SEOPerformanceReact/ViteAEO/GEO

How Yotvis Achieved 95/100 PageSpeed & Google Indexing in 3 Days

What did Yotvis achieve?

Yotvis built and optimized its own website from zero to Google-indexed in 3 days, achieving a 95/100 mobile PageSpeed score, 99/100 desktop, SEO 100/100, and Accessibility 100/100 — using React 18, Vite, react-helmet-async, JSON-LD structured data, and AEO/GEO implementation. The site covers 26 pages including service pages and blog articles.

Client: Yotvis (self)Timeline: Jun 23–26, 2026Sprint: 72 hoursPages: 26

Mobile Performance

95/100

↑ from 73

Desktop Performance

99/100

First run

Indexed In

3 Days

New domain

SEO Score

100/100

Perfect

01 · Overview

The Brief

Yotvis is a Chennai-based digital agency that builds complete digital presences for startups and growing brands. This case study documents a 4-day technical sprint in which Yotvis designed, developed, and optimized its own website — achieving Google indexing in 3 days, a mobile PageSpeed score of 95/100, desktop score of 99/100, and a perfect SEO score of 100/100 using React 18, Vite, JSON-LD structured data, and Answer Engine Optimization (AEO) strategies.

All metrics are sourced from Google PageSpeed Insights and Google Search Console URL Inspection snapshots captured between Jun 23–26, 2026. Screenshots in the Proof section are unedited source captures.

02 · Stack

Technology Stack

Frontend

React 18 + Vite (ESM build)

Routing

React Router + hover pre-fetching

Head Management

react-helmet-async

Structured Data

Hand-authored JSON-LD (6 schema types)

Crawlers

Static sitemap.xml + robots.txt

Hosting

Edge CDN · HTTP/2 · Brotli · Immutable cache

03 · Implementation

What We Built

Site Architecture

The build was configured with manual chunk splitting to separate vendor code from route-level application code. Each route exposes a statically known title and meta set so that crawlers receive a meaningful document even before JavaScript hydration completes.

Sitemap & robots.txt

A static sitemap.xml is emitted at build time from a single source-of-truth route list, ensuring no drift between the router and the sitemap. Both files are submitted to Google Search Console immediately after deployment.

Structured Data — JSON-LD

JSON-LD payloads were authored per page type and validated via Rich Results Test before deployment. Homepage carries Organization + WebSite schemas, service pages carry Service + BreadcrumbList, blog articles carry Article + FAQPage.

AEO & GEO Readiness

Three signals weighted heavily by AI search surfaces: (1) explicit Q&A blocks marked up with FAQPage schema, (2) concise factual summaries at the top of each service page, (3) entity-consistent NAP and service taxonomy so that models resolve Yotvis as a single well-defined entity.

PageSpeed Optimization

Largest gains from: (1) deferring non-critical JS via code-splitting and lazy-loading, (2) replacing render-blocking web fonts with system-font stack + font-display: swap, (3) serving all imagery through responsive srcset with lazy loading. Together: TBT from 220ms → 50ms, LCP from 5.1s → 2.8s.

04 · Results

Before vs After

Mobile Performance

7395

+22

Desktop Performance

—

LCP

5.1s2.8s

-2.3s

TBT

220ms50ms

-170ms

SEO Score

100

Perfect

Indexed In

03 Days

72h sprint

05 · Timeline

Indexation in 72 Hours

Jun 23 · 15:49 IST

Site deployed. URL not yet on Google.

Jun 23 · 15:51 IST

Indexing requested — URL added to priority crawl queue.

Jun 23 · 16:09 IST

Live test: Crawl allowed ✓, Page fetch successful ✓, Indexing allowed ✓

Jun 25 · 17:20 IST

Googlebot smartphone crawled yotvis.com

Jun 26 · 16:24 IST

PageSpeed: Mobile 95 / Desktop 99 / SEO 100 / Accessibility 100

Jun 26 · 16:51 IST

URL is on Google — Page is indexed ✓

06 · Proof

Unedited Screenshots

All screenshots are source captures from Google Search Console and PageSpeed Insights. Click to expand.

