Data is power, and every business wants to harness that power for both their customers and themselves. Opportunities to connect users with information about the systems they use, manage, and live within are all around us and often seem like quick paths to large value.

But where is the best place to start? Greenfield development can be overwhelming, especially in an IoT (Internet of Things) space where seemingly anything is possible. On an industrial scale, a traditional approach such as Condition Monitoring can feel like “boiling the ocean,” as there is more data than a user can possibly derive value from.

Value isn’t delivering the most data, it’s delivering the right data. In this article, we discuss four best practices to help you focus on what users actually need, navigate technical blockers, bake in a sound design foundation, and leave opportunities for your monitoring platform to grow as technology evolves.

1. Align your team on IoT Dashboard design strategy.

Throughout the development cycle, you’ll want to employ the only tenet of agile/scrum methodology I’ve found truly universal: Get honest feedback early and often. A great final product can’t happen without first getting to your users’ ‘jobs to be done’ and measuring how well your proposed tool meets their needs along the way. This ensures that the light at the end of your development tunnel is blissful rather than blinding.

• Value Opportunity Understanding: The initial persona/empathy work we bring to engagements like this grounds everything we do in user value.
• Agile Feedback Model: Take a collaborative approach and get demos in front of real users. Remember that negative feedback is just as valuable as praise and be ready to pivot if it turns out that your baby is ugly.
• Access to Users: Real Customers = Real Feedback = Real Value.

2. Ground your design & feature prioritization in user needs.

As my colleague James Wyler wrote in How to Prioritize Innovative Concepts, taking the time to prioritize can dramatically improve your outcomes, innovate faster, and increase the efficiency of your innovation efforts.

As best practice:

• Understand your Who, What, Where, Why and How: What are we building, measuring, and monitoring? What hardware capabilities do we have today? Where are users currently accessing or getting similar data today? Why does anyone care about this data? How will users access this data via our offering?
• Get to the heart of your users’ daily quick access needs: Data wishlists for industrial monitoring are often unending. Prioritization workshops often reveal what functionality can be moved down the totem pole or even dropped all together. Remember: Value isn’t delivering the MOST data, it’s delivering the RIGHT data.
• Avoid feature bloat: Too much information at once hinders your user’s ability to absorb any of it. Piling graphs on top of condition states on top of predictive analytics is a guaranteed way to prevent a user from having any idea what they’re doing on a given page.

3. Prevent adoption friction by calling out import/export needs early.

Meshing the hardware and software components in IoT systems is usually acknowledged but rarely fully understood by all constituents within a connected systems development effort. I’ll call out a few of the heavy hitters below.

• Data Types (Real Time, Predictive, State): Real-time data visualization almost always comes with a time window selection. Make sure you aren’t pigeon-holing users by dictating which chunks of data are available. Predictively derived data needs to be differentiated from live information so users understand what is a problem now versus what could be an issue six months down the road.
• Exporting / Customization: Rarely do IoT systems exist in a vacuum. Whether it’s an existing internal ERP system, corporate-managed data historian or a vendor-served external tool, today’s data aggregation and visualization systems need to learn how to play nicely with others. Data view/page customization is another common theme. Industrial organizations are diverse; their systems, practices and user expectations aren’t standardized. It is highly likely that new IoT monitoring tools will need to let users choose what they view on certain pages and how they would like to view it.
• Non-Automated Data: IoT monitoring systems can also incorporate internally driven goals from sales, production, or facility management. The intake and processing of information like this should be considered from both visualization and use/adoption viewpoints when reviewing potential designs.

4. Keep your IoT dashboard design clean and intuitive with these 6 principles.

We’ve grounded our effort in user understanding and considered a few external variables that will likely affect the tool you’ll be building. Now we’ll cover a few design tips you’ll want to make sure are top of mind when building/reviewing mockups throughout your development process.

1. Asset Navigation: Zoning is a common navigational theme for industrial systems. A company has a given number of plants, a plant has a given number of assets, etc. Mapping out these user flows ahead of time and determining what information is valuable in each area is essential to ensuring that users don’t get lost when zooming in on a problem area. This, in turn, generates a visual navigational strategy that lets users quickly navigate between macro and micro views for the systems they operate.
2. A word on navigation menus: Don’t be afraid to make these collapsible! Real estate is always a hot commodity in digital design, and even more so in dashboards/data visualization.
3. Less is More: Proper dashboard etiquette is all about the right information and the right place, at the right time. Don’t overwhelm your users with a sea of graphs; focus on the most pertinent information and save the rest for a breakout page. Maintain appropriate whitespace in your design by keeping graph legends within a tooltip or hover state. Don’t give users 20 default visualizations, let them select and configure the ones they need with dropdown menus and expandable graphs.
4. Data Hierarchy: Keep each area of your dashboard focused and precise by visually highlighting the “headline” information for that page. This is typically achieved with font sizing, page location or color highlighting.
5. Notification Strategy: Don’t wait until the end of your monitoring effort to think about how users will be made aware of state changes. In-platform notifications require real estate to execute and will need to have color indications consistent with the rest of your platform. External notifications like text and email are going to continually become more common as IoT monitoring systems get further integrated with existing platforms.
6. Color as an Indicator: Color is one of the most instinctive ways users group, prioritize and make decisions about data. Consistency is key here—don’t make red a positive state in one area of the platform and a negative in the next. Collapsed legends are also your friend, but avoid overloading a user with a rainbow graph. If you’re displaying more than five individual traces in a single visualization, it’s probably time to ask how that information might be better displayed.

Helping users make better decisions with data is a paradigm that isn’t going anywhere any time soon. Doing so effectively will separate market leaders from laggards for years to come. Thank you for reading and get in touch with us here at Method if you’d like to discuss your own IoT dashboard design or other experience design needs.