Let’s talk about dashboards
People have strong opinions about dashboards: are they dead? Are they here for good? Why isn’t there a better way to consume data? No matter where you land on those big questions, it’s worth confronting a basic fact: most businesses make big investments in business intelligence tools, most data teams produce a lot of dashboards, and most dashboards end up stale and unused.
The data visualization technology we all have access to today is top notch, so we’ve found ourselves asking, how did we get here and what can we do to make it better?
A brief history of BI
Business intelligence tooling emerged in the 1990s as part of a growing trend for companies to invest in analytics and insights. The promise of these softwares was a more simplified, accessible and efficient solution that allowed business people to work directly with the data they needed.
20 years later, BI tools sit on top of cloud storage solutions and promise to make data and analytics “self-serve” for their users. That means not only can data teams write SQL and visualize their results in charts and dashboards, there is an expectation that the business user can access and track the metrics they need. For some workflows this works quite well. For example, every company should aspire to have org and department level dashboards where the whole company can track performance against key KPIs.
But what about the longtail of questions data teams are responsible for answering? If you’ve ever been part of a BI team, you’re very aware of the endless dashboard requests that flow into Slack, email, or whatever ticketing system you’ve implemented to manage the constant influx. Many data people have spent most of their careers fielding these urgent requests, most of which become dashboards that are looked at once and then forgotten.
Where dashboards go wrong
How can this be? There are many reasons dashboards are imperfect delivery mechanisms for the type of information business users need, but we’re going to focus on two key (and interrelated) issues.
There are just too many.
Dashboards have become the de facto delivery mechanism for all things data. A new business question leads to a new dashboard (or at minimum, throwing a new chart on an existing dashboard). You may be able to keep pace with the business for a little bit with this approach, but very quickly, you’ll have a bloated BI implementation that’s very difficult to maintain.
Every dashboard requires maintenance, meaning it needs to keep pace with changes in business logic and the underlying data. The more dashboards you have, the more impossible this task becomes, and you end up with a mix of trusted dashboards and a graveyard of outdated numbers.
Most things shouldn’t be a dashboard.
Ideally, dashboards should be action-oriented and help the business focus on a few key metrics and their drivers. I’ve spent the majority of my career studying, teaching, building and evangelizing the value of visualizing data for decision making. A well designed visualization can represent multiple dimensions of complex information in a single view and empower fast and informed decision making.
In reality, the answers to most business questions don’t require mapping multiple dimensions of information onto a single view, or analyzing a complex trend over time. The reason we rely on dashboards to share information isn’t because it’s the best tool for the job, but because it’s often the easiest, quickest way to share the output of queries. Business users don’t necessarily have the time or ability to access the data warehouse directly, so data teams will write the query and dump information in a chart somewhere in their BI tool. Tools like Tableau are incredibly powerful for highly custom data visualizations, but most of us use them as makeshift data access solutions.
Using dashboards this way is less than ideal for both data teams and business stakeholders. Business users need quick and reliable access to information, and they don’t want to dig through dashboards to find it. And even when data teams spend their time crafting beautiful dashboards, business users often download the output of charts into spreadsheets and plug data into their own bespoke workflows.
Solving the data access problem
At Preql, we want to help business users access trusted metrics quickly. Our user interface exposes clearly defined and governed metrics to business stakeholders. From there, you can integrate with spreadsheets or, when data visualization is absolutely necessary, you can build off of pre-defined metrics in the BI tool of your choosing. We’re a window on top of your data that’s easy for business users to access and manage, all without submitting a request to the data team.
The path to making data more trusted and accessible likely involves fewer, but more meaningful dashboards, and more investments in tools that meet business users where they are.
To learn more about how Preql can help your business access the information you need, sign up for a demo today.