An IoT sensor catalogue can list a dozen things worth measuring on a single machine: stop time, vibration, temperature, speed, and more. Installing every available sensor is rarely worth it; the question that actually matters is whether a specific sensor's own monthly value, the cost or time it helps avoid, exceeds what it costs to install and run in the first place.
Cedar & Co., a domestic brand running 12 shops on a Make to Stock model, is reviewing which sensor to install on a warehouse conveyor motor, so this lesson works through how to tell a sensor worth installing from one that is not.
You need: no previous lesson is required, though reading machine data (Lesson 2.2) is useful preparation.
By the end of this lesson you can:
Not covered here: calculating a full payback period across a connected machine's whole cost picture (Lesson 2.4), or building the complete worksheet (Lesson 2.5). This lesson only covers choosing which sensor to install in the first place.
| Term | Meaning |
|---|---|
| IoT sensor | A small device attached to a machine that measures one specific fact, such as stop time, vibration or temperature, and reports it automatically. |
| Sensor installation cost | The one-time cost of installing one IoT sensor: the sensor unit itself, wiring or mounting, and connecting it to a reporting system. |
| Monitored fact | The single specific measurement a given sensor is built to capture, such as motor vibration or stop time. |
| Monthly value | The cost or time a specific sensor's own data actually helps a business avoid each month, once it is installed and used. |
| Downtime cost | The value of production lost while a machine is stopped, whether for a scheduled visit or an unplanned breakdown. |
| Low-value sensor | A sensor whose monitored fact rarely changes or rarely causes a real cost, so its own monthly value stays below its installation cost. |
Not every measurable fact on a machine is worth measuring; a sensor watching something that rarely changes, or that changes without ever costing real money or time, adds installation cost without adding real value. The decision is not "can this be sensed" but "does sensing this specific fact actually save more than it costs, on this specific machine."
| What can differ | The shapes you meet | Why it matters to you |
|---|---|---|
| Machine's own failure pattern | A conveyor motor that has stopped unexpectedly several times this year has a real, recurring stop-time cost; a motor with no history of unexpected stops has little for a stop-time sensor to actually catch. | A sensor watching a fact that already causes real, recorded cost is far more likely to earn its own installation cost than one watching a fact with no track record of causing loss. |
| Business model's tolerance for a stoppage | Cedar & Co.'s own Make to Stock warehouse can usually absorb a short conveyor stoppage by shifting staff to manual handling; an export Make to Order line depending on the same conveyor to hit a fixed ship date cannot absorb the same stoppage as easily. | The same sensor can earn its own cost on one business model and not on another, because the cost of the event it monitors differs by how much slack the business actually has. |
| How the data would actually be used | A stop-time sensor that triggers an immediate alert to a manager who can act on it creates real monthly value; the same sensor logging data nobody reviews creates none, regardless of what it measures. | A sensor only earns its installation cost if its data is actually used to prevent or shorten a real cost; installing a sensor without a plan to act on its data wastes the installation cost entirely. |
| Scale of the operation | A single warehouse conveyor at Cedar & Co.'s own scale has one motor worth reviewing; a larger export factory's finished goods store might have several conveyors, each with its own separate case to make. | At a larger scale, the same decision must be made machine by machine, since a sensor that earns its cost on one motor does not automatically earn it on every similar motor in the building. |
Cedar & Co. runs 12 shops on a Make to Stock model, buying from Delta Garments on a 60-day open account. Its own central warehouse relies on one conveyor motor to move stock between receiving and its shop-dispatch area. That motor has stopped unexpectedly four times in the past year, each time for about 45 minutes while a technician was called and the fault found.
Cedar & Co.'s own warehouse manager is reviewing two sensors offered by an installer: a stop-time sensor, reporting exactly when the motor stops and for how long, and a vibration sensor, monitoring wear patterns that could predict a mechanical failure before it happens. The vibration sensor is marketed as more advanced, but the motor has no history of the gradual wear failures it is built to catch; every stoppage so far has come from a sudden electrical fault the vibration sensor would not detect early.
The worked numberTen dollars a year in labour saving alone does not clear the sensor's USD 180 cost quickly through labour saving; the real monthly value comes from a separate fact: each 45-minute stoppage previously risked a missed same-day dispatch to one of Cedar & Co.'s own shops, a cost the warehouse manager estimates at USD 40 per missed dispatch, avoided in at least 3 of the 4 stoppages once the alert lets staff reroute stock immediately.
