5 IoT metrics to track

The metrics to track in IoT

IoT devices are often globally distributed and built on multiple hardware and software platforms. Interruptions in the data flow from any device might indicate a device failure, an application-level performance problem, a regional network issue or something else.

Different types of data are relevant depending on the type of IoT deployment you run, but there are some common data types you should track for your IoT fleet.

1. Percentage of devices online

This IoT metric tells you how many devices in your fleet are up and running. Devices usually send a heartbeat message back to the network to indicate they're still functioning. The occasional missed ping isn't indicative of an issue, especially if devices aren't permanently connected to the network. However, a series of missed pings is a problem that you should investigate.

2. CPU availability

CPU availability over time is an essential metric to track for edge computing IoT devices. It indicates how well the device and CPU cope with the workloads sent to and from it. Track this metric over time to determine when to add additional devices or replace outdated ones.

3. Latency

This KPI can help determine how the network handles data transmission workloads. It's especially useful if you use legacy technologies or expand your IoT fleet rapidly. Sometimes, the network cannot handle the added bandwidth, and latency starts to creep in.

4. Site availability

This KPI is useful if devices are deployed in clusters. It indicates how many or how much of the devices' services are being used at any given time. High numbers could indicate that the current deployment is overwhelmed and you need to install more devices. Low numbers could indicate low usage or another problem that requires investigation.

5. Active users

This KPI demonstrates how often users access the device. The churn rate of the device may indicate that users are no longer employing it or that there's an issue with remote users connecting to it.

The value of metrics in IoT

IoT metrics can help you keep track of your IoT fleet, ensure uptime, identify problems before they impact your business and tell you when it's time to replace a device. IoT metrics can also help unlock more business value -- the data is aligned with big-picture business objectives. McKinsey found that IoT could globally unlock $5.5 trillion to $12.6 trillion in value, with 65% of it in B2B companies alone. The report also found that, although the potential economic value of IoT in B2B is significant, it's being held back by the sheer size of many IoT deployments.

Many enterprises struggle to transition from IoT pilots to production, so they often cannot capture the value at scale. One reason for this is that organizations sometimes use the same metrics at all phases of the IoT deployment, yet any project manager would tell you that metrics change from the design to testing to deployment and maturity phases. The metrics that are relevant to your IoT devices and deployments change based on their lifecycle and usage; it's nearly useless to track the entire initiative with the same metrics. For example, metrics such as the number of IoT devices deployed and device costs are typically relevant only at the start of the project or in the installation phase. During the testing phase, track connection bandwidth rates, handshake failure rates and power usage. Once the devices are fully operational and the project is in the maturity phase, track metrics such as availability, update percentages, bandwidth rates and latency.