3 Rules for Realizing value from IoT Analytics

A few months ago, as the pandemic raged on, I was talking to a facilities manager at one of our automotive clients. He was lamenting how despite having invested millions in a SCADA system, he was unable to monitor the energy consumption across his sites in real time. The reason was not the system itself, but the fact that the intra-day audit process, at each of the geographically dispersed plants required significant human oversight. The audits were always prone to error and took a long time to collate, and in the pandemic the absence of sick workers led to significant productivity loss.

This past year has been a constant reminder that not just physical supply chains, but information supply chains too are fragile, beholden to their constituent human and analog links. Simply connecting remote assets, and field workers that throw off data onto static dashboards is not enough. Clearly more needs to be done for automating these links, building a real-time understanding of events, and adding levels of redundancies and security. There are three key rules that can help businesses in realizing value from all of their IoT investments – all the connected boxes and wires:

1. Purpose: Clearly define your business objectives

IoT driven analytics dashboards need to have clearly defined purpose. What is the business objective of a given KPI (Key Performance Indicator) on the dashboard? What actions is the person supposed to take when thresholds are breached? Are these human actions or do they need to be machine-driven?

Device management dashboards are typically the start of assembling a true picture of field assets and their operations. They provide a view into the real-time working of machines, valves, equipment, and other equipment connected via sensors. However, it is one thing to know whether a machine is on or off, quite another to be able to understand how efficiently the machine is running and how productive the throughput is. Dashboards should be designed not simply for operational purpose, and should be extended to function a descriptive, predictive and ultimately prescriptive dashboards. These dashboards should be able to foresee events, suggest actions, trigger workflows, and at a minimum, issue alerts.

Analytics cannot be general purpose and dashboards have to be designed with the user in mind – their role, their age and demographic, the level of sophistication of task, amongst others. The KPIs and the data narrative has to be simplified to be usable and actionable.

2. Process: Build trust in the data

‘Data pipelines’ are channels through which data flows from the myriad connected devices into staging areas, before it is visualized as insight on a dashboard(s). To derive real value from these IoT-driven analytics, these pipelines have to be designed with a view to keeping the data flowing through them clean and trustworthy.

A dashboard with faulty or misleading data leads to damaging actions. Data security is typically the first aspect to be addressed when connecting devices - via encryption, role based access, threat monitoring and other designs. However, security not the only aspect that needs to be addressed. A well-defined, automated process that addresses not only the security, but also privacy, and accuracy of data as it flows through the pipelines is critical. The analog component of this process should be minimized, and programmatic data feedback loops should be in place to check accuracy along each step of the information flow in the enterprise.

3. Resilience:

When ‘boxes and wires’ are attached to field assets (machines, equipment, etc.) they serve  a purpose that is a point-in-time. As businesses evolve, and add or deprecate assets, their monitoring needs will change. The IoT analytics information supply chain has to be resilient to support these changes.

If a manufacturing firm adds a few machines to their assembly line and retires a few others, the IoT analytics data pipelines and dashboards should be able to seamlessly adjust without requiring additional IT support. Such ongoing device changes, regardless of frequency should be invisible to the analytics users.  Scalability to add and delete devices is a critical aspect of building resilient information supply chains. Similarly dashboards should automatically alert the human elements in the supply chain, or to trigger workflows based on predictions.

To realize real value from IoT investments, businesses need analytics built on top of flexible, scalable information supply chains. Understanding the three rules: defining purpose, building trust in the data and resilience, are useful as firms look deeper into designing new IoT systems, or harnessing the power of existing ones.

Accelerity Infoventures is a software company focused on IoT analytics and machine learning. It provides a cloud-based analytics platform as a service (PaaS) called Aceiot for B2B, small and medium sized enterprises. Aceiot is a horizontal platform that can ingest streaming data from any IoT device/sensor/gateway and aggregate it. Aceiot also provides a business rules engine to trigger events based on data, and a machine-learning component that can apply algorithms to the ingested data to predict events. Key performance indicators (KPIs) and real-time and predictive insights are displayed on the aceiot self-service dashboards. Aceiot also works with all standard IoT protocols. Unlike device based operational analytics providers, aceiot focuses on the business end-user needs. Aceiot simplifies and subtracts data noise to display only relevant KPIs and insights to the end-user.

Find out more about Accelerity and Aceiot

Follow Vivek Mehra on LinkedIn