Best design practices for industrial IoT dashboards

We designed over twenty IoT dashboards. In this article, we share our experiences and best practices. Not all heroes wear capes; some of them just use good dashboards.

Table of Contents

First, the analysis of procedures
Second, an interview with operators
Third, priorities
Fourth, prototype & iterate
What does a good IoT dashboard do, anyway?

In this article, we focus on industrial IoT dashboards. With such a tool, users/operators gather multiple data about their work environment and manage it immediately from one place.

Thanks to information collected from sensors and presented on a simple, intuitive interface, engineers operating factories, airports, and power plants know precisely what is working and what can break down at any moment.

They have the power to react far before any severe issues appear.

Being IoT specialists, we've equipped thousands of operators with the superpower of a functional dashboard. I had the pleasure of managing the design of these systems (such as the complex industrial solution for storing thermal energy), and today, I will share my observations on how to do it the right way and why even do it at all.

First, the analysis of procedures

Hate it or love it, a well-designed IoT dashboard arises from a thorough analysis of existing routines, manuals, or other company procedures. Each industrial enterprise has a list of activities that different employees must perform in a specific sequence at a particular time to ensure that nothing fails, the people are safe, and the system is running smoothly.

We collect these routines and study them to draw a broad roadmap of what we will include in the interface. At the same time, these procedures provide precise information on the data that should be collected and who, when, and how should have access to it.

The first essential task when creating IoT dashboards is getting to know and understand these rules and transforming them into user stories.

Interestingly, at the stage of analyzing and automating procedures, we often catch that some of them contradict each other. Sometimes these cases have close to no consequences, and sometimes we dig out the source of severe daily problems of planners, machine operators, or blue-collar workers.

Second, an interview with operators

In most cases, IoT systems are used by a small percentage of one's company employees. A narrow group of operators, often with years of practice. It is to them that we follow in the second step because the theory and the safety procedures are one thing, but users' daily routine is another.

We conduct a series of interviews with people who manage a given environment daily. We juxtapose the theory from the procedures with the everyday practice of the users.

We aim to find out what could be improved in the current system, what would make their daily work more manageable, and what they feel they lack on a daily basis. We ask employees about processes and data and the preferred way of presenting them. We also look at the typical day of such operators; we watch their daily routines to understand their challenges and needs better.

Employee interviews are a critical stage in designing IoT dashboards because the operators have elusive, unique knowledge, and it's them who, at the end of the day, will use the system and train other people to use them.

Third, priorities

When we already know the theory from the regulations and the operators' routines, struggles, and needs, it's time to set priorities. We've learned that it's not possible o meet the needs of all system users equally. Hence, we collect suggestions from as many operators as possible and set priorities in consultation with them.

What data, where, and how should be displayed on the dashboard?

This stage taught us that you can't create a universal dashboard because one operator uses two and another uses four monitors. For one, the key is to watch engine oil temperature, and for another, the ability to open and close valves. John loves his dashboard to be clear and straightforward, while Josh can't live without eleven screens beeping and flashing constantly.

For this reason, every good panel should include a customization option. So we always build IoT dashboards with widgets that operators can freely change and adjust: move, stretch, and minimize.

Fourth, prototype & iterate

The last step before the actual app development is creating and testing the prototype with the users. We mix up the already gathered data, sketch the panel and then transform it into a functional MVP.

We give the prototype to the users' hands and gather their feedback. Then, in the next round of prototype development, we implement the input and provide the upgraded version to the users again.

And we do this until we all agree that the eagle has landed. :)

What does a good IoT dashboard do, anyway?

The above hierarchy of steps is a recipe for a good IoT dashboard that facilitates the operators' work and streamlines the company's internal processes. Now, there's one question that you might ask, and that is: why so much hustle over a simple interface?

Well, let's break this one down. From my experience, with a well-designed IoT dashboard...

You can foresee the future. The system collects data from various sensors and notices all kinds of deviations from the norm. A good dashboard will inform users that certain events are about to happen or that a failure is imminent and will allow them to react accordingly ASAP.

You get easy access to critical data. IoT dashboards display all devices and their status so that operators can check the inventory and all related data with a blink of an eye (or rather a couple of clicks).

You can draw business conclusions. Thanks to the dashboard, the operator can interpret many different data and thus notice trends and tackle strategic areas, such as purchasing new parts or hiring new employees.

You're less prone to human error. A well-designed dashboard will ensure that the operator won't forget any procedure, and many of the processes will happen automatically. The built-in notification system keeps your employees alert of important deadlines, upcoming threats, or unfinished tasks.

You can visually preview your devices. Although they primarily collect and analyze data, IoT dashboards also show the entire system map in space. The graphical representation of devices and their status gives users a better understanding of the current situation of the whole system, especially when the system contains hundreds (or sometimes thousands) of individual parts and sensors.

You manage the system faster. With a good IoT dashboard, operators can zoom in and out of the screen and select individual system components, such as valves on a pipe, to check their status, temperature, pressure, and so on. Then, they can remotely manage the system from this level, for example, open and close wires or turn sections on and off.

You have all data presented in real-time in a clear form. It's cool to have a flashy app, but most of all, IoT dashboards must be legible and practical. A good interface is simple, easily manageable, and doesn't tire the eyes.

You're safe. IoT systems often analyze and display sensitive data, and only specific people should have access to it. There are different ways to ensure the security of such systems; biometrics, local hosting on a physical server (not in the cloud), two-step verification, and physical tokens.

You're one step closer to a Digital Twin. DT is a virtual representation of a physical system. Thanks to it, you can simulate any situation and, thus, better plan investments. The research needed to create a digital twin largely coincides with the research required to make a dashboard. So, if you're thinking of starting a digital twin of your system, you're halfway there with a sound IoT dashboard.