Replacing Power BI with a live platform for digital asset yield

Founding Product Designer · Product, Design Systems

Problem

Yield positions across CeFi and DeFi venues were reported through Microsoft Power BI. The team had no live view of loans and collateral, and investors had no way to see how their money was being managed without asking.

Outcome

One live dashboard for the team that analyzes and manages the funds, with positions, risk and venue exposure in a single view. Investors can sign in to see how their money is managed, so reporting stops being a manual job.

Rebuilding the dashboard around the analyst

The old dashboard was a Power BI report: dense tables, no hierarchy, and nothing to say what had changed. We rebuilt it around the questions analysts ask every day, so the answer is on screen before they go looking for it.

Rebuilding the dashboard around the analyst

The old dashboard was a Power BI report: dense tables, no hierarchy, and nothing to say what had changed. We rebuilt it around the questions analysts ask every day, so the answer is on screen before they go looking for it.

Rebuilding the dashboard around the analyst

The old dashboard was a Power BI report: dense tables, no hierarchy, and nothing to say what had changed. We rebuilt it around the questions analysts ask every day, so the answer is on screen before they go looking for it.

After
Before

Putting what matters first

Every card answers one question: what needs attention, where the risk sits, how funds are spread across venues, and what changed recently. Distinct concepts get their own card, never a divider inside a shared one.

Putting what matters first

Every card answers one question: what needs attention, where the risk sits, how funds are spread across venues, and what changed recently. Distinct concepts get their own card, never a divider inside a shared one.

Putting what matters first

Every card answers one question: what needs attention, where the risk sits, how funds are spread across venues, and what changed recently. Distinct concepts get their own card, never a divider inside a shared one.

Needs attention

Exceptions only. Alerts and status colours appear when something is actually wrong, so a quiet card means a healthy portfolio, not an empty one.

Risk

Collateral health shown as a gauge with clear thresholds, so analysts can read how close a position is to its limit at a glance instead of calculating it.

By venue

Exposure broken down by venue across CeFi and DeFi, with APY, balance and status on every row, so concentration and stale data stand out.

Context where it’s needed

InfoTips sit beside every metric that needs explaining, defining terms like APY and collateral ratio without sending anyone to a separate document.

Where the old dashboard fell short

The existing dashboard showed everything and explained nothing. Every number carried the same weight, data lagged behind the venues, and spotting a problem depended on already knowing where to look.

Where the old dashboard fell short

The existing dashboard showed everything and explained nothing. Every number carried the same weight, data lagged behind the venues, and spotting a problem depended on already knowing where to look.

Where the old dashboard fell short

The existing dashboard showed everything and explained nothing. Every number carried the same weight, data lagged behind the venues, and spotting a problem depended on already knowing where to look.

The challenges we set out to solve

Working with the analysts surfaced three recurring needs. Each one shaped a part of the final product, from what loads first to how every number is explained.

The challenges we set out to solve

Working with the analysts surfaced three recurring needs. Each one shaped a part of the final product, from what loads first to how every number is explained.

The challenges we set out to solve

Working with the analysts surfaced three recurring needs. Each one shaped a part of the final product, from what loads first to how every number is explained.

Insight 01

Everything in one place

Positions, risk and venue data lived on separate report pages. Analysts needed one view that covered the daily check-in without clicking through anything else.

Insight 02

Risk is the main focus

Most of an analyst’s day went to watching collateral health and catching positions drifting toward risk, so that had to be the first thing on screen.

Insight 03

Not every number explains itself

Terms like blended APY and collateral ratio meant different things to different people. Every metric needed a plain-language definition right where it appears.

Making dense financial data feel calm

Financial data is dense by nature. Rather than hiding it, we gave it structure: a clear reading order, colour used only where it means something, and nothing on screen that doesn’t earn its place.

Making dense financial data feel calm

Financial data is dense by nature. Rather than hiding it, we gave it structure: a clear reading order, colour used only where it means something, and nothing on screen that doesn’t earn its place.

Making dense financial data feel calm

Financial data is dense by nature. Rather than hiding it, we gave it structure: a clear reading order, colour used only where it means something, and nothing on screen that doesn’t earn its place.

A new experience, built on old instincts

Analysts check this daily, with money on the line. Making them relearn where things live would cost speed and trust, so Power BI’s core concepts stayed, just clearer.

Built for a single glance

Analysts scan before they read. Following the eye’s natural path, key numbers and alerts sit above the fold, so the daily check-in takes a glance, not a scroll.

Colour that leads the eye

The eye catches contrast before it reads a word. A neutral interface lets amber and red stand out on their own, pulling attention to what needs action first.

Testing with analysts and iterating

We put prototypes in front of the analysts who would use the platform every day. One finding changed the plan: the dashboard couldn’t be a dead end, because the numbers still had to leave it.

Testing with analysts and iterating

We put prototypes in front of the analysts who would use the platform every day. One finding changed the plan: the dashboard couldn’t be a dead end, because the numbers still had to leave it.

Testing with analysts and iterating

We put prototypes in front of the analysts who would use the platform every day. One finding changed the plan: the dashboard couldn’t be a dead end, because the numbers still had to leave it.

Investors still need records

The team was still fielding investor requests for exports to support audits. A live dashboard didn’t remove that need, and without a clean export, they’d be back to pulling numbers out of Power BI by hand.

The revision: Export, built into the header

Export sits beside the date range, so every report matches the period on screen. One click, no rebuilding, and the dashboard stays the single source of truth.

Handing off a system, not just screens

The final designs shipped with a component library and user stories behind them, so engineering built from the same source as design and new views could be assembled from what already existed.

Handing off a system, not just screens

The final designs shipped with a component library and user stories behind them, so engineering built from the same source as design and new views could be assembled from what already existed.

Handing off a system, not just screens

The final designs shipped with a component library and user stories behind them, so engineering built from the same source as design and new views could be assembled from what already existed.

Figma design system + user stories

Cards, stat strips, gauges and InfoTips built as components with every variant and state, bound to shared colour and spacing tokens. User stories tied each screen to the job it supports.

Development ready

Hover, pressed and focus states documented in one approved sheet, alongside spacing and token specs, so nothing was left to interpretation in the build.

Learnings

Deciding who the product was for changed more than any visual choice. Once we committed to one primary user, the hierarchy, the card structure and what counted as an exception all followed. Next time I’d lock that decision even earlier.

Learnings

Deciding who the product was for changed more than any visual choice. Once we committed to one primary user, the hierarchy, the card structure and what counted as an exception all followed. Next time I’d lock that decision even earlier.

Learnings

Deciding who the product was for changed more than any visual choice. Once we committed to one primary user, the hierarchy, the card structure and what counted as an exception all followed. Next time I’d lock that decision even earlier.