Published August 4, 2026 · 9 min read
Start with one machine. Scale plant-wide. Same software, same data model, same dashboards at every step — no rip-and-replace when the pilot succeeds. This is the pitch behind PulseMQ’s micro/mini MES tier, and it exists because most small manufacturers who try MES fail for the same reason: the starter and the enterprise product are two different products.
Traditional MES buyers hit the same wall twice.
The first wall is the entry price. Enterprise platforms want $50,000 to $500,000 up front, and small manufacturers reasonably refuse. So a “small business MES” category emerged — simpler tools, per-machine pricing, often SaaS. These work for the pilot.
The second wall shows up when the pilot works. You want to add ten more machines, then twenty. Now you need shift scheduling, batch traceability, ISA-88 recipe control, work orders that flow from ERP. The starter tool cannot do those things. You have to migrate to an enterprise platform. That migration is another $50,000 to $500,000, and it starts by throwing away every dashboard, every alarm setup, and every workflow the pilot team just spent a year learning.
The lesson most small manufacturers take away is that MES is not for them. The real lesson is that most starter MES tools are dead ends by design.
Fully scalable means the code you deploy on one machine is the code you deploy on eight hundred. Not a lightweight version, not a preview, not a compatibility layer — the same binary, the same data model, the same UI.
A few practical implications follow from that:
The micro tier is meant for one line, one cell, or one pilot. Everything runs on a single edge appliance sized for a small cabinet: a fanless industrial PC with 16 GB of RAM. No cloud dependency, no plant IT approval — you plug it into the machine network, point it at the PLCs, and you have live dashboards inside the day.
What you get: live OEE per machine, downtime tracking with reason codes, alarm capture and history, a job/work-order overlay so counters know what they are counting, and full AlarmIQ. The 60-minute replay buffer means an operator can ask what happened during the last shift and get a real answer.
What is deferred: multi-site rollups, ISA-88 batch execution, ERP integration, and formal recipe management. None of these are missing from the code — they are turned off because they add complexity a pilot does not need.
The mini tier is the pilot after it works. Typical shape: one plant, several lines, a real production team asking real questions. The single edge appliance grows into either a bigger server or a small cluster, still on-prem, still no cloud dependency.
What turns on: shift management (shift definitions, per-shift OEE roll-ups, shift-based alarm views), batch tracking and traceability, work-order flow with quantities and material genealogy, PM scheduling for maintenance, and mobile access from the plant floor. Reports start shipping to email automatically.
What stays off: multi-site orchestration, cloud data warehousing, formal ISA-88 procedural control. Again — not missing, just off.
Plant tier is one facility fully deployed. Every machine is instrumented. Every shift is scored. Every alarm has a history. Maintenance runs off the schedule the system generates. Quality events are traced to the material lot and the PLC state that produced them.
Full ISA-88 batch execution turns on: recipes, formulas, procedural control, unit allocation. Full ISA-95 turns on: work-order flow from ERP down to the machine and material consumption reported back up. The mobile app becomes a first-class interface, not a companion.
This is where most starter MES tools would have forced a migration. Here, you turn on modules in the license.
Enterprise adds the things that only matter once you have multiple plants: central rollups across sites, cross-plant benchmarking, shared master data (recipes, work orders, materials), multi-tenant security policies, and a cloud layer for reporting and analytics that does not depend on any one plant being online.
Each plant still runs its own local instance. The cloud layer aggregates — it never mediates. If corporate IT goes offline, production keeps running. If a plant goes offline, its data replays to the aggregator when it comes back.
The reason the tiers work this way is that they share one underlying data model. A machine on the micro tier is described by the same fields, in the same schema, as a machine on the enterprise tier. A downtime event captured on day one is the same shape as one captured five years later.
This has a subtle but important consequence: historical data collected during the pilot is directly usable at every larger tier. Your first year of OEE trends does not become a set of import artifacts — it is baseline data for the plant-wide dashboards.
Contrast that with a typical migration from starter MES to enterprise, where historical data usually needs mapping tables, format translation, and often gets abandoned because the effort is not worth the questionable payoff.
The code is one thing. The way it is deployed changes with scale, and the platform handles that transparently.
An operator standing in front of a machine cannot tell which deployment they are in. That is the point. The UI, the dashboards, and the AlarmIQ chat behave identically. All that changes underneath is where the data is aggregated for the roll-up view above them.
A pilot that runs on the same software as the eventual full deployment is a pilot you can actually learn from. The dashboards you like, you keep. The workflows the operators adopt, they keep. The AlarmIQ conversations that turned out to be useful, they carry over.
Contrast the alternative: a pilot on a “small business MES” teaches you almost nothing about what running enterprise MES will feel like — because the enterprise product is a different product. The lessons you take away are lessons about the starter tool, not about MES.
We built the tiers this way because we noticed how often a successful pilot turned into a stalled enterprise decision. The pilot succeeded; leadership said “so let’s do this everywhere;” the team looked at the enterprise migration cost and said “let’s wait.” The wait usually turned into a “never.”
Per-machine pricing that stays per-machine at every tier. No step-function jumps when you cross an artificial boundary. Adding a machine at 5 machines costs the same as adding a machine at 500.
Modules that only exist at higher tiers (ISA-88 execution, ERP integration, multi-site) are add-ons. If you do not need them, you do not pay for them — regardless of your machine count. A 300-machine plant that does not need batch execution runs at the same base rate as a 30-machine plant.
The typical starting point is a single-machine or single-line pilot on the micro tier. Two-week deployment on your own PLCs. Live dashboards inside the first day, AlarmIQ turned on the same afternoon. If the pilot succeeds — and the shape of a “successful pilot” is usually obvious within a few weeks — adding machines is a license count change and a few hours of configuration per machine.
A two-week pilot on your own PLCs. If it does not deliver, walk away.
Scope a pilot See the platformRelated reading: Affordable MES Software · How to Calculate OEE · Deployment topologies