Stop Shipping Dashboards That Don’t Matter

Three Questions to Ask Before You Build That Dashboard

If you work in data, you’ve probably lived this story:

A stakeholder pings you: “Can you build us a dashboard?”

Maybe you they give you a requirements doc. Maybe you hop on a call to hash out the details.

You spend a month pulling data and building the dashboard.

You ship it. They thank you.

Then… silence.

A month later, the usage metrics show no one’s touched it since demo day.

It’s deflating. Not because you want applause, but because we want our work to matter. We want it to drive decisions, not sit in a dashboard graveyard. We’ve likely all had this happen.

Over time, I’ve learned a few small mindset shifts and questions that make a difference in how people view your dashboard.

Whether you’re delivering a dashboard or deploying a fraud detection model, it starts with a better understanding of why your stakeholder is even asking for the dashboard. But just asking why is not enough, so let’s talk about how you can build data products that are actually helpful.

Ask These Three Questions

One of the most common pieces of advice you’ll hear in data: Push back and ask why more often.

It’s good advice, but incomplete.

In practice, asking why doesn’t always get you the insight you need. Sometimes, the answer is just, “My boss's boss asked, and I’m not about to tell them I don’t know.”

That might be fine for quick ad-hoc asks and if you’re team wants to be reactive instead of proactive.

But if you’re about to spend the next few weeks building pipelines to support their dashboard, it’s worth going deeper.

At a minimum, ask these three questions:

1. What decision are you trying to make (or action are you trying to take)?

Plenty of dashboards track vanity metrics, or worse, fluff. Numbers that look interesting but don’t lead anywhere. If no one can act on the data, why did we build the dashboard in the first place?

Start by asking: What decision are you hoping to make with this?

Or what action will this help you take?

You want to build for decisions, not just data. Here’s are a few subtle examples:

Good examples - Metrics presented in a way that that could lead to a decision:

• Customer acquisition cost by channel → Helps decide whether to increase spend on Meta or Google Ads next month. Of course then you’ll likely have further questions like which cohorts from said channels are more likely to respond?

• Churn rate by customer segment → This can provide so much information. In fact I was recently having a conversation with a data leader who discussed a past project where they found that although churn was starting to increase, when they dug down into the metrics they found there was a specific segment that was actually causing the increase. In turn, they were able to build a plan and experiment on how to reduce it for just that group.

Less effective - Metrics without a clear decision:

• Page views over time (without segmentation or context) → Interesting, but what do you do with that? So your page views are going up, are people converting? Are they coming from specific areas and why?

• Revenue from the last 12 months → Good for historical context, but not actionable unless tied to forecasting or more granular information. Great, your revenue is going up, will it keep going up, can you invest more money some where to make it go up more or at least not drop?