Reducing user friction by 39% through clearer data visualization
Transformed a confusing analytics experience into a system users could reason about, clarifying data scope and logic so analysts could trust and act on what they saw.
Blackbird’s Constellation platform helps intelligence teams analyze narratives and signals across large datasets.
Two core surfaces — Analyze and Network Graph — displayed different Twitter metrics derived from different data scopes (Local vs Global data).
This created confusion even among experienced analysts and led to repeated support tickets.
Users assumed both views used the same dataset. In reality, the Network Graph only used Local Data — creating a hidden mismatch.
Users assumed both pages represented the same dataset. When numbers didn’t match, they questioned platform reliability. The risk wasn’t only UX friction — it was trust in the intelligence output.
I led the design effort in partnership with a PM and Tech Lead, owning:
- Problem framing
- Research direction
- UX strategy
- Solution design across both surfaces
The problem wasn’t just a “data mismatch.” It was a mental model mismatch.Even trained analysts struggled to distinguish:
- Local data (project-scoped)
- Global data (platform-wide signals)
If experts were confused, external users would be even more.
1) Align mental models before aligning interfaces
Instead of “fixing visuals,” I focused on making data logic legible to users.
2) Reduce cognitive load through structure
On the Analyze page, I:
- Reorganized data into Posts, Engagements, and Signals
- Introduced clearer naming aligned with Twitter conventions
- Added visible data counts to set expectations
This made the system feel more predictable. The data organization changed from this:
To this:
3) Make the Network Graph educational, not just visual
We added a data preview layer before graph generation.
This:
- Clarified what would be visualized
- Showed how Analyze data translated into Network data
- Reduced back-and-forth navigation
The graph became less of a “black box.”
Quantitative
- 14% reduction in support tickets after month 1
- 39% reduction after month 2
Qualitative
- Increased user trust in the platform’s data
- Reduced ambiguity around data interpretation
- Helped analysts move from confusion to confidence when building narratives
The solution didn’t just reduce frustration — it improved how users reasoned about intelligence data.
This project reinforced that in data-heavy platforms, clarity of logic is as important as clarity of UI. Designing for understanding, not just usability, can directly affect product trust and support burden.