The next time you’re starting to work on a Power BI report, think of Healthcare.gov.

Yup, you read that right. Healthcare.gov, the website designed to help people across the United States sign up for health insurance after the passage of Obamacare.

If you are an American, you probably remember when it first launched and it failed spectacularly. (heck, I’m Canadian and I know about it). Healthcare.gov crashed repeatedly in its first days/weeks of existence.

This post is NOT a post about Obamacare, or about the American style of healthcare. This post is about trying to do too much.

There is a LOT out there about what went wrong with Healthcare.gov, from bad product management to tight timelines to inefficient contractors…

Recently on the Freakonomics podcast, I heard Jennifer Pahlka speaking about it. Pahlka founded Code for America (and wrote the book [Re]Coding <America/>) and was one of the people President Obama brought in to fix Healthcare.gov, and also helped found the U.S. Digital Service.

Jennifer said the following (and more):

“Product management is the art of deciding what to do.

HealthCare.gov just tried to do all the things.

It didn’t have somebody who was empowered to say: “I don’t think we can launch a system this complex, that handles this many edge cases, and have it work for everyone on Day One.”

…we need to have that idea of what is it that we’re actually deciding to do, and then empower someone to make those choices instead of say, Here are literally thousands of requirements, and have them all work the first day that the site launches.

I mean, it’s in retrospect insane.“

The beautiful thing about Power BI (and dashboards made in any software) is that we can build dashboards that are all things to all people.

But we should NEVER do that.

If you want a dashboard that people actually USE, it has to be focused and purpose-driven.

There needs to be questions that have to be answered. It can NOT be just something that throws all the data in an organization into a reporting tool that your users have to figure out.

You ARE allowed to make more than one dashboard with different goals, and you NEED those goals BEFORE you start building.

You can look at the data available … but before you put time and effort into building something, step away from your computer and go talk to the people who will be using what you will be building.

Those people do NOT need to examine every piece of data in your organization. They need certain insights and metrics and trends to do their job or to report to their boss, their board, or the public. That’s it. Full stop.

Don’t build the (first version of) Healthcare.gov of dashboards. Focus that sucker. Define a SCOPE and ROCK the HELL out of that scope.

Your dashboard users will love you rocking that scope. They’ll get a purpose-driven dashboard they can use easily.