Skip to content
Faisal Mahmud TusherDigital marketing, SEO & analytics
with Faisal Tusher

Conversion Rate Optimisation › Troubleshooting

Conversion Rate Dropped After Redesign

Why conversion falls after a website redesign: broken tracking, removed proof, changed forms, slower pages, lost message match and moved landing URLs.
Troubleshootingintermediate8 min readUpdated

First, confirm the drop is real

Before blaming the design, make sure like is being compared with like. Pull orders, leads or bookings from the system that records them, not from analytics, for comparable full weeks before and after launch. If the business figures held while the reported conversion rate (CVR) fell, the problem is measurement and the first cause below comes first. Then confirm that the conversion definition is the same on both sides: a rebuilt template may count only completed checkouts where the old one counted form views, or the reverse.

Next, hold the traffic mix still. Compare the rate by source and device, before and after, rather than the overall figure; a relaunch is often timed with a new campaign or a pause in an old one, and a different crowd converts at a different rate whatever the page looks like. Finally, list the pages that carried most conversions before the relaunch and pull up their old versions from screenshots, an archive or the staging history, because the comparison between old and new is the core of every check that follows. If it is the number of visits from search that fell rather than the share who convert, the traffic drop after migration diagnosis covers that separate problem, and the conversion rate troubleshooting hub lists the other conversion symptoms.

Diagnosis table

Diagnosis

SymptomCheckLikely causeFix
Reported conversions fell while orders or leads in the business system heldWalk the conversion path with a debugger; compare event names and triggers with the old siteTracking broken or redefinedRestore and verify each key event; annotate the gap
Visitors reach the decision point and hesitate; recordings show scrolling in search of somethingOld and new versions of key pages: reviews, guarantees, credentials, policies near the buttonTrust and proof removed or movedPut proof back beside the decisions it supports
Form starts held but submissions fellField count, validation rules and form tool, old against new; errors by deviceForms changedRemove added fields; fix validation; test the new tool
The drop is concentrated on phones; exits rise before the first screen rendersField data for loading, responsiveness and layout shift, before and afterPages slower or harder to use on mobileFix the heaviest assets and tags; restore layout stability
Paid traffic converts worse than before while organic holdsAd headline and offer against the new page's first screen, per campaignLost message matchRestore the ad's words and offer in the headline
Campaign traffic now lands on the homepage or a generic pageFinal URLs in ad accounts and emails; where old landing addresses redirectLanding URLs moved or redirectedMap every campaign URL to its equivalent page

Fix, verify and prevent each cause

Tracking that broke or changed meaning in the rebuild

A new template is a new set of buttons, forms and confirmation pages, and tags that relied on the old selectors, data layer keys or page addresses stop firing or fire on the wrong action. Google documents that Analytics measures interactions as events, some collected automatically and others defined by the site, so a rebuilt page can stop sending an event or send it under a new name.

Fix: work through the tracking verification checklist on the live site, confirming that every Key Event fires once per real action and that ad platform conversions still record.

Verify: for a full closed day, reported conversions and the business system's count agree within their usual margin.

Prevent: run the checklist on staging before launch and again on launch day, with the measurement owner signing it off.

Proof that disappeared from the decision points

Redesigns often favour clean layouts, and testimonials, accreditation, guarantees and delivery terms are the first casualties: removed, collapsed into a carousel or moved to the footer. Visitors who needed that reassurance at the moment of deciding no longer find it.

Fix: compare the proof on old and new versions of the main pages and put each element back beside the decision it supports, as the trust and proof guide describes: credentials near the headline, evidence beside each claim, the guarantee by the price and a testimonial by the form.

Verify: the rate on the restored pages recovers against comparable weeks, and scroll data shows the proof is now seen on phones.

Prevent: keep an inventory of proof elements per key page and make their placement a review item for any future template.

Forms that grew, changed rules or changed tool

A rebuild can add fields that marketing always wanted, apply stricter validation, or move the form into a new tool that behaves differently on some devices. Each change adds cost at the last step.

Fix: compare the old and new forms field by field and error by error, remove what the next step does not need, and relax validation to accept what people actually type, following the forms guide and Nielsen Norman Group's form usability recommendations.

Verify: submissions over starts return to their previous level on each device, and the share of failed submit attempts falls.

Prevent: require any new field to be justified against the completions it costs, and test every form on a phone after each release.

Heavier pages that load and respond more slowly on phones

New designs bring larger images, more fonts, animation and fresh third-party tags, and the cost lands on mid-range phones over mobile networks. web.dev describes Core Web Vitals as field measures of loading, responsiveness and visual stability, assessed on real visitors.

