Redesigns tank conversion because they change everything at once: layout, copy, URLs, tracking and speed. Some of that was quietly working. Without a baseline and a staged rollout, you can't tell what broke, so you fix by guessing. Redesign in measured steps and protect what already converts.
- A big-bang redesign changes too many variables to diagnose. A staged one lets each change earn its place.
- The usual damage is invisible on launch day: dropped tracking tags, broken redirects, rewritten copy that used to convert, and a heavier page.
- Get a baseline before you touch anything. If you can't state your current conversion rate by page and device, you can't call the new one better.
The new site looks better. Everyone on the team says so. Three weeks later, leads are down and nobody can point to why.
This is the most common way a redesign fails, and it's rarely about taste. It fails because a redesign is not one change. It is dozens, shipped on the same day, and the number you care about moves with no explanation attached.
What actually breaks
Most conversion loss traces back to a short list. None of it shows up in a design review.
- Tracking. Tags and conversion events live in templates. Rebuild the templates and the events can vanish. Your dashboard then shows a "drop" that is partly a measurement gap, and you can't separate the two.
- Redirects. Change a URL without a 301 and every link, bookmark and search ranking pointing at the old address lands on an error. Paid landing pages break the same way.
- Copy that was working. A headline or proof point that quietly converted gets rewritten because it "sounded dated." Nobody checked whether it was carrying the page.
- Weight. New fonts, sliders and video make the page heavier. Google's own Core Web Vitals thresholds still apply: a Largest Contentful Paint over 2.5 seconds is outside the "good" band, and the new design has to clear that bar too.
- Friction on mobile. A layout approved on a desktop monitor is often the one that fails on a phone, where many of your visitors will actually meet it.
Why nobody can diagnose it afterward
When everything changes on one date, you get one data point: before and after. Any of the five causes above, or a real design problem, or seasonality, could explain the movement. You're left arguing opinions, and the loudest one wins.
A staged rollout gives you a series of clean comparisons instead. Change the navigation, measure. Change the product page, measure. If a step hurts, you know which step, and you revert one thing instead of a whole site.
How we approach a redesign
This is the sequence we use before recommending a rebuild at all. The first step often changes the answer. Sometimes the page needs three fixes, not a new site, which is the same diagnose-first logic behind why page speed is costing you conversions.
- Baseline first. Record conversion rate by page, by device and by traffic source before anything moves. This is the number the new design has to beat.
- Keep what converts. List the pages, headlines and proof elements carrying revenue. Carry them across intact unless there's a test showing the change wins.
- Map every URL. Build a one-to-one redirect map before launch, and test it.
- Audit tracking on staging. Every conversion event fires on the new build before it goes live, not after.
- Roll out in stages. Highest-traffic template first, with the old version kept for comparison where traffic allows.
- Set a speed budget. The new pages must pass Core Web Vitals, or they don't ship.
The point
A redesign is a bet on a hypothesis, and a hypothesis needs a control. Protect the pages that already earn, measure each change, and the new site gets judged on revenue instead of on how it looks in a screenshot.
Related service: Web Design & Development