Cloud MES: How a Cloud-Based Manufacturing Execution System Works, 5 Compared (2026)

Cloud MES buyer checklist: 6 things to check

These six checks decide whether a cloud MES works on your floor. For the IT side in more depth, read our cloud MES buyer's guide on security, uptime and data ownership.

1. Offline buffering at the line

Internet connections drop. Ask what the system records at the line while the link is down, where those records wait, and how they sync when it comes back.

Test it in the pilot: disconnect the gateway for an hour during a run, then check that every stop and count from that hour appears with the right timestamps.

2. Data residency, hosting region and certificates

Ask which cloud provider hosts the data, in which region it is stored and processed, and where the backups are kept. In the EU, data residency is often the deciding question for IT and legal teams.

Then ask for the certificates themselves: ISO/IEC 27001 for information security, ISO/IEC 20000-1 for IT service management, and a SOC 2 Type II report where the vendor has one. Check who issued each one and what scope it covers.

3. Integration with the ERP and the PLCs

Ask what flows down from the ERP (orders, products, standard cycle times) and what flows back (counts, scrap, material use). Then ask how each machine connects: from the PLC over protocols such as OPC UA or MQTT, through an added sensor, or through operator entry.

Have the vendor show the method on your oldest machine. On SMT lines, ask whether it reads IPC-CFX and the Hermes standard.

4. Mobile on the floor

Operators use the system on a tablet or terminal at the line, often with gloves on and a machine waiting. Ask how many taps it takes to tag a stop reason, time it in the demo, and check what a technician sees when a machine needs repair.

5. How OEE is calculated

Two products can show different OEE for the same shift. Ask where the ideal cycle time comes from, whether planned stops are excluded, and whether short stops count as downtime or disappear into slow running.

Ask the vendor to work out one shift by hand from the raw stop events. Our guides to Overall Equipment Effectiveness and the OEE formula show what to look for.

6. A pilot plan with written criteria

Pick one bottleneck line, such as a filler, a press or a CNC cell, and agree before the start what a pass looks like: counts match a manual count over a shift, short stops appear as separate events, and the supervisor uses the data in the daily meeting.

Set an end date and a decision owner, and avoid the three mistakes that most often stall a pilot:

• Connecting every line at once instead of one bottleneck.
• Judging the product on demo data instead of your own machines.
• Leaving stop reasons undefined until go-live, so every shift invents its own.

5 cloud MES products at a glance

We picked products that describe themselves as cloud MES or cloud manufacturing software on their own websites. Product names are trademarks of their owners, and features change often, so confirm details with each vendor.

1. Fabrico

Disclosure: Fabrico is our product.

How it positions itself

Fabrico describes itself as a cloud-based MES and OEE platform for manufacturers, with maintenance management (CMMS) built in. The losses you measure on the line and the maintenance work that fixes them sit in one system.

What it covers

• Logs stops automatically from the PLC, an IoT sensor or a camera, and calculates OEE from that machine data.
• Operators tag stop causes on a tablet at the line.
• A technician opens the work order straight from the logged stop, in the built-in CMMS.
• Hosted on AWS in an EU region, with ISO/IEC 27001, ISO/IEC 20000-1 and ISO 9001 certification from DNV (see the Fabrico security page).
• SAP Silver Partner, which matters if your plant runs SAP.

Things to consider

• Production planning and scheduling are on Fabrico's roadmap, not in the product today. How will your ERP or current planning tool sit alongside it?
• Which connection method fits each of your machines: PLC, IoT sensor or camera? Put any sensors or cameras into the rollout plan.
• How does a stop become a work order, and how are the spare parts for that repair tracked? Ask to see it end to end.

Where Fabrico fits

Fabrico fits plants whose first problem is lost time on the line and slow follow-up from maintenance. For OEE with hardware, connecting the machines takes about a month and the full rollout 3 to 4 months.

Results include 20% less downtime at ABB, and factory-wide OEE up from 53.1% to 59.6% at Ficosota.

If maintenance still runs on paper, start with how a CMMS changes maintenance work. Read how Fabrico's MES and OEE platform works, or request a demo to see a logged stop turned into a work order.

2. Autodesk Fusion Operations

How it positions itself

Autodesk describes Fusion Operations as a cloud-based MES that digitizes the shop floor and connects planning and production in real time. Its site says the product brings MES and ERP capabilities together, for users from production managers and schedulers to QA teams.