07 · Reproduce It

The 10-Step Checklist

Follow these steps on any new React + Vite deployment to reproduce this SEO baseline.

Configure vite.config.js with manualChunks for react-vendor and helmet, enable cssCodeSplit, minify with esbuild

Install react-helmet-async and wrap the application root with <HelmetProvider>

Create a route registry as single source of truth for both React Router and sitemap generator

Generate sitemap.xml and robots.txt at build time from the route registry; emit to /public

Author JSON-LD payloads per page type and validate each with the Rich Results Test

Add a <Helmet> block to every route — title, meta description, canonical, robots, Open Graph

Run Lighthouse on production build; iterate on LCP, TBT, and CLS until mobile Performance ≥ 90

Deploy to edge CDN with Brotli compression and immutable cache headers for hashed assets

Submit sitemap.xml and request indexing for homepage in Google Search Console

Verify via GSC URL Inspection: crawl allowed = Yes, page fetch = Successful, indexing allowed = Yes

08 · Takeaways

Key Findings

A React + Vite SPA can pass Core Web Vitals on mobile if the build is configured for code-splitting, pre-rendered <head>, and lazy-loaded below-the-fold routes.

Submitting sitemap.xml plus a manual indexing request on day one compresses first-index time from 1–2 weeks to roughly 3 days for a clean new domain.

JSON-LD authored per page type (Organization, Service, Article, FAQPage) is the single highest-leverage action for AEO/GEO readiness.

react-helmet-async with build-time pre-rendering is sufficient for SEO on an SPA without the operational cost of full SSR.

Hitting SEO 100/100 is achievable on day one by enforcing canonical URLs, descriptive meta, alt text, and crawlable link text.

TBT dropped 77% (220ms → 50ms) by deferring GTM load and adding ScrollTrigger.config({ ignoreMobileResize: true }).

Want this for your website?

We'll do the same — SEO 100, indexed fast, performance-optimized.

Book a Call

Questions
About This
Project

Detailed answers on how we optimized and indexed the Yotvis platform.

By combining three actions on day one: submitting a complete sitemap.xml to Google Search Console, requesting manual indexing via URL Inspection, and ensuring the page fetch returned a clean 200 status with crawl allowed and indexing allowed both set to Yes. The blog articles also provided internal linking signals that Googlebot followed during its first crawl on Jun 25.

Three fixes drove the improvement: (1) preloading the LCP element (the Yotvis logo) directly in index.html with fetchpriority="high", (2) delaying Google Tag Manager by 3 seconds so it doesn't block the initial render, and (3) adding ScrollTrigger.config({ ignoreMobileResize: true }) to stop GSAP from triggering forced reflows every time the mobile address bar resized the viewport. Together these cut TBT from 220ms to 50ms and LCP from 5.1s to 2.8s.

Answer Engine Optimization (AEO) targets AI assistants like ChatGPT and Perplexity that answer questions directly rather than serving blue links. Generative Engine Optimization (GEO) targets large language models that cite sources in generated content. Yotvis implemented both by adding FAQPage schema to every content page, writing concise factual summaries at the top of each service page, and maintaining entity-consistent business information so that AI models can resolve "Yotvis" as a single, well-defined company.

react-helmet-async with build-time pre-rendering of the head shell is sufficient for most SEO use cases on a React SPA. Googlebot executes JavaScript and can index client-rendered content, so the critical requirement is that each route has a correct, unique title, meta description, and canonical tag — which react-helmet-async handles per route. Full SSR adds operational cost and is only necessary if you're targeting crawlers that don't execute JavaScript.

The 10-step reproduction checklist in the page is a framework any React + Vite project can follow. The core principles — code-splitting, per-route meta, JSON-LD, sitemap submission on day one — are stack-agnostic. What Yotvis adds is the specific implementation judgment: knowing which GSAP configuration causes forced reflows on mobile, why GTM needs to be deferred, and how to structure JSON-LD for AEO rather than just classic search. That's the expertise we bring to client projects.

Want to see results like this for your brand?

Back to Explore