, re-measure Friday” is an action. The insights article, coming soon, develops the distinction.
It is for multi-line discrete and process plants in FMCG, food and beverage, packaging, consumer goods and components, running at or below about 70% OEE on an honest basis, with a capacity question in front of them: a new-line request, structural overtime, a plan that breaks by Tuesday, or a maintenance budget that cannot show its effect on output. Groups with several such plants gain a second benefit, which is comparable loss data and consolidated reporting across sites with the same categories.
It is not for a two-machine job shop, where the owner already knows where the time goes. It is not for a plant with a mature MES already running at 85% on an honest basis, which has a different problem. And it is not for a buyer whose question is still maintenance: a plant that wants work orders on a phone and a PM calendar should buy a CMMS and come back when the question changes, which the CMMS comparison, coming soon, addresses.
Eight questions, each with the answer that should disqualify a candidate.
Fabrico is a manufacturing performance platform (MES, OEE, CMMS & AI) built on the six stages above, with machine, sensor and camera connectivity, OEE and loss analysis, production scheduling and replanning, SKU-level output value, AI actionable insights and maintenance in one system on one asset model. Fabrico calls its method the Manufacturing Performance Engine: detect losses, measure them in money, turn them into actionable insights, and recover output with results verified against production data. It was built group-first, so a multi-site manufacturer compares lines and plants on the same loss categories, and for the shop floor, with iOS and Android apps, so the people who clear the stops and do the work use it at the line. It is hosted on AWS in the EU, is certified to ISO/IEC 27001, ISO/IEC 20000-1 and ISO 9001, and is an SAP Silver Partner listed on the SAP Store.
One FMCG group running it across its plants increased output by 14% (see the Ficosota customer case). The entry point in the United States is a fixed-fee, six-week pilot on one representative line, with the fee credited in full against a first-year subscription if the plant rolls out.
This article is the hub for twenty pieces on the losses, decisions and economics a manufacturing performance platform exists to address. Each is written for the person who owns the problem.
Capacity and capex (COO, VP Manufacturing)
Loss analysis and OEE (Plant Manager, CI Director)
Scheduling and replanning (Planning Manager, VP Operations)
Output value and economics (CFO, COO)
AI and the category (all roles)
What is a manufacturing performance platform? Software that brings MES, OEE, CMMS and AI together. It connects machine data, OEE and loss analysis, production scheduling, SKU-level output value and maintenance in one system, so a plant can find where capacity is lost, quantify what it is worth, decide what to do about it, and sustain the gain. It covers six stages, Connect, Measure, Plan, Value, Act and Maintain, as a loop.
How is a manufacturing performance platform different from an MES? A traditional MES records production: what was made, when, in what quantity, with genealogy and compliance where required. A manufacturing performance platform includes the MES functions most plants use day to day (live production tracking, stops, waste, changeovers) and adds why output was lost, what the loss cost, what to do about it and whether the fix worked. Where a compliance-focused MES is already in place, the two coexist.
How is it different from OEE software? OEE software reports a ratio. A manufacturing performance platform starts from machine-verified OEE and adds station-level loss attribution, scheduling on measured rates, SKU-level value, actions with owners, and maintenance on the same asset model, so the ratio becomes a dollar figure and a plan.
How is it different from a CMMS? A CMMS manages maintenance work and knows nothing about output. A manufacturing performance platform includes maintenance as one stage of six and expresses every maintenance decision in lost or recovered output. A plant whose question is still maintenance should buy a CMMS; a plant asking where capacity is lost needs the platform.
Do I need an MES first? No. A manufacturing performance platform covers the core MES functions itself, so most plants run it with an ERP above and PLCs or sensors below, without a separate MES. Where an MES exists for compliance or genealogy, it stays, and the platform takes production counts from it or from the machines directly.
Does it replace my ERP? No. The ERP remains the system of record for orders, materials and finance. The platform takes orders and due dates from it, returns planned start and finish, and leaves master data where it is.
What does a manufacturing performance platform cost? Platform pricing is typically per line and per site on an annual subscription, with a short fixed-fee pilot on one line as the entry point so the plant has a measured number before committing. The pilot fee is usually credited against the subscription on roll-out.
The category exists to answer one question: how much capacity does the plant already own, and what is it worth. The way to find out is not to buy a platform; it is to measure one line for six weeks and see the number.
The fixed-scope pilot does that: one representative line, an industrial sensor and hub installed with your team, machine-verified OEE on a stated basis, the loss Pareto with recoverable and structural losses separated, the recovered capacity valued using the plant’s own SKU economics, and the three highest-value interventions with owners. The fee is fixed and credited in full against a first-year subscription if you roll out. The pilot runs on Fabrico’s manufacturing performance platform (MES, OEE, CMMS & AI), which connects machine data, OEE and loss analysis, production scheduling, SKU-level output value and maintenance in one system, so the recovered capacity is measured, planned and sustained rather than reported once.
Request a demo or explore the platform.
Login
Blog
