Machine monitoring replaces whiteboards and paper slips with live data on whether each machine is running, stopped or idle, how much it produced and why it stopped. Start with a few key machines, capture run status and count, add downtime reasons, then build simple dashboards. The value comes from acting on the data every shift, not from the sensors alone.
- Begin with a handful of bottleneck machines rather than the whole plant.
- Run status, output count and downtime reason answer most early questions.
- Older machines can be connected with low-cost sensors or signal taps.
- Dashboards only help if supervisors review them in shift meetings.
- Involve operators early, or the data will be ignored or gamed.
Why do factories still run on whiteboards and paper?
In most plants the day's production is chalked on a board or written on a slip, collected at shift end and typed into a spreadsheet the next morning. It works until someone asks why output fell last Tuesday, and nobody can say whether the cause was a power cut, a tool change, a missing operator or waiting for material.
Chennai is well known for its manufacturing and automotive belt, and plants across India, large and small, face the same blind spot. The information exists on the shop floor, but it is trapped in memory and paper. By the time management reads the report, the moment to fix the problem has passed.
What should you measure first?
Resist the urge to measure everything. Three signals answer most early questions: is the machine running, how many parts did it make, and when it stopped, why. From these you can calculate utilisation, output against plan and the main causes of lost time.
Many plants then move on to overall equipment effectiveness, which combines availability, performance and quality. That is useful, but only once the basic data is trusted. Start with the simple view and let the team decide what else they want to see after a few weeks of use.
- Machine state: running, idle, stopped, in setup
- Part count or cycle count per hour and per shift
- Downtime reason picked by the operator
- Rejections and rework counts
- Energy or temperature where it explains problems
How are machines actually connected?
Newer CNC and PLC-based machines can often share data over standard industrial protocols. For older machines without any interface, a sensor on a signal tower light, a current clamp on the motor or a counter on a stroke can still show when the machine is working. The approach is chosen machine by machine.
A small gateway at the machine or line collects the signals and sends them to a local server or the cloud. Think about network reliability and what happens when the connection drops, since the gateway should store data and send it later. Good partners will survey a few machines before quoting a plan.
How do operators fit into the picture?
Machines can tell you that they stopped, but not always why. A simple screen or tablet at the machine where the operator picks a reason takes seconds and turns raw downtime into useful categories such as waiting for material, tool change, maintenance or quality hold. Keep the list short and in the operators' language.
Be open about the purpose. If operators think the system exists to punish them, they will choose vague reasons or skip entries. Present it as a way to expose problems that are not their fault, such as late material or repeated breakdowns, and show them when the data leads to a real fix.
- Keep the reason list to six or eight plain choices
- Show operators their own shift output
- Share fixes made because of their entries
- Never use the data to blame individuals
How do dashboards turn into decisions?
Put a live board on the shop floor and a summary on the owner's phone. Show current status of every machine, output against target, and the top downtime reasons for the shift. Alerts for long stoppages can notify supervisors or maintenance by WhatsApp or email so they react quickly.
Then build the habit. Review last shift's numbers in a short stand-up, pick one loss to attack, and track whether it improves. Over time, connect the data to maintenance schedules, production planning and your ERP, so a pattern of breakdowns triggers preventive work and plans reflect real machine capacity.
What mistakes should you avoid?
The usual mistakes are starting too wide, buying hardware before deciding what questions to answer, and ignoring the people side. Another is building beautiful dashboards that nobody checks. Choose one plant area, define three questions, and deliver answers within weeks rather than months.
Also think about ownership. Name someone responsible for the data, for sensor upkeep and for the shift review. Plan for security on the factory network, access control on dashboards, and backups. A small, well-used system beats an ambitious one that quietly goes stale.
Step by step
- Pick pilot machines. Choose three to five bottleneck or problem machines and list the questions you want answered.
- Survey and connect. Decide per machine whether to read controller data or add sensors, and install gateways.
- Capture downtime reasons. Add a simple operator screen with a short list of reasons.
- Build the first dashboard. Show status, output against plan and top downtime reasons for each shift.
- Hold shift reviews. Review the numbers in a short daily meeting and choose one loss to reduce.
- Extend and integrate. Add more machines, then connect with maintenance and ERP planning.
Frequently asked questions
Can old machines be monitored?
Often yes. Signal tower lights, current clamps or counters can show when an older machine runs, even without a digital interface.
Do we need cloud for this?
Not necessarily. Some plants prefer a local server, others use cloud for remote access. The choice depends on network quality, security policy and where you need to view data.
Will operators resist it?
They may if it feels like surveillance. Involve them early, keep input simple and share the improvements the data makes possible.
What is OEE and do we need it?
OEE combines availability, performance and quality into one view. It is helpful later, but start with status, count and downtime reasons.
How does this connect to our ERP?
Production counts and downtime can feed the ERP or reports through integration, so planning reflects what machines really do.
Need help with this? Ask us a question about it — we reply within one working day.
