The website redesign checklist that actually lifts conversions

Most website redesigns lose money. That isn't cynicism — it's what the data says.

Studies from Google, HubSpot, and our own client audits consistently show that around 60% of redesigns result in flat or lower traffic and conversions in the first 90 days. A surprisingly large share never recover.

The sad part is that this is mostly avoidable. Redesigns fail in predictable ways. After rebuilding 30-odd SMB sites across India, we've distilled the playbook into a 14-point checklist. If you do nothing else, do these in order.

Why most redesigns lose traffic

The three recurring culprits are:

• SEO gets broken on launch — 301 redirects are missing, titles change, URLs change, Core Web Vitals drop.
• Redesign is opinion-led, not data-led — someone's boss didn't like the old homepage, so it was thrown out. Nobody checked what was converting.
• Launch is treated as "done" — no monitoring, no fixes, no iteration. The site's best month is week one; it's been dying quietly since.

The checklist below prevents each of these.

Phase 1 — Audit

You cannot improve what you haven't measured. Before any design work happens, collect:

1. Current GA4 data (12 months minimum)

• Top 25 landing pages by organic traffic
• Top 10 pages by conversions
• Bounce rate and time on page for each
• Mobile vs desktop split
• Conversion rate by device and by page

2. Search Console baseline

• Top 50 ranking keywords
• Top 50 URLs by impressions
• Click-through rate by query

3. Heatmap / session recording snapshot

Install Hotjar or Microsoft Clarity (free) two weeks before you start. Watch 20 real sessions. You will learn more than any competitor audit will tell you.

4. The "why do they leave?" interview

Call three customers who chose you, and if possible two who didn't. Five 15-minute calls will beat any analytics dashboard.

If you haven't done an audit, you're not redesigning. You're redecorating with your eyes closed.

Phase 2 — Keep / Kill / Fix decision matrix

Every page on the old site goes into one of three buckets.

Keep

• High-traffic, high-converting pages — leave the URL, meta, and H1 alone. Improve only layout and speed.

Kill

• Pages with no traffic, no links, and no commercial value. Delete them. Set up 301s to the nearest relevant page.

Fix

• Medium-traffic pages that don't convert. These are the lift opportunity — rewrite copy, simplify the CTA, restructure the layout.

Most SMB sites have about 40% keep, 30% kill, 30% fix. The exact mix varies but the discipline of putting every page in a bucket is what prevents the "oh we forgot the pricing page" launch-day panic.

Phase 3 — SEO preservation

This is the single biggest source of post-redesign traffic loss. Treat it as engineering, not marketing.

5. Build the 301 map before the design is signed off

A spreadsheet: old URL, new URL, status code. Every old URL that exists today must be accounted for. Orphans die.

6. Preserve H1s and title tags on top-traffic pages

If a page ranks #3 for "B2B industrial fasteners Pune," do not change its H1 on launch day. Improve the layout around it, not the signal Google has been reading for three years.

7. Match canonical tags and meta descriptions

Keep canonical structure identical where the URL stays the same. Update meta descriptions only where you're confident the new one will lift CTR.

8. Preserve internal link structure

Footer links, main nav, in-body links — if /about linked to /services/cro from every page, keep that link. Internal links are SEO currency that's easy to spend accidentally.

9. Submit a new XML sitemap day one

And watch Search Console's "coverage" report daily for the first two weeks. Any "discovered but not indexed" URL is a signal to investigate.

Phase 4 — Design

This is where most redesigns go wrong, because clients have strong opinions and teams default to pleasing them.

10. Design for the user, not the boardroom

The user wants to know, in this order:

1. What do you do?
2. Who is this for?
3. Is it any good? (social proof)
4. What does it cost / how do I start?

Every page should answer those four in the first scroll. If your new homepage needs three scrolls to tell a user what you sell, you have designed for yourself, not them.

