Home About
Dealer Permissions AI Assistant Analytics and Reporting Restructuring Data Model Finance Calculator Marketplace Checkout
HOME ABOUT

Analytics and Reporting

Evora Global

Role
Product Designer, then Senior Product Designer
Year
2023 to 2025
Team
UX researchers, product managers, engineers and reporting consultants
Tags
Data visualisation, Data quality, Workflow design, B2B SaaS
Glass office towers seen from street level.

Context

The platform stored ESG data well but interpreted it poorly. Fund managers could see figures, but not answer the questions that mattered: are we on track, which asset is the problem, can I trust this number? Reporting lived outside the product entirely, drafted in slides, signed off by email and shared as PDFs that drifted out of sync with the live data.

Approach

I led design for both halves, analytics first and reporting second, because a report is only as good as the data under it. Interviews with fund managers, consultants and analysts were affinity-mapped into six problem areas. Rather than run six workstreams, I sequenced them into four phases by dependency.

For analytics, that meant drill-downs by scope, sector and asset, and a new data quality module. For reporting, I modelled the workflow before the screens: every report gets a state, a named reviewer and an audit trail.

Key decisions

Coverage and completeness are shown apart, because a meter can report every month and still be mostly estimated, and one blended score hid that. Quality sits beside every number, so each figure shows how much of it is actual, estimated or missing. Approval works as a lifecycle, moving from draft to in review to approved or rejected, with a named reviewer at each step. One thing we cut was one-click report generation: the AI writes a draft, a consultant edits it and a senior approves it.

Outcome

Managers could spot a problem, trace it to the meter behind it and see how far to trust the figure, without leaving the platform. Reports became tracked items with a state, a named reviewer and a visible history. Approved reports could be shared as a PDF, a revocable link or straight into Power BI.

The platform moved from storing data to answering questions about it, and reporting moved out of email threads and into the product.

Full process

Problem

The platform stored ESG data well but interpreted it poorly, and the reports built from that data lived outside it.

Where it hurt

Users exported raw figures into Power BI and spreadsheets to make sense of them. Reports were drafted in PowerPoint from Excel exports, signed off over email and shared as PDFs that drifted out of sync with the live data.

Who felt it

Fund managers and asset owners who needed to know what was wrong, the analysts maintaining the data behind them, and reporting consultants running several reports across frameworks with no shared status or named reviewer.

The end goal

Data people could act on, data quality they could see, and reports as first-class objects with a state, an owner and an audit trail.

The old dashboard: figures with no direction.
What fund managers inherited: data without direction.
The old reports area: a list of uploaded files.
A report was a file, with no state, owner or trail.

Research

I ran structured interviews and review sessions with fund managers, ESG managers and sustainability consultants, and shadowed live report cycles from brief to disclosure. Every session was tagged in Dovetail but listened back to in full, because the hedges that shape a real insight rarely survive a summary.

Mapping the investigation journey

I mapped the experience across user types in four phases: spotting a problem, analysing it, drilling down, and what comes next. Every note on the map is something a user needed and couldn't do.

An experience map of sticky notes under four phases: spotting problems, analysing data, drilling down and future areas.
Experience map: what users needed at each phase of an investigation.

“I'm looking at a wall of numbers with nothing telling me where to start.”

Asset manager, customer fund

The swim-lane showed what happened when a consultant spotted a problem. A simple data question became a chain of escalations across three roles, with a spreadsheet at every broken step.

A swim-lane across consultant, fund manager and data team, with high-friction steps highlighted.
Where a simple query became an escalation. High-friction steps are in orange.

“I have to ping three people on Slack to find out if this report shipped. And then I still have to open the PDF to check the version.”

Reporting consultant, customer fund

Discovery

The interviews were affinity-mapped into root-cause categories, and each one was reframed as a How Might We to set the brief with product and engineering. Six categories became four phases, sequenced on dependency rather than on how loud each complaint was: performance first, then dashboards, then data quality, then site-wide fixes.

A roadmap of sticky notes: columns for home page, performance, dashboard, data quality and site-wide, rows for four phases.
Roadmap: each column a data domain, each row a phase.

The report as an object, drawn before any wireframe

For reporting, I modelled the workflow before the screens. The flowchart traced every path a consultant or the reporting team took, from opening the reports section to disclosure, and showed where a report left the platform for Microsoft tools or Power BI and came back as a file. Draft, in review, approved and rejected became states with a named reviewer, and that lifecycle was the contract every screen had to honour.

A flowchart from logging in to disclosing a report, through view, edit, download, upload, format checks and final checks.
Reporting flow: every path from the reports section to disclosure.

One metric set, every framework

GRESB, EPRA, INREV and SFDR ask for the same numbers in different shapes. A report type per framework would have rebuilt the spreadsheet problem inside the platform, so the metric set was defined once and every framework reads from it. The AI drafts a first pass from those validated numbers; the consultant edits, decides and signs off, and every edit is logged.