The stop-time sensor's own USD 130 annual value does not fully clear its USD 180 cost within one year either, but it is close and trending toward a genuine payback the next lesson calculates in full; the vibration sensor's own value is close to zero on this specific motor, since it is built to catch a failure type this motor has never actually shown.
A vibration sensor sounds like a more sophisticated upgrade than a simple stop-time sensor, and a business impressed by the more advanced option can install it without first checking whether the machine has ever actually shown the failure type it is built to catch. Spending more on a sensor that monitors a fact the machine does not have a history of is money spent on a feature the business does not need, while the simpler sensor that matches the machine's real failure pattern goes unconsidered.
Not every fact a sensor can measure on a machine is worth measuring on that specific machine. Before choosing a sensor, check what has actually gone wrong with the machine in the past, and pick a sensor that watches for that, not the most advanced option on a supplier's list. A sensor only earns its own cost if what it watches actually happens often enough, or costs enough when it does, to add up to more value than the sensor itself costs to install.
Cedar & Co.'s warehouse manager is presenting the sensor decision to Omar, the owner, ahead of budget approval.
Should Omar approve the stop-time sensor only, both sensors, or neither? Decide before opening the answer.
Approve the stop-time sensor only. The two facts that decide this are the direct match between the stop-time sensor and the motor's real failure history, and the vibration sensor's near-zero value on this specific motor.
1. The stop-time sensor's own USD 130 annual value comes close to its USD 180 cost and matches a failure type the motor has actually shown four times, making it the sensor most likely to earn its own cost within a reasonable period. 2. The vibration sensor's own value stays near zero on this motor because the motor's failures have all been sudden electrical faults, not the gradual wear pattern the vibration sensor is built to catch; its cost would be spent with almost nothing to show for it. 3. Installing both adds USD 260 in cost for almost no additional value, since the vibration sensor's own contribution does not change based on whether the stop-time sensor is also installed.
Approving both sensors because "more monitoring is always better" ignores that a sensor's own value comes from matching a real failure pattern, not from the number of sensors installed on one machine.
USD 260 spent on a sensor with no matching failure pattern on this motor is money that could instead fund a stop-time sensor on a second machine with its own real stoppage history, which is exactly the kind of trade-off a business makes when it chooses sensors by catalogue appeal rather than by the machine's own actual record.
Use the T26-C02 MES and machine-data worksheet, "Sensor selection" sheet. For a specific machine, list each candidate sensor, its own installation cost, the failure type it monitors, and whether that failure type matches the machine's own real history, then let the sheet flag which sensor is worth installing. Start with Cedar & Co.'s own example above, then add a machine of your own.
[Download the MES and machine-data worksheet](../../templates/T26-C02-mes-and-machine-data-worksheet.xlsx)
1. What actually decides whether a specific IoT sensor is worth installing on a specific machine?
2. Why did the vibration sensor create almost no value for Cedar & Co.'s own conveyor motor?
| EN | AR | FR | TR | Meaning |
|---|---|---|---|---|
| IoT sensor | مستشعر إنترنت الأشياء | Capteur IoT | IoT sensörü | A small device attached to a machine that measures one specific fact, such as stop time, vibration or temperature, and reports it automatically. |
| Sensor installation cost | تكلفة تركيب المستشعر | Coût d'installation du capteur | Sensör kurulum maliyeti | The one-time cost of installing one IoT sensor: the sensor unit itself, wiring or mounting, and connecting it to a reporting system. |
| Monitored fact | الحقيقة المرصودة | Fait surveillé | İzlenen gerçek | The single specific measurement a given sensor is built to capture, such as motor vibration or stop time. |
| Monthly value | القيمة الشهرية | Valeur mensuelle | Aylık değer | The cost or time a specific sensor's own data actually helps a business avoid each month, once it is installed and used. |
| Downtime cost | تكلفة وقت التوقف | Coût des temps d'arrêt | Duruş maliyeti | The value of production lost while a machine is stopped, whether for a scheduled visit or an unplanned breakdown. |
| Low-value sensor | مستشعر منخفض القيمة | Capteur à faible valeur | Düşük değerli sensör | A sensor whose monitored fact rarely changes or rarely causes a real cost, so its own monthly value stays below its installation cost. |
Next you will take a sensor's own installation cost and monthly value and calculate the full payback period for connecting one specific machine. Back to the course map