11. One primary CTA per page, always visible

"Talk to us," "Get a quote," "Book a call" — pick one. Put it in the nav, above the fold, and at the end of every section. The "call now + WhatsApp + enquiry form + chatbot + downloadable PDF" approach spreads conversion too thin.

12. Design for mobile first, desktop second

70%+ of Indian SMB site traffic is mobile. Design the mobile layout first and up-spec from there. Reverse order is how you end up with homepages that look great on a 27" monitor and cramped on a phone.

Phase 5 — Build

13. Build fast and build clean

In 2026, that means:

• Core Web Vitals in green on launch day: LCP < 2.5s, CLS < 0.1, INP < 200ms
• Images compressed and served in next-gen formats (AVIF / WebP)
• Fonts preloaded and subset
• No render-blocking scripts above the fold
• Structured data (Organization, LocalBusiness, FAQ, Service, Article) where appropriate

We build on Next.js because it makes the above mostly free. But the stack isn't the point — the performance target is.

Phase 6 — Launch

The night before launch is where the wheels fall off. Work the list.

14. The launch-day runbook

The checklist we run, every time:

• Staging QA complete on mobile, tablet, desktop, across Chrome, Safari, Edge
• 301 redirects tested from a sample of 20 old URLs
• Forms tested end-to-end (including the email/CRM destination)
• Tracking tested: GA4, Meta Pixel, Google Ads, GTM events all firing
• SSL certificate valid on all subdomains
• DNS TTL lowered 48 hours before cutover
• Robots.txt and XML sitemap reviewed
• noindex tags from staging removed — seriously, check this
• Search Console and Bing Webmaster Tools re-verified
• Analytics annotation added on launch day so the data spike (or dip) is labelled

The single most common post-launch bug we find on sites we take over? A stray noindex from staging that was never removed. It kills organic traffic overnight.

Phase 7 — Monitor

Launch isn't done. It's the start of the real work.

30-day check

• Compare GA4 traffic to the 30-day pre-launch baseline
• Fix any URLs showing high bounce and low time-on-page
• Review Search Console "coverage" for indexing errors

60-day check

• Compare keyword rankings to pre-launch
• Identify any pages that lost top-10 positions and investigate
• Review session recordings on the new top-converting pages
• A/B test one hero headline if you have traffic volume (>10,000 sessions/month)

90-day check

• Compare CPL / conversion rate to baseline
• Identify two pages to iterate further
• Lock in new title tags and meta descriptions based on what Search Console is telling you

Eight mistakes we see agencies make

Even experienced teams fall into these.

1. Changing URLs "because they're cleaner now" — with no redirect map. Bye-bye rankings.
2. Shipping without Core Web Vitals testing. Elementor/WordPress redesigns routinely launch with an LCP above 5s.
3. Removing "unsexy" pages that were quietly ranking for long-tail keywords.
4. Hiding the price page because someone higher up "wants leads to talk to us first." Your bounce rate will double.
5. Auto-sliders in the hero. They tank conversion. Every heatmap we've run for the last six years shows users barely interact with slide 2 or 3.
6. Generic stock photography on a local Indian business. It signals "not real."
7. No launch-day tracking QA. Silent tracking failures are worse than broken tracking — you don't know you can't trust your data.
8. Declaring victory at launch. The site is a hypothesis. The next 90 days prove it or disprove it.

When you need help

If your current site has 6+ months of traffic and isn't converting, a redesign framed as "CRO + SEO continuity" will almost always outperform a scratch rebuild. That's precisely what our Revamp service is built for — preserving the signal your old site has earned while fixing the layout, speed, and conversion paths that are losing you money.

See our TrustTec case study for a before-and-after example — same domain, same products, completely rebuilt presentation, and a 7× lift in monthly B2B enquiries within 90 days of launch.

---

If you're planning a redesign and want a second pair of eyes on the plan before you sign with anyone, get in touch. We're happy to tell you whether your brief is sound or whether it has the usual holes.