A dashboard redesign focused on four friction points: empty states, data hierarchy, navigation, and cross-section consistency. No added features. Just clarity doing its job.
Mixpanel is a B2B analytics platform used by product teams to track behavior, revenue, and retention. The tool itself is powerful. The interface reflects years of accumulated decisions.
This redesign focuses on four places where that complexity creates real friction: empty states, data hierarchy, navigation, and cross-section consistency. No feature additions. No structural overhauls. Just the interface making better use of what already exists.
The audit covered the full dashboard experience, from first login to active use. Four distinct issues emerged, each affecting a different user type and a different part of the interface.
The dashboard looks broken before it's begun
New users land on a screen structured like a fully populated dashboard. Hollow card grids, competing sections, no clear first action. Nothing indicates what to do or what the end state looks like.
The primary KPI competes with everything around it
The MRR value (9.28M) sits in a three-column grid where supporting charts carry the same visual weight. Nothing establishes priority. The eye enters and has nowhere to land.
Browsing by type is a configuration step, not a default
100 unsorted results on every visit. Type filtering is pushed into a persistent right panel requiring deliberate setup. The page is labeled Discover but the interaction model is filter management.
Events, Users, and Lexicon feel like different products
Each view under Data follows completely different layout logic despite sitting under the same navigation parent. Different structures, different interaction patterns, no shared visual language.
Each finding pointed directly to a design decision. No exploration needed. The problems were specific enough that the direction was already implied.
Give new users one clear action
Replace the hollow grid with a focused setup state. One headline, one line of context, one button.
Establish hierarchy before effort is required
Structure carries the hierarchy first. Size and position do the work before color or styling enter.
Replace passive filtering with active defaults
Type filtering moves to the top as tab pills. Secondary filters stay accessible without being permanent.
One consistent template across data views
Events, Users, and Lexicon get the same layout shell. Same header, toolbar, table structure.
Five screens across the core dashboard experience. Each one maps directly to a finding from the audit. Mixpanel's original color language is preserved. The argument is structural, not aesthetic.
A skeleton preview behind one clear action
A skeleton preview of the populated dashboard sits behind a focused setup modal. One headline. One line of context. One action. The surrounding structure shows what the product becomes once data is connected.
Three tiers. One entry point.
One dominant KPI card, one supporting chart, three secondary charts in a subordinate row. Hierarchy is structural. Inline explanatory text replaced with contextual tooltips via an info icon.
Context and metrics grouped as one unit
Context text and metric cards wrapped inside a single container. A horizontal divider separates grouped metrics from deep-dive charts below.
Browsing first, filtering on demand
Type filter tabs at the top make category browsing the primary interaction. Secondary filters move behind an on-demand button. A page that actually helps you find things.
One template, three views
One consistent layout template across Events, Users, and Lexicon. A top-level tab bar makes the relationship explicit. Same header. Same toolbar. Same table structure.
Original on the left. Redesign on the right. Same product, different decisions.
Each screen change maps directly to a finding. Nothing was redesigned for its own sake. The argument throughout is structural: hierarchy, grouping, defaults, consistency. Mixpanel's color language and core interaction patterns are preserved.
The result isn't a reimagined product. It's the same product with less friction in the places where friction compounds most.
Scope & Limitations — This is a five-screen redesign based on heuristic analysis and live product audit of the demo environment. It doesn't account for Mixpanel's internal product goals, A/B testing data, or full user research. Findings are based on visual inspection, interaction testing, and established UX principles.
The most useful part of this project was learning to audit a product I didn't fully understand domain-wise. Mixpanel is built for analysts. I'm not one. But the hierarchy problems, the missing visual relationships, the defaults that don't serve the majority use case, those don't require domain expertise to identify. They require eyes and a frame. The audit framework was the unlock.
The other thing: Figma Make accelerated the UI phase significantly, but every useful output still required manual reasoning. The tool generates. The decisions still have to come from somewhere.