PMFKit v1.7: From One-Off Diagnostics to Strategic PMF Tracking

February 15, 2026 · 15 min read

Today we're launching PMFKit v1.7, a complete transformation from a one-off diagnostic tool into a comprehensive Product-Market Fit tracking platform. This release represents the culmination of a 2-week development sprint that touched 156 files, added 16,787 lines of code, and introduced enterprise-grade infrastructure while maintaining our core principle: see value before we ask for anything.

The Problem We Solved

Before v1.7: Ephemeral Diagnostics

PMFKit was powerful but ephemeral. Users would paste a URL, wait 8-12 minutes for analysis, receive comprehensive PMF diagnosis, and then lose access to it. There was no way to save projects, track changes over time, or compare analyses.

For a tool designed to help founders make strategic decisions about Product-Market Fit, this was unacceptable.

After v1.7: Strategic Tracking Platform

Now you can save unlimited projects (free tier), track multiple analyses per project, view complete history with timestamps, rerun analyses with one click, and export everything as PDF or JSON.

New Features

1. Multi-Project Dashboard

The new dashboard is your central hub. See all projects at a glance with names, URLs, run counts, and latest DRL scores. Load time: 80-150ms (was 400-600ms before optimization).

User flow: Dashboard → View all projects → Click project → See all runs → Click "View" for full report → Click "Rerun" for fresh analysis → Click "Export" for PDF/JSON.

2. One-Click Rerun

Every analysis now has a "Rerun" button. Click it to start a fresh analysis with the same URL and context. Perfect for monthly PMF health checks, quarterly investor updates, or tracking improvement over time.

3. Analysis Runs System

Every analysis is tracked as a "run" within a project. See your progression over time:

• Run #1 (Jan 15) — DRL 92/100 (L2)
• Run #2 (Feb 1) — DRL 94/100 (L2)
• Run #3 (Feb 15) — DRL 96/100 (L2) ← Improvement!

Each run includes timestamp, DRL score, analysis mode, and pipeline version (which algorithm generated it).

4. Component-Level Feedback

Give thumbs up/down on individual report sections (Brand, Product, Pricing, Market, Distribution). This helps us improve quality based on real user feedback.

Within 24 hours of soft launch, we received 47 feedback events and identified that pricing analysis needs improvement (lowest thumbs-up rate) while brand analysis is working well (highest thumbs-up rate).

Architecture Evolution

Database Schema Transformation

We evolved from a flat structure to a hierarchical, normalized schema with proper relationships, indexes, and Row Level Security (RLS) policies on every table.

10 database migrations were carefully staged over 2 weeks. Each migration was idempotent (safe to re-run), reversible, tested on staging first, and transaction-wrapped.

Zero Downtime Migration

We had 1000+ existing analyses in production. We couldn't break them. Our solution: staged migrations with lazy backfill.

Instead of migrating all data at once (slow, risky), we backfill each user's data when they first visit the dashboard. Their historical analyses are automatically organized into projects. Zero downtime. Zero data loss.

Performance Improvements

75% Faster Dashboard

Before: 400-600ms
After: 80-150ms
How: Eliminated N+1 queries, added strategic indexes, implemented caching, and used server components for data fetching.

Minimal Bundle Impact

Despite adding 11 components, 8 APIs, and massive functionality, bundle size increased by only 12KB (+6%). We achieved this through code splitting, lazy loading, and tree shaking.

Quality Assurance

Every commit passes TypeScript strict mode, golden baseline regression (50+ test products), smoke tests, RLS verification, and consistency checks. We wrote 13 new developer scripts that automate database health checks, RLS verification, data consistency, and performance testing.

Philosophy: Prevention > debugging. Quality gates enable velocity.

Lessons Learned

1. Boring Technology Wins

We used standard PostgreSQL, normal Next.js patterns, boring migrations, and simple auth. Result: Shipped in 2 weeks instead of 2 months. Choose boring technology. Save creativity for product, not infrastructure.

2. Quality Gates Enable Velocity

Catching bugs at commit time is 10x faster than debugging in production. No emergency rollbacks. No hotfixes. Prevention > debugging.

3. Scripts Compound

We wrote 13 scripts (~8 hours investment) and saved 15+ hours in this sprint alone. Over a year? 100x ROI. Automate everything you do more than twice.

4. Version Aggressively

We shipped 10 versions in 2 weeks (v1.6.1 → v1.7). Each version was a checkpoint, revertable state, and testable milestone. When v1.6.4 had N+1 query bugs, we reverted to v1.6.3 while fixing.

5. Users Want Reliability > Features

We could have shipped report comparison, trend charts, team features, and API access. Instead, we focused on solid persistence, fast queries, data integrity, and reliable backups. Trust first. Excitement second.

What's Next

Immediate (Next 2 Weeks)

• Report Comparison UI: Side-by-side diff view highlighting what changed
• Trend Charts: DRL over time, component scores, traffic estimates
• Mobile UX: Better dashboard on mobile, responsive layout, touch controls

Near-Term (Next Month)

• Team Collaboration: Share projects, comment threads, @mentions
• API Access: RESTful API, webhooks, rate limiting
• Advanced Analytics: Cohort analysis, conversion funnel, accuracy trends

Try It Now

The Numbers

Conclusion

PMFKit v1.7 transforms a one-off diagnostic tool into a comprehensive Product-Market Fit tracking platform. We built multi-project dashboard, historical tracking, one-click rerun, component feedback, and enterprise data architecture—all while achieving 75% performance improvement.

The vision: PMFKit is becoming the system of record for Product-Market Fit. Not just "get a diagnosis once," but "track your PMF journey continuously" from idea → traction → PMF → scale. With evidence. With confidence. With clarity.