About this demo

What ProductPulse is, and how it's built

A read-only executive product-analytics demo built on synthetic data — every number on every screen is fictional, and every status is computed, never asserted.

Product context

ProductPulse is an executive product-analytics dashboard for a fictional B2B SaaS product. It tracks adoption, engagement, and revenue — and, critically, maintains an initiative registry that connects specific shipped work (AI features and standard product work alike) to the metric movement that followed. Most growth dashboards answer what happened; ProductPulse also answers what did we ship that caused it, and did it work?

ProductPulse is the natural second half of AgentOps: AgentOps asks “is this AI initiative safe to launch?” before it ships. ProductPulse asks “did it actually work?” after. Governance before, accountability after.

Architecture

Next.js App Router with TypeScript strict mode; every route is statically prerendered.
No backend, no database, no APIs — a fully typed synthetic dataset lives in src/data/.
Impact statuses come from a deterministic pure function in src/lib/impact.ts, covered by a vitest suite. Same inputs, same status, always — no AI calls, no scoring drift.
Tailwind CSS with shadcn/ui components and Recharts, deployed on Vercel from GitHub main.

Security posture

Read-only by design: no forms, no write paths, no authentication, no uploads, no cookies.
All data is synthetic and labeled as such in the UI.
No secrets, no API keys, no environment variables exposed to the browser.
A STRIDE threat model and security policy ship with the repo — see docs/threat-model.md and SECURITY.md.

Fischer Product Lab

One suite, one thesis: turn noisy, high-stakes reality into governed, decision-quality executive signal.

AgentOpsAI agent governance control tower — decides if an initiative is safe to launch.TrustDeskCustomer trust automation — security questionnaires with a deterministic confidence engine.VulnBoardExecutive vulnerability metrics — exposure as decision-quality signal.

Source for this demo: github.com/Fischer-Product-Lab/productpulse

Glossary

Every metric on these screens, in plain language. The same definitions power the tooltips on the stat cards.

DAU: Daily Active Users — unique users who took a meaningful action on a given day.
WAU: Weekly Active Users — unique users active in the last 7 days.
MAU: Monthly Active Users — unique users active in the last 30 days.
Stickiness: DAU ÷ MAU — the share of monthly active users who show up on a given day; a measure of how habitual the product is.
Activation rate: The share of signed-up accounts that reached their first core workflow.
Habit rate: The share of activated accounts that made the product part of their regular routine.
Adoption lift: Percentage-point change in adoption of the relevant workflow, from the pre-launch baseline to today.
MRR: Monthly Recurring Revenue — subscription revenue normalized to one month.
Net new MRR: The total change in Monthly Recurring Revenue this month: new + expansion − contraction − churned.
NRR: Net Revenue Retention — recurring revenue kept from the existing customer base after expansion, contraction, and churn. Above 100% means the base grows even with zero new sales.
AI cost savings: Hours of manual work eliminated by AI initiatives, valued at a blended fully-loaded hourly rate ($95/hr in this demo).
Impact status: Deterministic rating of every shipped initiative: Strong Impact (adoption lift ≥25 pts plus a measurable business outcome), Underperforming (lift under 5 pts), Monitor (everything between). Computed by a pure function — never asserted.
Cohort retention: Of the users who activated in a given month, the share still active N months later. Each row of the cohort table follows one activation class through time.
Churn risk: Deterministic per-segment rating from usage signals: High (inactive 30+ days, or under 0.5 sessions/week while inactive 14+ days), Elevated (under 2 sessions/week, inactive 7+ days, or seat utilization under 30%), Low otherwise.
Seat utilization: The share of purchased seats that are actively used across a segment's accounts.
Attribution caveat: A flag raised when another initiative launched within six weeks of this one — adoption movement inside that shared window may belong to either launch, so neither can claim it cleanly.