Most factory managers aren’t comparing monitoring platforms. They’re deciding whether to monitor at all. The machines are running, parts are shipping, and the whiteboard in the office has some OEE calculations on it from last Thursday. Good enough, right?
Here’s what “good enough” actually costs, and whether £150 per machine per month changes the equation.
The real cost of not knowing
The headline figures on manufacturing downtime are enormous — IDS-INDATA projects UK and European manufacturers will lose over £80 billion to unplanned downtime in 2026. Fluke’s research found that 68% of UK manufacturers suffered unplanned downtime in the past year, with incidents costing an average of £1.36 million per hour.
Those numbers come from automotive plants and pharmaceutical facilities. They’re real, but they’re not your numbers. Scale them down to a real-world scenario: a manufacturer running 10 to 30 machines across one or two sites.
The RS Components and IMechE Maintenance Engineering Report found that unscheduled downtime costs UK manufacturers an average of £5,121 per hour. For smaller firms, 39% reported costs under £500 per hour. For larger operations, 24% reported costs above £10,000 per hour. The average UK manufacturer loses 49 hours of production per year to unplanned downtime — roughly an hour a week.
Take a modest example. A shop with 15 CNC machines, each generating £400 per hour in production value. One unplanned stoppage per week averaging two hours, across the whole shop. That’s £800 per week, or £41,600 per year — in direct lost output alone. Before overtime to catch up. Before scrapped parts from mid-cycle failures. Before the late delivery penalties. Before the production manager spent half a morning rescheduling jobs instead of running the floor.
Manufacturers who start tracking downtime properly consistently discover it costs two to three times their initial estimate. The number you have in your head right now is almost certainly too low.
The hidden costs nobody tracks
Lost output is the number everyone calculates. It’s also the smallest part of the real cost. When a machine goes down unexpectedly, a chain reaction starts that doesn’t show up on any maintenance report.
The planner has to rejig the schedule. Operators stand idle or get moved to machines they’re less efficient on. Overtime gets booked later in the week to recover lost hours. If the failure happened mid-cycle, the part in the machine is probably scrap. Downstream processes that depend on that machine’s output start backing up. And if the delivery date slips, your customer’s view of your reliability just took a hit — the kind of cost that compounds invisibly over months and years.
Then there’s the time spent figuring out what happened. The supervisor walks over when someone reports the stoppage — often 20 minutes after it actually occurred. A fitter is called. The fault is diagnosed. Parts are ordered. By the time the machine is running again, four hours have gone. And nobody logged the first 20 minutes because nobody knew about it.
This is the core problem that monitoring solves. Not predictive maintenance, not AI-driven analytics — just knowing that a machine has stopped, when it stopped, and getting that information to the right person immediately instead of whenever someone happens to walk past it.
What “doing nothing” actually looks like
It looks like normality, which is why it persists. Most factories that don’t monitor their machines don’t think of themselves as neglecting anything. They have maintenance schedules. They have operators who report problems. They have a production manager who walks the floor. It works. Mostly.
But “mostly” has a cost. The maintenance schedule is calendar-based, not condition-based — so you’re either servicing machines that don’t need it or missing the ones that do. Operators report problems when they notice them, not when they start. The production manager can only be in one place at a time and can only see the machines in front of them.
The whiteboard OEE numbers? They’re collected manually, entered hours or days after the fact, and they’re optimistic — because nobody records a five-minute stoppage, and those five-minute stoppages add up to hours over a month.
Over 80% of industrial businesses experienced unplanned downtime in the last three years. The average incident lasts four hours. Ageing assets are the leading cause, cited by 28% of maintenance professionals, followed by mechanical failure at 18% and operator error at 10%. None of these causes announce themselves in advance — unless you’re watching.
What monitoring changes
Machine monitoring doesn’t eliminate downtime. Nothing does. What it eliminates is the gap between a machine stopping and someone knowing about it, and the gap between knowing something went wrong and understanding what happened.
With RoboVigil, when a machine’s state changes — running to faulted, — an alert goes to your phone. Not to a screen in the control room that nobody’s watching at 2am. Not to an email inbox that gets checked at 9am. To the phone in your pocket, in real time.
The production manager who used to find out about the 6am breakdown at 8:30am now knows about it at 6:01am. The maintenance engineer who used to diagnose faults by walking the floor can pull up live data and camera feeds from their phone before they’ve even left the car park. The operations director running three sites can see every machine across every factory without being physically present at any of them.
The data accumulates, too. After a month of monitoring, you know which machines stop most often, which shifts have the most stoppages, which fault codes keep recurring. That’s not AI — it’s just visibility. And visibility is what turns reactive maintenance into planned maintenance.
What monitoring costs vs what it saves
RoboVigil costs £150 per machine per month. No hardware costs, no installation fees, no per-user charges.
For a 15-machine shop, that’s £2,250 per month, or £27,000 per year.
Put that against £41,600 per year in direct lost output, using conservative assumptions. If monitoring helps you catch even one stoppage per week 30 minutes earlier — reducing average downtime from two hours to 90 minutes — you’ve saved £400 per week, or £20,800 per year. The monitoring has nearly paid for itself on that single improvement alone.
And that’s before the overtime you didn’t need to book. The scrap you didn’t create. The delivery you didn’t miss. The customer who didn’t start looking for a backup supplier.
For a larger operation — 50 machines across two sites — the numbers scale accordingly. The monitoring cost is £7,500 per month. But those 50 machines represent a much larger attack surface for unplanned downtime, and the value of early detection multiplies with every additional machine.
There’s no contract lock-in. If monitoring doesn’t pay for itself within three months, cancel. But in 25 years of working in factories across 20 countries, I have never seen a manufacturer start monitoring their machines and then decide they preferred not knowing.
What RoboVigil doesn’t do
RoboVigil is not an MES. It won’t schedule your jobs, manage your inventory, or run your quality system. It’s not a SCADA replacement — if you already have a SCADA system, RoboVigil adds cloud and mobile access alongside it. It’s not a predictive maintenance AI that claims to know when your spindle bearing will fail next Tuesday.
It’s machine monitoring and alerting. Live data from your machines, live video from your cameras, alerts on your phone, historical data you can actually use. Done simply, done cheaply, deployed in hours not months.
For most small and mid-sized manufacturers, that’s 80% of the value for 20% of the cost and complexity of a full MES or SCADA deployment. And it’s infinitely better than the alternative — which is not knowing what your machines are doing right now.
Who should keep doing nothing?
If you run one or two machines and you’re standing next to them all day, you don’t need monitoring software. You are the monitoring system.
If your machines have no network connectivity and no OPC-UA, MQTT, or other data interface, software-only monitoring can’t read data that doesn’t exist. RoboVigil’s camera-based monitoring still gives you visual oversight, but the real value comes from machine data. For legacy machines with no modern interface, that’s a genuine limitation but you can add sensors and a MQTT box for a few hundred pounds.
If you’ve already deployed a full MES or SCADA system that gives you everything you need, including mobile access and real-time alerts, then you’ve already solved this problem. Frankly you probably already have several people employed specifically to handle this stuff for you.
Everyone else — manufacturers running 5 to 200 machines who know they should be monitoring but haven’t found something quick to deploy, affordable, and free of IT overhead — that’s who RoboVigil is built for.
Get started
Sign up at robovigil.com, download the app, and connect your first machine. If your machines speak OPC-UA or MQTT, you can be monitoring within the hour. No hardware to order, no engineers to book, no IT department to persuade. Point a camera at the machine, connect to the data source, and start seeing what you’ve been missing.