Case Study: How One Neighborhood Directory Cut TTFB by 60% and Doubled Engagement

Case Study: How One Neighborhood Directory Cut TTFB by 60% and Doubled Engagement

Hook: Faster pages mean more eyes and bookings. In this case study we document the exact technical changes and product tradeoffs that reduced TTFB, improved Core Web Vitals, and doubled engagement in three months.

Starting point — the problem

A community directory with 200k pages had slow TTFB on repeat visits. The team saw high bounce rates on mobile and inconsistent search placements. The goal: cut TTFB 50–70% without changing hosting provider immediately.

What we changed

1. Edge caching for HTML fragments: push critical snippets to the CDN and render the rest client‑side.
2. Query cost optimization: batch requests and add selective denormalization; we followed guidance from the engineering playbook at Engineering Operations: Cost‑Aware Querying for Startups.
3. Static fallbacks for rarely changing pages: export monthly snapshots to an archive bucket.
4. Asset optimization and an AI upscaler: we integrated an automated image pipeline and tried the new webp→jpeg upscaler news tool to preserve legacy assets — see the announcement at JPEG.top Launches Native WebP‑to‑JPEG AI Upscaler for context on options.

Process & team changes

Performance wasn’t just engineering work — product and ops aligned on KPIs. We ran weekly performance sprints, created a TTFB playbook, and added alerting to watch query costs. Migrating processing off the critical path also allowed us to consider a free hosting migration later; see the practical roadmap at Migrating from Paid to Free Hosting.

Results

• TTFB reduced by an average of 60% on mobile.
• Core Web Vitals improved and mobile search impressions rose 22%.
• Engagement doubled — saves and event RSVPs increased significantly.

Lessons learned

• Start with the critical path: identify slow database queries and cache them near users.
• Measure query costs to avoid runaway bills — use the toolkit at query cost toolkit.
• Automation for images saves long‑term time; the new upscaler approaches can be useful for legacy archives (JPEG.top upscaler).

Next steps for the directory

The team plans to explore hybrid free hosting for static assets and keep transactional work on modest paid compute, following the migration patterns in the migration roadmap.

Conclusion: Performance tuning is high ROI for local products — start with the critical path, manage query costs, and automate media pipelines to protect both UX and budgets.