I audit the pages blocking revenue, form a falsifiable hypothesis for each, rebuild and wire the analytics, then prove the lift before anything else changes.
- Design
- Build
- Launch
fixed-price project
Who this is for
SaaS founder or e-commerce operator with steady traffic and a conversion rate that hasn't moved in months. Heatmaps show where visitors drop off. GA4 shows the funnel break. Nobody internally has the time or the skill mix to redesign, ship, and run a split test in the same sprint.
The pain today
- A redesign ships, looks sharper, and conversions drop anyway
- Stakeholders debating button colors instead of reading the funnel data
- Heatmaps and session recordings that nobody acts on
- No staging setup, so every change goes straight to production
- Analytics too shallow to separate design impact from traffic-mix shifts
The outcome you get
- Primary-CTA click-through lifted 25–60% within 90 days of launch
- A written hypothesis for each page change — falsifiable, with a target range
- GA4 + Plausible dual setup that separates signal from noise
- Split-test infrastructure baked into the template, not bolted on after
- Full handover on day one: Figma source, code, analytics docs, test results
Why conversion rate optimization redesigns fail
The most common way a redesign hurts conversion is by changing too many things at once. Brand refresh, new copy, restructured navigation, new CTA placement — all live on the same day. Traffic drops, leads drop, and nobody can say why because every variable moved simultaneously.
My approach to website redesign for conversions starts by separating the work into layers. Brand and visual refresh on one track. Structural conversion changes on another. I rebuild the pages that drive 80% of revenue first — typically the pricing page, the top organic entry points, and the signup step — with a single, written hypothesis per page. Everything else follows only after the first lift is confirmed and measured. You keep what's working, and the team can explain every result to the board.
Primary-CTA click-through lifted 25–60% within 90 days of launch
The CRO-first redesign framework: four stages
Audit. I review heuristics, heatmaps, session recordings, and the GA4 funnel for each target page, then run short interviews with three to five recent buyers. The output is a prioritized list of friction points scored 1–5, with a recommended fix for each.
Hypothesis. One primary change per page, written as a falsifiable statement: "Moving the primary CTA above the fold on the pricing page will increase click-through by 20–35% at 95% confidence." No vague goals.
Design and build. Wireframe, copy, visual, then the staging build — all matched to the hypothesis. The staging environment mirrors production, including redirects and structured data.
Test. A split test runs against the live page for a defined sample size. Results go into a living doc so future redesigns build on what actually worked. The framework came directly from my work at Cuez and GigEasy, where every design decision had to move a number, not just look cleaner.
45%: Bounce rate reduction after launch.
Incremental redesign versus full-site rebuild
The right choice depends on how broken the current site is and how much traffic you have to run a test. If you have fewer than 1,000 monthly visitors per target page, a full redesign is often the faster path because you will not accumulate statistical significance quickly enough to run meaningful split tests on individual components.
With enough traffic, I strongly recommend starting with the three to five pages that drive 80% of conversions and leaving everything else alone. Fewer simultaneous changes mean cleaner data and lower risk. A full-site rebuild is still on the table when brand alignment or a platform migration forces it, but I treat it as a last resort, not a default. In 17 years I have watched more full-site projects tank conversion than incremental ones.
Analytics completeness as a conversion blocker
Most of the sites I audit have at least one major analytics gap — a form completion event that never fires, a CTA click that routes to an external page without a referral parameter, or a GA4 setup that tracks pageviews but not micro-conversions.
Before any design work starts, I map every conversion event against what GA4 and Plausible actually record. Gaps get fixed first. Designing against incomplete data produces exactly the redesigns that fail: the team optimizes for a metric that doesn't reflect what's actually happening.
The Cuez project is a clear example. The API was running at 3 seconds on the signup flow. That latency was invisible in the standard analytics dashboard but showed up the moment I added a custom event tracking load time per step. The fix — 10x faster response time, down to 300ms — moved downstream conversion without a single design change. Fix the measurement first, then optimize.
SEO protection during a conversion-focused rebuild
A CRO redesign that destroys organic rankings is a net negative even if conversion improves on the remaining traffic. I protect SEO as part of the build spec, not as an afterthought.
This means: URL structure preserved wherever possible; structured data from the old pages mapped to the new ones before launch day; canonical tags audited; 301 redirects staged and tested against a crawl; and Search Console monitored weekly for the four weeks following launch. If a keyword ranking drops after the redesign, I want to know within days, not months.
For the LAK Embalagens project, the B2B manufacturer site rebuild kept every existing URL, updated structured data for the new product categories, and added missing image alt text. Bounce rate dropped 45% and Search Console impressions tripled — conversion and SEO moving in the same direction.
Scope, pricing, and what the handover includes
A conversion rate optimization redesign fits the Websites Redesign tier, starting at $4,999 fixed-price for a single-page rebuild and scaling up for multi-page projects with analytics integrations. Typical timeline runs three to five weeks from audit to the first live variant, then four to six additional weeks if I stay through the measurement phase.
Every build ships with the 14-day money-back guarantee and a one-year bug warranty. Work Made for Hire, so the Figma source, code, analytics configuration, and test documentation are yours on day one after launch. No retainer lock-in, no change orders for scope that was in the original brief, no surprises.
Recent proof
A comparable engagement, delivered and documented.
Turned a B2B manufacturer into a digital showroom
Designed and developed a high-performance institutional website to showcase packaging solutions and generate qualified leads.
Read the case studyKeep reading
Frequently asked questions
The questions prospects ask before they book.
The typical range is 25–60% on primary-CTA click-through within 90 days, assuming more than 1,000 monthly visitors to the target page. I share a target range after the audit, not before, because the number depends on how broken the current page actually is. A page with a buried CTA and no social proof has more room to move than one that's already well-optimized.
No. This service covers strategy, heuristics audit, hypothesis writing, design, build, analytics wiring, and the first round of split testing. If you later want an ongoing experimentation specialist, the handover is clean because all hypotheses, test results, and analytics events are documented and the split-test infrastructure is already live in your template.
Not if the build spec includes SEO protection from the start. I preserve URLs wherever possible, map structured data before launch, stage 301 redirects on URL changes, and monitor Search Console for four weeks post-launch. The LAK Embalagens rebuild kept all existing URLs and saw Search Console impressions triple after launch, so conversion and SEO can move in the same direction.
Start with the three to five pages that drive 80% of conversions. Fewer simultaneous changes mean cleaner measurement and lower risk. A full-site rebuild is sometimes forced by a platform migration or major brand shift, but for CRO purposes it's rarely the right first move. I can redesign the full site once the incremental approach proves the framework on your highest-traffic pages.
If within the first two weeks of the engagement you decide the fit is wrong — scope mismatch, communication style, early output quality — you get a full refund. After two weeks we are in execution and the guarantee transitions to the one-year bug warranty on all shipped code. No fine print beyond that.
Every project ships with a dual analytics setup: GA4 for event-level data (CTA clicks, form starts, form completions, drop-off by step) and Plausible for clean, cookie-light traffic data. Results are reported with confidence intervals, not just raw percentage changes, so the next decision is based on signal. The split test stays live until it reaches statistical significance at 95% confidence or a pre-agreed maximum run time.
Yes, and fixing the analytics is the first deliverable. Designing against incomplete data is one of the most reliable ways to produce a redesign that moves the wrong metric. I audit every conversion event before the design phase starts, fill the gaps, and give you a clean measurement baseline. That work is included in the project scope.