Case study
In production
Revenue dashboards, rebuilt as an app.
Three intermediate reporting tools — Coefficient, Sheets, Looker Studio — collapsed into a single Claude Code app that pulls directly from HubSpot. No more weekly manual sync.
- Build team
- Solo
- Intermediate tools collapsed
- 3→1
- Weekly manual syncs now
- 0
- In production
- Live
Context
The operating team needed sales and marketing measurement from scratch. HubSpot had been set up, but without the data structure, reporting layer, or shared metric definitions to actually run the funnel.
The job wasn't "build a dashboard." It was "define what we measure, build the pipes, make the numbers reliable, and put them somewhere the team can use."
The problem
The first version of the reporting layer put three intermediate tools between HubSpot and the team: Coefficient pulled data out, Google Sheets cleaned it, Looker Studio rendered the charts. It worked — but every metric update required a manual weekly sync across all three. Fragile, slow, time-tax every Monday.
That stack also fragmented ownership. Each tool held a slice of the logic, with no single place to look if a number moved.
What I built
I interviewed sales and marketing leads to define the metrics that actually drive decisions, then stood up the first-gen flow (HubSpot → Coefficient → Sheets → Looker Studio) so the team had numbers while I rebuilt it underneath.
The second-gen build is a single Claude Code app that pulls from HubSpot on its own, computes every metric in one place, and renders a dashboard with tabs for Marketing, Sales topline, Sales meetings, Opps, Stage, and Win rate. No weekly manual sync. The team opens the app, the numbers are current.
In the app
Click any screenshot to expand. Use ← → to navigate, Esc to close.
Marketing view — MCLs, MQLs, SQLs, and Meetings Booked at a glance, alongside the weekly trend and the full MCL → MQL → SQL → SAL funnel.
ICP fit distribution, acquisition channels, first conversion points, and meetings booked by rep — the breadth the team uses to tune the funnel each week.
Sales topline — SALs, meetings, open opportunities, win rate, and closed-won revenue. ICP fit on SALs tells ops what's actually qualified.
Where the revenue lands — qualified buyers converted to customers, closed-won revenue across projects and accounts.
Outcome
The three-tool intermediate pipeline is gone. Metrics live in one app, update automatically, and stay consistent across sales and marketing because they're defined once in the codebase — not copied across a spreadsheet and a BI tool.
The build was faster than traditional BI work because Claude Code was the primary tool the whole way — from prototyping the data model to shipping the production UI. The same loop would transfer directly to portfolio company operating work.