An ecommerce website redesign loses rankings and revenue for predictable reasons: URLs change without redirects, content that ranked gets cut, internal links to key categories disappear, and the new templates are slower or convert worse than the old ones. None of that is inevitable. A redesign that keeps its traffic starts with a URL map and a baseline, launches in stages where it can, and measures the right things for the first eight weeks.
What actually causes the drop
When stores come to us after a redesign that went wrong, the cause is usually one of these, in roughly this order of frequency:
[Inline image to add: Sample URL map spreadsheet: old URL, new URL, type, traffic, status, test result.]
Sample redesign URL map with old and new URLs and test results
-
Missing or wrong redirects. Old product and category URLs return 404s or redirect everything to the homepage.
-
Content removed. The new design dropped category copy, buying guides or product details that were doing the ranking.
-
Internal links changed. A simpler navigation removed links to categories that relied on them.
-
Technical regressions. Staging settings left on (noindex, blocked robots.txt), canonical tags pointing to the wrong place, missing structured data.
-
Slower pages or a worse buying flow. Heavier templates, new apps, a checkout or cart change that adds friction.
The first four hurt rankings; the fifth hurts revenue even when rankings hold. Plan for both.
Step 1: take a baseline before you touch anything
You cannot tell whether a redesign cost you anything without knowing where you started. Record, for the last three months at minimum:
-
Organic sessions and revenue by landing page, from analytics.
-
Clicks and impressions by page and query, from Search Console.
-
Conversion rate, average order value and add-to-cart rate by device and by template.
-
Core Web Vitals and lab speed scores for each main template.
-
A full crawl of the current site: every URL, its status, title, H1, canonical and internal links.
Export it. Analytics setups change during redesigns, and you want the old numbers in a file you control.
Step 2: build the URL map
The URL map lists every current URL that has traffic, links or rankings, and where it will live on the new site. It is the single most important document in a redesign, and the one most often done in a hurry in the last week.
-
Start from the crawl, plus URLs from Search Console, analytics and backlink data, since some valuable URLs are not linked internally any more.
-
Keep URLs the same wherever you can. The safest redirect is the one you do not need.
-
Map each changed URL to its closest equivalent: product to product, category to category. Do not send everything to the homepage.
-
Decide deliberately what happens to discontinued products: redirect to the closest replacement or category, or return a helpful 404 page, depending on whether an equivalent exists.
We cover the mechanics in detail in a redirect map that survives re-platforming; the same rules apply to redesigns that change URLs.
Step 3: plan redirects properly
Use permanent (301) server-side redirects, one hop from old to new, and keep them in place long term. Google’s guidance on site moves with URL changes and on redirects is the reference. Common mistakes we fix:
-
Chains, where an old redirect points to a URL that now redirects again.
-
Redirects that drop query strings needed for tracking or pagination.
-
Case or trailing-slash mismatches that make half the map fail.
-
Redirects added to a plugin or app with a row limit that silently ignores the rest.
Test the whole map automatically before launch: request every old URL and check that it returns one redirect to a live page.
Step 4: protect content and internal links
Compare the new templates with the old pages that rank. If a category page ranked with buying copy below the grid, the new category template needs a place for it. If product pages ranked partly on specifications or FAQs, keep them. Designers sometimes remove text to make pages look cleaner; the answer is usually to present it better, not to delete it.
Check internal links the same way: crawl staging and compare the number of internal links to your top categories with the current site. If a category lost most of its links because the menu got simpler, add them back through the menu, footer, related categories or guides. Our article on category page SEO explains why those links matter so much.
Step 5: launch in stages where you can
A big-bang launch is sometimes unavoidable, especially with a platform change. When it is not, reduce the risk:
-
Launch template by template. Ship the new product template first, measure, then categories, then the homepage.
-
Test against the old design. Where your platform allows it, split traffic between old and new templates and compare conversion before switching everyone.
-
Launch at a quiet time, not before a sale or peak season, with the team available for the following two weeks.
-
Keep the old theme ready to restore if something serious goes wrong.
Step 6: the launch-day checklist
CheckHowSite is indexableNo noindex tags or robots.txt blocks left from stagingRedirects workEvery mapped URL returns one 301 to a live pageCanonicals are correctPoint to the live domain and the right URLStructured data is validProduct, breadcrumb and organisation markup testedSitemaps updatedNew URLs listed and submitted in Search ConsoleTracking worksAnalytics, conversion tags and consent all firing correctlyCheckout worksReal test orders on each payment method and deviceSpeed holdsEach main template at least as fast as the baseline
Step 7: what to measure for eight weeks
Some movement in rankings after a large change is normal while search engines recrawl. What matters is the trend and where the drops are. Watch weekly:
MetricCompare withInvestigate ifOrganic clicks by page (Search Console)Same weeks in the baselineA page or template drops more than [X]%404s and redirect errorsZeroAny URL with traffic or links returns 404Indexed pagesPre-launch countBig drops or unexpected new URLsConversion rate by template and deviceBaselineDown more than [X]% for two weeksCore Web Vitals by templateBaselineAny template moves out of “good”
Speed deserves special attention because new designs often add larger hero images, carousels and apps. Our article on where ecommerce LCP actually goes wrong lists the usual causes.
Who needs to be involved
Redesigns lose traffic when SEO is consulted at the end. The people who should review the plan before design is signed off:
-
Whoever owns SEO: the URL map, content that ranks, internal links and technical checks.
-
Merchandising: which categories and products drive revenue, and how the team edits them day to day.
-
Customer service: the questions customers ask, which usually points to content the new design must keep.
-
Whoever manages tracking: analytics, conversion tags, consent and feeds.
-
The developer: redirects, performance budgets, and what the platform can and cannot do.
A realistic timeline
For a mid-sized store, we plan roughly: two to three weeks for the baseline, crawl and URL map alongside early design; the build with SEO and speed checks at each template; a week of staging tests including the full redirect test; launch; and eight weeks of monitoring. The URL map and redirect testing are the parts most often squeezed, and the parts that most often decide whether traffic holds.
If you are also changing platform
A replatform adds data migration (products, customers, orders, reviews) and URL structures you may not fully control. Treat it as a migration project with a design component, not the other way round, and give the URL map and data mapping their own time in the plan.
Our ecommerce design team runs redesigns with this process built in, from baseline to eight-week review. When the redesign includes a platform change, our store migration team handles the data and redirects, and ecommerce SEO covers the monitoring after launch.