Blog
Why data apps confused the hell out of me (and probably confuse you too)
Dashboards show you what happened. Data apps let you act on it.
When I joined Hex as a Product Evangelist, one of my first questions was: "What is a data app and how is it different from a dashboard?"
The answer I got was... nebulous. "They're the same but different." Not super helpful when you're trying to explain this to anyone curious about why Hex doesn't have traditional dashboards in their product.
After a week of listening to sales calls, I kept hearing the same confusion from prospects. Analytics leaders would ask variations of: "So how is this different from Tableau?" or "Is this just another dashboard tool?" The question made sense — we've all been trained to think about data delivery in terms of dashboards. It's the universal vocabulary. Every company has them. Every analytics team builds them.
But something fundamental is shifting, and the old vocabulary doesn't quite capture it anymore. I needed to understand this for myself before I could explain it to anyone else. Hex defines our data apps as “…a curated selection of cells from your project that you can publish and share with stakeholders.” But not to be rude…isn’t that a dashboard?
This is my attempt to work through that thinking, because if it confused me — someone who's built dashboards professionally for over a decade — it probably resonates with some of you too.
My dashboard education
I need to tell you where I'm coming from, because my confusion about data apps vs. dashboards came from a deep love for dashboards themselves.
My data career started in 2012 and spanned several tools in the BI world. I built dashboards in Tableau, Power BI, and Qlik by virtue of my various analysts roles and consulting stints.
I became really good at dashboards. I mastered the tools. I could build almost anything stakeholders asked for. I was genuinely proud of the work.
And yet, I kept running into the same walls.
Let me share two experiences that's stuck with me
I built a dashboard for a children's hospital marketing team tracking website attribution. The GA data was challenging to work with but we made it work. The team was thrilled. They finally had visibility into conversion rates and visit patterns they'd never seen before.
Then came the natural next question: "This is great. Can we explore user journeys? Can we click into specific visitors and see what paths they took through the site?"
I tried. I hacked together a Sankey diagram showing a few journey steps. But they couldn't DO anything with it. They couldn't click into a journey segment to see the underlying users just filter by page URLs.
This wasn't unique to that project. I heard variations of it constantly:
"Can I click here and explore this on my own?"
"Why can't I filter by product sub-sub-sub categories? You've only given me the top two levels."
"Can you just export the raw data so I can pivot it myself in Excel?"
That same hospital's marketing team had inconsistent UTM parameters everywhere — "email", "Email", "e-mail", "emale". Campaign structures with launch dates sometimes included, sometimes not, sometimes camel-cased, sometimes delimited by dashes. A mess.
I thought: what if I build them something where they can construct UTM parameters from dropdown lists? Consistent inputs, clean data going forward.
It was fine but the UI felt clunky and incomplete. It kind of worked, but it wasn't really solving the problem. The data lived in Tableau I needed it to feed other systems.
What I needed — though I didn't have the language for it then — was an app.
When I finally understood "data apps"
So when I joined Hex and asked "What's a data app vs. a dashboard?", I was bringing all of this history with me. The nebulous answer frustrated me because I needed clarity to do my job.
Here's what I've figured out: a data app is what I always wished dashboards could be.
It sounds simple, but the implications are significant. Let me show you what I mean with an example:
A sales person opens their Hex app and inputs a new lead with forecasted revenue. That information writes directly to the underlying tables, which connect to Salesforce — one point of entry. In that same app, their pipeline report updates immediately, showing how this new lead affects their attainment percentage.
One interface. One workflow. Immediate feedback.
This isn't easily achievable in traditional BI tools without significant custom development. The salesperson has to use 2-3 different systems to accomplish this task — something they likely do multiple times per day.
Hex’s data apps allow dashboard owners to break constraints we’ve all experienced.
Dashboards | Data Apps |
|---|---|
Exploration boundaries | Deeper exploration within boundaries |
Users could drill down and filter — but only where I, the designer, had allowed it weeks ago. | Through features like Hex's Explore mode, end users can add groupings, change breakdowns, and answer follow-up questions — exploring within the governed data model. |
The freshness problem | Data when you need it |
Live connections existed, but they came with significant trade-offs in performance and user experience at scale. | Data apps use a smarter connection, refreshing when users need it, then caching results for quick subsequent interactions. |
The workflow gap | Embedded workflows |
Dashboards showed you data. They didn't let you easily act on it. Want to update a forecast? You needed a different tool. | Apps are designed to accept inputs, write back to databases, trigger processes, and enable complete workflows as part of their core functionality. |
Tool sprawl | End-to-end in one platform |
Modern BI tools build great dashboards but trap you the moment you need deep SQL or Python analysis — forcing constant context-switching that kills momentum. | With Hex, SQL, Python, visualizations, and app components live in the same notebook. No exporting between tools, just build and publish when ready. |
Dashboards let you look. Data apps let you do.
You can read about the difference all day, but nothing beats experiencing it yourself. We built a Customer Account 360 app that's open for you to explore so please get your hands dirty.
Honestly? Five minutes playing around with it will probably tell you more than the rest of this article.
The mental model shift
Hex’s data apps require you to think a bit differently than dashboards.
Dashboards are linear — extract, transform, visualize, publish. There's a clear sequence.
Apps are compositional — one query feeds another, which creates parameters for a Python model, which updates a visualization that can trigger new inputs. It's building blocks that interact with each other, not an assembly line.
AI is making this shift happen much faster. You don't have to build all of this manually anymore. Hex's Notebook Agent can scaffold this entire workflow from a simple prompt. Describe what you want — "Create an app that lets sales reps input new leads and see updated pipeline metrics" — and the agent creates the cells, connects them, and gets you to a working first draft in minutes. You can then refine, adjust, and publish.
Where I landed
After working through all this, here's my answer to "What's a data app vs. a dashboard?":
A dashboard presents information for consumption. A data app embeds information alongside workflows to drive action.
The reason "data apps vs. dashboards" is so hard to explain is that apps do everything dashboards do — they just do more. It's not a clean either/or comparison. It's more like: data apps are what dashboards would have been if they'd been designed for today's data infrastructure instead of 20 years ago.
Both can be beautiful. Both can be well-designed. Both can display the same charts and metrics. The difference is in what happens next. Can users act on what they see? Can they explore their follow-up questions? Can they complete their work without leaving for another tool?
This shift in thinking takes time. But I'd encourage you to experiment. The next time you're scoping out a "dashboard" project, ask yourself: would this be better as an app? Could the users benefit from deeper exploration? Is there a workflow we could enable, not just information to display?
The answer won't always be yes. But asking the question is worth it. If you want more examples, see how ClickUp, StubHub, and Om1 all used data apps to help their stakeholders act on decisions.