Datadog gives you raw RUM and a wall of widgets. What product and leadership needed was one place that showed our products as products—who's searching, where they're stuck, and whether demos turned into real use. So we built a product view on top of the same data.
We started with native Datadog dashboards: North Star Overview, Weekly Overview. They assumed more instrumentation than we had. In November 2025 we audited what RUM actually contained, retired the legacy dashboards, and shipped a minimal "Current Signals" dashboard that only showed what we could observe. Then we took the same idea offline: a Node script (build-live-report.mjs) that runs the search, audit, usage, and post-demo scripts, bakes the JSON into HTML, and writes reports/live-dashboard.html. One command, one static snapshot.
What made it shareable was moving that pipeline into a Netlify function. Every request runs the same fetchers, renders the same HTML, and serves it. No build step, no static site—just a serverless function (netlify/functions/dashboard.mjs) and a redirect so the whole site is the dashboard. We added cookie-based password protection, a five-minute cache so we don't hammer the Datadog API, and a Refresh button so stakeholders can pull fresh data on demand.
Refinements followed: product tabs, a link to the Datadog overview for drill-down. In parallel we built a full Next.js analytics app with date ranges, workflow filters, and user lists—deployed to its own Netlify site for deeper analysis. Two artifacts, one story: product metrics that start from what we actually observe.