Jobsly v0.1 had a beautiful Kanban board. Columns for every stage. Color-coded cards. Drag-and-drop. It looked great in demos. And nobody used it.
Not because the UI was bad — it was polished. The problem was deeper: a dashboard assumes that the user’s job is to manage a process. But hiring managers don’t want to manage a process. They want to find the right person.
Killing the dashboard was one of the hardest product decisions we’ve made. It was also one of the best. This is the story of why we built it, why it failed, and what we learned about the difference between feeling productive and being productive.
Why we built it in the first place
Every hiring tool has a dashboard. It’s table stakes. When we started building Jobsly, we didn’t question it. Of course there would be a pipeline view. Of course there would be stages. Of course candidates would be cards you could drag between columns.
We spent three months perfecting it. Custom stages, configurable columns, batch actions, filters, keyboard shortcuts. It was genuinely well-built. Our early beta users said it looked great. Investors loved the demo. And then we watched people actually try to use it.
The average hiring manager opened the dashboard once, glanced at it, and then asked Rob for a summary instead. The dashboard was a beautiful answer to a question nobody was asking.
Dashboards are anti-solutions
A dashboard gives you the illusion of control without the substance. You can see all your candidates, move them between columns, add notes. But none of that gets you closer to a hire. It just makes you feel productive.
The real work of hiring — evaluating fit, making decisions, having honest conversations — happens outside the dashboard. The dashboard is just bookkeeping.
There’s a deeper problem too. Dashboards create busywork that masquerades as work. Moving a card from “Phone Screen” to “Onsite” feels like progress. But it isn’t. It’s data entry. The progress happened when you decided to move the candidate forward. The card is just the receipt.
We tracked dashboard interactions for two months. The average user spent 12 minutes per session in the dashboard. Of those 12 minutes, 9 were spent on status updates, reorganizing columns, and reading information they already knew. Only 3 minutes involved an actual decision. That’s a 75% overhead rate.
The conversation as interface
Once we accepted that the dashboard wasn’t working, we had to answer a harder question: what should replace it? The answer was already in our data. Users were bypassing the dashboard and talking to Rob instead.
They’d ask: “Who should I interview next?” “What’s the status of the design role?” “Are there any candidates I need to make a decision on?” Rob already had all the pipeline data. The dashboard was just a visual intermediary — one that added friction without adding value.
We realized the conversation itself could be the interface. Instead of the user navigating to information, the information comes to the user. Rob proactively surfaces what needs attention: candidates ready for a decision, interviews to schedule, offers to extend. The user’s job shifts from “manage the process” to “make decisions when prompted.”
What replaced it
We replaced the Kanban with Rob. Instead of showing you a board and letting you manage your own process, Rob tells you what needs to happen next. “This candidate scored 82. Here’s why. Want to schedule an interview?”
The interaction model flipped from “you manage the pipeline” to “Rob manages the pipeline and asks you for decisions.” Usage went up. Time-to-hire went down. Nobody missed the Kanban.
We did keep one visual element: a simple list view that shows all active candidates with their current score and stage. No drag-and-drop, no columns, no batch actions. Just a read-only view for the moments when you want to see everything at once. It’s used by about 20% of our users, and they spend less than two minutes per visit.
The lesson
The lesson isn’t that dashboards are always bad. They work for operations teams managing hundreds of items, for data analysts exploring datasets, for project managers coordinating complex workflows. Dashboards are great when the user’s job is to manage a process.
But hiring managers aren’t process managers. They’re decision-makers. And decision-makers need context, recommendations, and clear action items — not a board full of cards.
If you’re building a product, ask yourself: is my user’s job to manage a process, or to make decisions? If it’s the latter, the dashboard might be the thing standing between them and their actual work.