Executive Analytics Platform

A firm with $4B under management ran executive reporting on Excel workbooks, emailed PDFs, and a legacy BI tool three people knew how to operate. Every Monday, a two-person analytics team spent six hours assembling a 30-page PDF for the C-suite. By the time executives read it, the data was stale.

The CEO had wanted "a dashboard" for two years. Previous attempts died because nobody could agree on what belonged on it — seven executives, seven sets of metrics, and an analytics team drowning in custom report variants.

I shadowed the analytics team through their Monday ritual, and the six hours of copy-paste turned out to be full of judgment calls: which source to trust when numbers conflicted, how to handle unreconciled data from the previous trading day, when a variance deserved a flag. Invisible in the final PDF, critical to its accuracy. Any replacement had to encode that judgment, not just automate the paste.

Interviews with all seven executives revealed the pattern that unlocked the design: despite receiving 30 pages, each executive read the same 3–4 metrics first. The CFO went to cash position. The CRO checked pipeline velocity. The 30-page report existed because nobody had the authority to say "you don't need pages 8 through 22." I also cataloged 47 follow-up questions over three months — 80% were "can you break this down by region / product / period?", exactly what a drill-down interface answers instantly.

The first design problem was political: one platform for seven executives with different mental models, without building seven dashboards. The answer was a shared data architecture with personalized entry points — each executive lands on a focus view with their metrics, and every path beyond it uses identical interaction patterns.

Card-sorting with each executive surfaced surprising overlap: the CFO and CRO cared about many of the same underlying numbers, just at different aggregations and time horizons. That validated the shared spine. From there: progressive disclosure — summary cards, sparklines on hover, full drill-down on click. I tested six chart types for the portfolio view alone before landing on small multiples.

The platform shipped with three core views: a personalized executive dashboard, a cross-functional metrics explorer, and an anomaly feed with plain-language explanations of what moved and why it might matter.

The biggest risk wasn't the interface, it was trust in the numbers. A beautiful dashboard showing inaccurate data dies in a week. I spent more time on validation rules and error states with the analytics team than on visual polish — delayed feeds and half-reconciled trading days happen every week, and the design had to tell the truth about them.

The adoption lesson: we ran a 4-week parallel period, and within two weeks five of seven executives had stopped opening the PDF. The holdouts converted after I added a weekly digest email replicating their favorite PDF sections with live links back to the dashboard. Meeting people where they are — even when that's their inbox — is sometimes the fastest path.