Core Web Vitals for Indian SMBs — 90-day plan

Core Web Vitals for Indian SMB websites: a 90-day plan.

Core Web Vitals are three scores. They are LCP, INP, and CLS. Google uses them to rank pages. Most Indian SMB sites fail all three on phone. This plan fits one engineer over 90 days.

The thresholds.

Here are the three numbers to hit.

LCP must be under 2.5 seconds. INP must be under 200 ms. CLS must be under 0.1. If any number is too high, that metric is failing.

How to measure.

Use real user data. Not lab data. Here is how.

First, check the Chrome User Experience Report. It is free. It shows your speed over 28 days.

Second, add the speed data lib to your site. This gives you data from your own users.

Third, use PageSpeed Insights. It pulls your CrUX data. It adds Lighthouse data. Our free Core Web Vitals tester runs the same three checks in your browser.

Lighthouse and WebPageTest are lab tools. They run on fast links. They will under-report what a user on a cheap Android phone sees.

The five biggest fixes.

Here are the five changes that move the needle most. These same fixes feed into our broader performance budget playbook, which covers how to enforce JS and image budgets in CI so regressions get caught before they ship.

1. Image optimization.

Most Indian SMB sites have 2 to 5 MB hero images in JPEG. You can fix this fast.

Switch to AVIF or WebP. Both are 50 to 70 percent smaller. Add a srcset so the screen picks the best size. Load below-fold images on demand. Load the LCP image first with high priority. The image optimizer converts a whole media folder in one pass.

This alone can fix LCP. It is the best win.

2. JS bundle size.

Most Indian SMB sites ship 1 to 3 MB of JS. JS slows sites down.

Turn on code split. Load below-fold parts on demand. Swap heavy libs for light ones. Remove third-party tags no one uses. Then set a hard ceiling per route. The performance budget playbook shows how to fail a build when the budget slips.

This can fix INP. Do this next.

3. Font loading.

Web fonts often cause CLS and slow LCP. This is easy to fix.

Self-host your fonts. Use swap so text shows while the font loads. Load fonts needed above fold first.

This can fix CLS. Do this too.

4. Third-party tags.

Chat widgets, tracker scripts, and tag managers add 500 KB to 1.5 MB of JS. Most sites have too many of these.

Audit every tag. Remove the ones no one uses. Load tag managers with defer. Gate them behind user consent.

This can fix LCP and INP. Keep only key tags.

5. Cache headers.

Most Indian SMB sites miss basic cache. This is a fast win.

Set immutable cache for hashed assets. Set public cache for HTML. Turn on Brotli.

Next time, this can cut LCP in half. Users will thank you.

Indian-specific issues.

Indian users face four specific challenges. You must plan for these.

Low-speed 3G and 4G are common outside big cities. Keep page weight under 1.5 MB. Cheap Android phones have slow CPUs. Test on a 4x slow CPU. CDN coverage varies. Use Cloudflare, BunnyCDN, or AWS CloudFront with a Mumbai edge.

The 90-day plan.

Here is the week-by-week calendar.

Weeks 1 to 2: Measurement.

Track real users. Pull CrUX data for your top pages. Build a simple chart.

Weeks 3 to 8: Quick wins.

Start with image optimization. Remove third-party tags no one uses. Self-host your fonts. Add cache headers. These four fixes give big wins fast.

Weeks 9 to 12: Structural fixes.

Add code split and on-demand loading. Swap heavy libs for light ones. Move to a CDN with a Mumbai or Singapore edge. Switch image formats to AVIF or WebP.

Weeks 13 to 14: Polish.

Inline critical CSS. Load the LCP image and fonts first. Set up a Lighthouse CI check on every pull request.

After 90 days, you should be in the green on all three metrics for most real users.

What to skip.

Do not use a performance plugin that claims to auto-fix everything. These tools hide the problem. The real wins are structural changes.

Also skip AMP. It is no longer a ranking factor. A separate AMP page is not worth the cost.