Skip to the case study

B2B SaaS • Trust Diagnosis • Conversion UX

CodSphere Built an AI Platform. Its Own Homepage Didn't Show It.

No organic search traffic before. 75 organic users after.

View Live Website (opens in a new tab)
CodSphere logo

CodSphere

Role
Product Designer, design owner
Team
1 developer, 1 stakeholder
Timeline
2 weeks
What I owned
  • Homepage hierarchy and IA
  • Trust signals before CTAs
  • Handoff and WCAG 2.1 AA

Problem

CodSphere had a working product and no way to prove it.

Multiple offerings competed for attention with no hierarchy, so visitors could not tell what CodSphere does. That cost trust before anyone explored further.

Where it started

0
found them through Google
122
people visited
329
visits
1,474
pages opened

GA4, 1 Aug to 1 Nov 2025 — active users, sessions and views. Organic search does not appear in the property at all, which is why the first number is zero rather than small.

“You build websites, but your own site feels complex.”
Prospective lead, unprompted feedback

Proof + Process

GA4, 25 Nov 2025 to 25 Feb 2026, against the 92 days before launch. Tracking on the baseline property only starts around 21 Aug, so the before window holds about 72 days of collection against 92 after — the comparison understates the change rather than flattering it. Organic search is now 24% of all sessions.

A trust audit of the live site and a competitor scan came first; both fed the decisions below. All three shipped in the same release, so each number is the event that decision is closest to, not an isolated result.

Decision

Hero rebuild and proof placement — 122 of 367 people scrolled to 90% of the page

Shipped
After the redesignBefore the redesignBeforeAfter

Before -- product buried below text · After -- product-first hero, proof below fold

Drag the handle, or focus it and use the arrow keys, to compare

The old hero led with copy and pushed the product below it. I rebuilt it product-first and moved proof directly under the fold, so the thing being sold is the first thing seen.

Decision

Trust signals and interaction fix — 166 of 367 people interacted with the page

Shipped
Before the redesign
Before -- placeholder testimonial copy
After the redesign
After -- real Google Reviews

The page ran placeholder testimonial copy next to a carousel that felt stuck. I replaced the copy with verified Google Reviews and the carousel with a pattern people already know how to operate.

Decision

Success Stories and Free Trial pages — 355 visits, 58 form starts from 23 people

Shipped
Success Stories page, full scroll
Success Stories -- scroll to explore
Free Trial page, full scroll
Free Trial -- scroll to explore

Neither page gave a visitor enough to commit. Success Stories was rebuilt around real outcomes and Free Trial around the objections that were stopping signups. Both shipped in the same release as the homepage, so neither has a result I can isolate from it.

Learning + Trade-off