Should budgets be different for desktop and mobile?

Yes. Budgets are about real users. Mobile CPU/network is your true bottleneck, so set stricter JS and TTFB budgets there. Keep desktop budgets, but optimize primarily to mobile p75.

How do we handle big launches or seasonal spikes?

Don’t change targets; change procedures. Use canaries, relax alert sensitivity windows, and pre-warm caches/CDN. If you must breach temporarily, file an RFC with rollback and a buy-down plan in the next sprint.

What if Lighthouse passes but RUM is red?

Synthetic ≠ real. Check third-parties, geo latency, long-tail devices, and cache hit rates. Treat Lighthouse as a gate and RUM as the truth. Tune caches and prioritize the hero resources for the real world.

How do we keep Marketing from adding heavy tags?

Create a third-party budget and a change process. Every new tag: async/defer by default, consent-gated, measured in a staging sandbox, and must offset cost by removing or deferring something else. CI and RUM both enforce the line.

Performance-optimization · Oct 4, 2025 · 10 minute read

The Day Marketing Added Pixel #13: Performance Budgets That Keep LCP Green

Budgets aren’t scorecards — they’re guardrails that keep your UX fast when scope, traffic, and third‑parties creep. Here’s how to make them real, measurable, and enforced in CI.

Alex Ramirez

Principal Engineer, GitPlumbers

20 years shipping and fixing web stacks from jQuery-era monoliths to Next.js at scale. Ex-platform lead, CDN tinkerer, and the person teams call when perf breaks on launch day.

If you can’t enforce it in CI, it’s not a budget — it’s a wish.

Back to all posts

The Day Marketing Added Pixel #13: Performance Budgets That Keep LCP Green

Key takeaways

Implementation checklist