Fix: compare field data for the old and new templates, find the largest element, the slowest interaction and the shifting component on the pages with most paid traffic, and fix them in the template as the speed and mobile guide sets out.

Verify: field measures return to their previous bands and the mobile conversion rate recovers over the following weeks.

Prevent: set a performance budget for templates and test new tags on a throttled phone before publishing them.

Headlines that no longer echo the ads

Redesigns tend to rewrite page copy in the new brand voice, and in doing so drop the exact promise, offer or price that the ads and emails make. Visitors arriving from those links no longer recognise what they clicked for.

Fix: audit each campaign's ad against the first screen of the page it lands on, as the message match guide describes, and restore the ad's words, offer and any price in the headline and subhead.

Verify: the conversion rate from paid sources returns towards its pre-launch level, while organic traffic is unaffected by the change.

Prevent: include the live ads and email offers in the copy review for any redesigned Landing Page.

A rebuild often changes page addresses. If the ad accounts, emails and partner links still point at the old ones, visitors land on redirects that end at the homepage or a generic category, or on a missing page.

Fix: map every old address to its closest new equivalent, never to the homepage, as the website migration checklist requires, then update final URLs in ad accounts and links in active emails directly.

Verify: a crawl of every campaign URL returns the intended page in a single hop, and landing-page reports show campaign traffic arriving on the expected pages.

Prevent: export all campaign destinations before launch and include them in the redirect map.

A prevention routine for redesigns and relaunches

Almost every cause on this page can be caught before launch. Before the rebuild starts, record the pages that carry most conversions, take screenshots of them on a phone and a desktop, list the proof elements and form fields on each, export every campaign destination and note the current field data for speed. On staging, run the tracking checklist, compare each key page with its screenshots for missing proof and changed forms, read each ad beside the page it will land on, and load the main templates on a mid-range phone over a throttled connection. On launch day, run the tracking checklist again, crawl every campaign URL and reconcile a closed day of conversions with the business system. For the following weeks, compare conversion by source and device with the same weeks before launch, and investigate the first segment that diverges rather than waiting for the overall figure to settle. Keep a dated log of everything the relaunch changed so that, if the rate does fall, the list of suspects already exists. To judge the next redesign against a comparison group rather than after the fact, the guide to evaluating a website redesign covers template tests before launch, staged rollouts and rollback rules agreed in advance.

Illustrative example — not a real result

A furniture retailer's relaunch

A furniture retailer relaunched its site on a new platform and saw its reported conversion rate fall from 2.4% to 1.6% over the next three weeks, although orders in the shop system fell only from about 1,900 to 1,700 a month. The debugger showed the purchase event firing only when the confirmation page loaded twice, so a third of real orders were missing from analytics. The order shortfall had two causes: delivery charges and the returns promise, once beside the basket button, were now in a collapsed panel, and 140 campaign URLs redirected to the homepage. The team fixed the event, restored the delivery and returns lines beside the button, and mapped every campaign URL to its product range page. Within a month orders were back to about 1,950, and reported conversion read 2.5%.

Common mistakes

Redesigning again before checking the tracking. Comparing a launch month with a different season. Judging the overall rate instead of rates by source and device. Treating removed reviews and guarantees as clutter. Redirecting old campaign pages to the homepage. Changing the conversion definition during the rebuild without saying so.

Frequently asked questions

Should we roll back to the old website?

Rarely all of it, and only if the drop is large, confirmed in business figures and the cause cannot be found quickly. A full rollback discards whatever the redesign improved and usually reintroduces the problems it was meant to solve. It is normally faster to compare the old and new versions of the pages that carry most conversions, restore the specific elements that were lost, and keep the rest.

How long after a redesign should we wait before judging conversion?

Long enough to cover at least one normal business cycle with the same traffic mix as before, but not so long that a real fault keeps costing sales. Tracking and redirect faults should be checked on launch day. Changes to proof, forms and message are better judged over full weeks, compared with the same weeks before launch and, where seasonality matters, with the same period a year earlier.

Our organic traffic also fell after the relaunch. Is that the same problem?

It is a separate problem with separate causes. Fewer visits from search after a relaunch points at redirects, indexing and lost content, which the migration diagnosis in the search section covers. This page is about a lower share of the visitors you do receive going on to convert. Check both, because a relaunch often causes a little of each.

Terms used on this page

Where this page is used

Also related

Sources

  1. Web Vitals — web.dev
  2. Website Forms Usability: Top 10 Recommendations — Nielsen Norman Group
  3. [GA4] About events — Google Analytics Help

Related