An introduction to IoT Analytics

A tour around the landscape and the various terms used

"IoT Analytics" is now a recognised product category, with its own box in Matt Turck's annual IoT Lumascape, though across the industry it's still referred-to by several alternative names. Here we survey the landscape and highlight the areas where DevicePilot particularly shines:

  • IoT Analytics: This broad category covers the idea of taking IoT datastreams (plus maybe metadata) and turning it into higher-level information and insight. DevicePilot focuses on the live use-cases such as telemetry visualisation and operational management. Non-IoT-specific analytics tools such as general BI (Business Intelligence) tools may be a better fit for generalised, offline analysis (e.g. analysing your pricing strategy). And of course there are multiple IoT Analytics providers, for example Mnubo. A distinguishing feature of DevicePilot is its ability to work with existing IoT platforms.
  • IoT Telemetry (or Time Series) Visualisation: If you're a developer with your first IoT devices sitting on a lab bench, you want to immediately see what data they're producing, debug them and demonstrate them to your colleagues (even before an application has been written). DevicePilot solves this problem, for free, in seconds, with just one line of code. Alternatives ways to visualise time-series include the likes of Kibana, though this is a relatively technical tool to use, whereas DevicePilot is aimed at regular business users.
  • IoT Dashboard: You might use this term if you need to pull information from multiple devices into one place, whether you have 5 devices (displaying data from each on a dashboard) or 50,000 (displaying summarised performance metrics for the entire device fleet). Generic dashboarding tools such as Tableau, Qlik and Looker can't easily digest the large volume of data that IoT produces, but you can use DevicePilot as an intermediary to turn it into higher-level metrics, then send the onto these dashboards to combine with other business metrics.
  • IoT Operational Management (or Monitoring): Does what it says on the tin. As an IoT business grows beyond 1,000+ devices it becomes apparent that someone needs to be put in charge of all the devices in the field - are they working, are customers happy? That becomes the #1, daily mission for the company, and the person in charge of that (and in time their Operations team) needs a tool to help them do that job for an growing number of devices.
  • IoT Customer Support: If Operations is the "back-of-house" activity (sometimes it's called 2nd/3rd-line support) then Customer Support is the "front-line" activity: engaging directly with customers. Sometimes these two teams are merged into an overall "COPS" team. Whereas Operations will mainly look at issues which could affect the whole device estate, Customer Support personnel will often be looking at issues with a specific device, and need a tool to quickly show its state and recent performance. The automation DevicePilot provides can be very helpful too, e.g. to raise a ticket in your customer support tool when a device goes wrong, allowing you to move support from a reactive process (wait for the phone to ring) to a proactive one (be the first to know about problems, and fix them before the customer even notices).
  • IoT Product Management: The Product Manager, who owns the overall product proposition, is another kind of person who needs to be able to see how users are interacting with connected devices. DevicePilot provides an aggregated view of how all users are interacting with the product, e.g. which buttons are they pressing, their level of engagement etc. (rather like what Google Analytics does for a website).
  • IoT Service Assurance (or Network Assurance): Along with the acronym OSS, if you use these words to a telco person they understand immediately what DevicePilot is about - but no-one outside of that industry uses them!
  • IoT Fleet Management (or Estate Management or Asset Management/Tracking): As soon as you have more than a few hundred devices deployed, that's too many to treat them as individuals most of the time, so you will often be thinking at a macro scale about all of them - your device estate or fleet. You'll want information about all of them, or at least about a large cohort of them (e.g. "how many of my sensors deployed in Walmart have gone offline in the last day?").
  • IoT Orchestration: This musical analogy captures the idea of high-level IoT visualisation and management quite nicely: the individual musicians in an orchestra are the connected devices, and DevicePilot helps you be an effective conductor.
  • IoT Device Management: We saved the most problematic term for nearly last. It's perhaps the most obvious term for what we do, and is problematic only because lower-down in the IoT software stack there is already a layer/function with the same name! In embedded-software-land device-management is the function which can e.g. securely and robustly upgrade the firmware on one device. This is a vital and complex task, but it's a much lower-level concept than the idea of generally managing your devices. For example, if you want to upgrade your fleet of 1,000 devices, then at some point you will need to trigger that device management function on each of those devices individually. But something needs to manage that process overall - which devices are you going to upgrade? under what conditions? did it work? That something is DevicePilot.

Most product companies are now looking to connect their products, so the term "IoT" is becoming rather too broad to be useful (for the same reason that no-one calls themselves "Web" companies any more). So companies deploying connected products today don't use the term IoT, but perhaps call their products "Smart" or "Connected". So the above terms are often allied with those words to form phrases such as:

  • "Device Fleet Management for smart parking sensors"
  • "Operational Management for connected streetlights"
  • "Dashboards for connected cameras"
  • etc.