In Part 1, we looked at our Job to be Done and why it wasn’t building Dashboards. The key takeaway is to focus on user value and solve their pain points. Now, let’s turn to why what building dashboards doesn’t deliver the user value we expect.

Dashboards Fail by not Matching Users’ Needs

The primary way we deliver data and try to solve problems is the modern, visual, interactive dashboard. At times, we may call them other things (reports, scorecards, infographics, etc), but in the day-to-day of business, they are all shades of the same basic object and there are a lot of them. We ask a lot of these dashboards, that can include (but not limited to):

• Be easy to scan
• Allow exploration
• Be performant
• Show the details
• Tell stories
• Be visually appealing
• Allow ad-hoc queries
• Drive users to action
• … And more.

In short, we’ve built solutions with a wide range of features at users’ requests, but often failed to understand their actual need. That has left us with Dashboards that are often hard to find, hard to put into action and hard to incorporate in our users’ work flows.

This isn’t universal. For some users, the dashboard hits the goldilocks zone where it’s just the right fit for what they need. For others, though, it’s not enough. They still need to download data and move it to another system where they can use another to reshape it, enhance it, filter it or process it in some other way.

Yet, for others, the dashboard is too much. They need something that starts at a higher-level. They need an easier way to process the information and understand where they should focus their work efforts.

Thankfully, the data industry is evolving and there are clearly two trending solutions that are arising to address this complexity spectrum.

User Needs — When Dashboards Aren’t Enough

A common question I get is, “How do I get my users to stop downloading to Excel?” It’s such a common experience to spend time building a dashboard only to be asked for a data export. Sometimes the problem is training, but more often it’s the fact that the dashboard isn’t enough to solve the user’s problems. Data Apps are the emerging solution to address when a Dashboard aren’t complex enough.

Current tools are addressing this gap by adding new features to extend beyond the dashboard. For Tableau, this includes embedding APIs that enable tools such as Curator and extension APIs that enable tools such as Apps for Tableau,

There’s also a new generation of tools like Observable, Hex and Evidence that seek to make these targeted user experience easier to build, deploy and maintain.

Across these tools, you’ll see some common patterns including data input and data activation.

Data input (such as write-back) is when we move from a read-only experience of a dashboard to allowing users to enhance the data with their own knowledge or comments. These additional controls allow for a wide-range of new capabilities, including:

• Discussion of specific data points
• Experts adding needed context
• More complex what-if analysis and simulation
• Dynamic target setting
• … And more

Data activation takes this input one step further by putting it to work in other systems. With data activation, we can action data and make it part of our workflow. This is the most powerful technique for user adoption. This can be done through features such as Tableau External Actions, Curator webhooks or by using some of the next-generation data app tools mentioned above. Use cases include:

• Discussing data points in chat apps (like Slack or Teams) for discussion
• Connecting data insights a task management system for tracking
• Sending customer lists to your Sales CMS for targeted promotion
• Saving model settings for future reference and historical analysis
• Creating more data driven discussion with employees about goals
• Implementing review and approval processes
• Speeding up work flows
• …And more

User Needs When Dashboards are Too Much

On the other side of the spectrum are users who just need the headlines. They don’t need to drill down extensively — they want important KPIs highlighted for them, alerts for when things change or need attention, and to get in and out quickly, trusting that the information they have is accurate.

Tableau Pulse, Power BI Metrics and ThoughtSpot KPIs are all examples of answers to this need. However, there is also a growing list of next generation metric focused tools including Steep, Hashboard and LightDash that are all examples of tools that focus on helping you build (or use) a metric layer in your BI workflow.

For end users, all these tools mean a simplified experience with predefined metrics they can browse and follow. The experience moves away from a single website and more into email digests, notifications and chat alerts. In some cases, the tools will offer a simplified self-service experience for building reports leveraging the pre-built metrics.

One key to all of this will be the sematic layer (or metric layer) which we see dbt and start-ups like Cube helping companies centralize. Many (but not all) of the tools in this section are either using these layers as sources or integrating with them to provide better interoperability.

Create Targeted Experiences

For Analytics to Mature, we need to move from delivering Dashboards to a portfolio of options. These will include more complex data apps that connect user work flows and systems, as well as more simplistic metric reporting and consumption options. The good news is that there are lots of options in the industry to solve these needs and perhaps more importantly, these are building blocks that will help us as we move into the era of AI, but we’ll save that for a future installment.

In Part 3, we’ll look at user-adoption and the ways delivery mechanisms impact it.