What it covers

• Production tracking, with live views of orders, work in progress, machines, materials and labor.
• Scheduling and MRP, inventory management and operations maintenance.
• Quality and traceability from materials through operations to finished goods.
• Insights on cycle times, lead times, throughput and OEE.
• Connected tablets in place of paper forms, plus integrations.

Things to consider

• Which of your current ERP functions would move into it, and which would stay where they are?
• How does machine data reach it from your oldest machine, and what do operators still enter by hand?
• How does a machine stop become a maintenance task? Ask to see it in the demo.

3. Katana

How it positions itself

Katana describes itself as cloud inventory management software: one shared system for stock, orders and production, built for product businesses that sell across several channels and locations. Its manufacturing pages describe software for small and midsize businesses that have outgrown spreadsheets.

What it covers

• Real-time inventory across locations, with batch, lot and serial number tracking.
• Manufacturing orders, bills of materials and subassemblies, scheduled from demand and stock levels.
• A Shop Floor App with operator task lists and barcode scanning for materials.
• Integrations with e-commerce and accounting tools such as Shopify, QuickBooks and Xero, plus an open API.

Things to consider

• If OEE and machine stops are part of your brief, how would they be captured, and where would that data live?
• How do operators get their tasks in the Shop Floor App and report progress?
• If maintenance is in scope, is it handled in the product, in a separate CMMS, or through an integration?

4. Plex (Rockwell Automation)

How it positions itself

Rockwell Automation presents Plex as a smart manufacturing platform that unifies MES, ERP, quality and supply chain in one cloud system. It calls its MES an elastic MES: a cloud-based production management platform with a modular architecture, aimed at discrete, hybrid and regulated industries.

What it covers

• An MES that connects to an existing ERP, or MES, ERP and supply chain management in one platform.
• Operator control panels and error-proofed workflows on the plant floor.
• Quality procedures run from the control plan, inventory tracking with traceability, and finite scheduling.
• Production monitoring, asset performance management and machine data collection from PLCs.
• Mobile access, and deployment options from cloud-only to hybrid with a resilient edge.

Things to consider

• Are you buying the MES on its own or with the ERP? Settle it before the demo, because it changes the size of the project.
• Which deployment would each of your sites run, and what keeps running locally during an outage?
• Which modules does a first phase on one line need?

5. Tulip

How it positions itself

Tulip describes itself as a frontline operations platform and offers a composable, AI-native, app-based MES. Instead of fixed MES screens, teams assemble apps from building blocks and ready-made templates that share one data model.

What it covers

• A no-code app editor with AI-assisted authoring, and a library of ready-made apps, including an MES app suite.
• OEE tracking, with operators selecting downtime reasons at the terminal.
• Machine connections over OPC UA, MQTT and Modbus, with edge devices for older equipment.
• Work instructions, quality inspection, and genealogy and traceability apps, with role-based permissions and version tracking.
• GxP-ready features, a SOC 2 Type II report, and hosting on AWS and Microsoft Azure.

Things to consider

• Who on your team will build and maintain the apps, and who approves changes across sites?
• Which library apps match your first use case, and what would you still need to configure?
• Which edge hardware would your older machines need?

Frequently asked questions

What does cloud based MES actually change?

Mostly who runs the infrastructure and how fast you can start. The execution functions are much the same, but servers, backups, upgrades and availability move to the vendor. People reach the system from a browser at any site, and a pilot can begin on one line without buying servers.

What happens if the internet connection drops?

This is the question to press hardest. A credible cloud MES buffers data locally at the line and syncs when the link returns, so production keeps recording through an outage. Ask exactly what happens during a two hour loss of connectivity and what is lost, then test it in the pilot.

Is cloud MES suitable for regulated plants?

It can be, but the burden shifts to documentation: data residency, validation evidence and the vendor's change control for each release. In the EU, where data is stored and processed is often the deciding factor, so put the hosting region on your checklist next to the certificates.

How long does implementation take?

The software is rarely the long pole. Agreeing definitions, connecting equipment and changing how people work is where the time goes, which is why narrow first scopes finish and broad ones stall. See why MES implementations fail, and start with one line.

Do you need full MES, or something lighter?

If the actual problem is that nobody knows why the line stopped, a full execution system is a large answer to a smaller question. Dedicated OEE software or production monitoring software measures the losses first and shows whether you need scheduling and genealogy at all.

What should an aerospace or defense plant look for in a cloud MES?

Traceability comes first: genealogy by serial and lot, digital travelers, and quality records that stand up to an AS9100 audit. Then check hosting. If export-control rules such as ITAR apply to your data, confirm where it is stored and who can access it before you compare features.