Introduction

KPI Tree View helps reveal how a high-level overall KPI is derived from underlying detailed KPIs. Read this first for a brief refresher on KPIs.

Most IoT device estates exist in one or more hierarchies. For example, if fifty smart meters are all connected to the internet via one cellular gateway, if that gateway goes offline then so will those fifty meters. Using DevicePilot, an operations manager who has just seen the  key “uptime” KPI for their entire device estate (perhaps 50,000 deployed meters) drop by a few percentage points can quickly discover that it’s due to fifty meters in particular - but their next question will be “Why?” - what do those meters have in common? Viewing the devices according to their connectivity hierarchy will quickly reveal the gateway as the common "parent" and therefore the likely culprit.

There can be many hierarchies, all useful. Viewing devices in a hierarchy of "device vendor" or "software version" might help spot a software upgrade issue. Or viewing devices by "installer" might reveal an issue tied to a particular person's installation practices.

For more information about hierarchies and IoT you might also like to read our Digital Twin article.

How to use KPI Tree View

To use KPI Tree View go to the View page, select a KPI at the top of the page (e.g. "Availability by site" in the screenshot below), and select the new Tree View icon at the top right of the Device List pane:

As below, this reorganises your flat list of devices into a hierarchical tree, according to the grouping property of that KPI (e.g. “gateway” or "vendor"). The KPI for each group is shown (e.g. the KPI for vendor 1, the KPI for vendor 2 etc.) and by clicking on the little down-arrow at the left of each row, you can expand any group to reveal the individual KPIs for each device in that group. Now you can understand how each device is contributing to that vendor's overall KPI.

In the screenshot below we see a KPI measuring the uptime of smart meters, grouped by vendor. By using KPI Tree View we can see that one vendor (Itron) has a much worse uptime than the others, and then by expanding that group we can see that indeed all of the meters from that vendor seem to be suffering (so it's a systemic manufacturer problem, not just one "bad apple" ruining the KPI).

A similar example would be a KPI which sums the energy delivered by all the charging points at each customer site. At a high level you’ll see total energy for each site, and then when you drill down and open-up one site you’ll see all the numbers from each charging point which got added to produce that total, over the selected time period. An unexpectedly high or low KPI may be driven by extreme performance from just one device, or just by slightly changed performance from many devices.

REMEMBER: To start exploring KPIs in Tree View, you can "drill down" to the View page from any KPI chart on a Dashboard or the Cohort page - just select a column on the chart and then click the "eye" icon which appears at the top to drill-down into the View page, which will now show those devices, with the relevant KPI already selected - all ready for you to do some Tree View exploring!

Benefits

KPI Tree View allows users to “drill-down” from any KPI to the level below, helping to uncover the general "shape" of the performance of a device estate, then drill-down into the details. Combined with the drill-down capabilities already present in the Dashboard and Cohort pages, this makes it now possible to drill-down all the way down from the most-general KPI to the most-specific KPI and its cause:

• from global vanity metrics such as service delivery rate (that the CEO cares about)
• to per-region or per-vendor contributions (that the country manager cares about)
• down to e.g. per-machine or per-firmware-version contributions (that a technical person cares about)

Now read about another new feature - KPI rules