IoT for Manufacturing — Where to Start in 2026

Most manufacturers have no idea how much their machines actually run — not because they don't care, but because nobody measures it. IoT for manufacturing starts exactly there: turning "we think the cell runs at 70%" into a hard number from a sensor.

This article shows where to realistically begin an Industrial IoT rollout — without replacing your machine park or budgets in the hundreds of thousands.

What IoT in production really means (no buzzwords)

IoT on the shop floor is a network of sensors and controllers that collect data straight from machines and lines, then push it to one place: a dashboard your production manager and board can see.

In practice we usually measure:

• machine run / stop time (the most important and cheapest to capture),
• units produced (cycle counting),
• process parameters — temperature, vibration, pressure, power draw,
• downtime reasons (the operator picks a code on a panel).

From this data we build OEE (Overall Equipment Effectiveness) — one of the most important metrics in manufacturing, combining availability, performance and quality.

Where to start — 4 steps

1. Pick one bottleneck, not the whole factory

The most common mistake is trying to "digitize" everything at once. Start with a single machine or cell that is your bottleneck — every point of availability you recover there directly increases the throughput of the whole plant.

2. Measure availability before changing anything

The cheapest, fastest pilot is a sensor detecting whether the machine is running (a current clamp, a vibration sensor, or a PLC signal). A simple "run / stop" chart over a week is often a shock — real OEE is usually 15–25 points lower than management assumes.

3. Add context — downtime codes

Once you see when the machine stops, it's time to learn why. A simple operator panel (a tablet or an industrial button) where they pick the reason turns raw data into a concrete list of problems to fix.

4. Automate reports and alerts

The final step is a report that lands in the manager's inbox each morning, plus an alert (SMS / Slack / email) when a machine stops for longer than X minutes. From then on you react in real time, not at month's end.

What does it cost?

Good news: getting in is cheaper than most firms assume.

Item	Ballpark cost
Sensor + controller per machine	€70–350
Gateway / data concentrator	€180–700
Pilot deployment (1–3 machines)	from a few thousand €
Dashboard + alerts (SaaS)	monthly subscription

A single-machine pilot can pay for itself in weeks — just by cutting one recurring downtime thanks to the data.

What to avoid

• Vendor lock-in — beware systems you can't export your own data from.
• Collecting data "just in case" — measure what you'll act on.
• Ignoring operators — they enter the downtime codes; if the system gets in their way, the data is worthless.

FAQ

Do I have to replace old machines to deploy IoT?

No. Even a 30-year-old machine with no controller can be instrumented externally — with a current, vibration or proximity sensor counting cycles. That retrofit approach is one of the biggest advantages.

How fast will I see results?

You get the first data on real machine availability within days of mounting a sensor. Optimization decisions — within the first month.

Is the data secure?

Yes — the system can run fully on-premise, with nothing sent to the cloud, if company policy requires it.

Summary

Deploying IoT in a manufacturing company doesn't have to be a massive digital-transformation project. Start with one machine, measure real availability, add downtime codes and automated reports. The rest is scaling a proven pattern.

At Kajpa Studio we design these rollouts end-to-end — from sensor and firmware, through the gateway, to dashboard and alerts. Get in touch and we'll show you which machine to start with.