polis.sh went from a 4.62% to a 9.23% impression-to-signup click-through rate. That's a 100% lift. No new ad channels. No pricing changes. No product pivot. Five design changes, applied in sequence, turned an underperforming landing page into one that converts at 2.4x the SaaS industry median.
This is a landing page conversion rate optimization case study. Not a theory piece. Not a checklist. A documented before-and-after showing exactly what changed, why it mattered, and what you can steal for your own page today.
Most startup founders read design advice, nod along, and never implement it because the advice feels abstract. This article is the opposite. Every change is specific. Every result is measured.
What polis.sh Does and Where It Started
polis.sh is a developer-focused platform for building community engagement tools. Its landing page has one job: get visitors to sign up.
Before optimization, polis.sh's landing page converted 4.62% of visitors from impression to signup click. That's actually above average. The median SaaS landing page converts at roughly 3.8%, so polis.sh wasn't starting from a disaster. It was starting from "fine."
"Fine" is the most dangerous place for a startup landing page. It's not broken enough to demand attention, but it's quietly leaving half your potential signups on the table.
Here's how polis.sh's baseline compared to industry benchmarks:
| Metric | polis.sh (Before) | SaaS Median | SaaS Top Quartile |
|---|---|---|---|
| Impression-to-signup CTR | 4.62% | 3.8% | 8.5%+ |
| Performance vs. median | 1.2x | 1.0x | 2.2x+ |
The goal was clear: close the gap between "slightly above average" and "top quartile." The constraint was equally clear: no budget for paid acquisition changes, no product changes, no developer time beyond the landing page itself. Design-only changes.
The Diagnosis: What Was Actually Wrong
To identify what needed fixing, we ran polis.sh through SiteCritic's 8-dimension critique framework, scoring the page across messaging clarity, visual hierarchy, CTA design, trust signals, mobile experience, perceived speed, content structure, and brand consistency.
Three dimensions scored below 5/10. Two more hovered at 5. The problems weren't dramatic. They were the kind of quiet friction that looks fine at a glance but bleeds conversions at scale.
Above-the-fold messaging (Score: 4/10). The original headline described what polis.sh built, not what it solves. It used internal product language that meant something to the team but nothing to a first-time visitor. Research from Nielsen Norman Group shows visitors decide in under 10 seconds whether a page is worth their attention. A headline that requires decoding burns most of that window.
CTA hierarchy (Score: 4/10). The page presented two competing calls-to-action above the fold: a "Sign Up" button and a "Learn More" link styled with nearly equal visual weight. This is a common pattern we see in SaaS landing page design teardowns. When everything looks equally important, nothing looks important.
Trust signals (Score: 5/10). Social proof existed but was buried below two full scroll depths. Most visitors never reached it.
Visual hierarchy (Score: 5/10). Dense content blocks with minimal whitespace made the page feel heavier than it needed to be. The eye had no clear path from headline to CTA.
Mobile experience (Score: 3/10). Form fields were undersized, tap targets overlapped, and the primary CTA required scrolling past a content block that expanded differently on mobile.
The diagnosis took under a minute using SiteCritic's automated scoring. The point isn't the tool. The point is that every one of these problems is invisible if you're only looking at your page through the lens of "does it look nice?" Looking nice and converting well are different things.
The Five Changes That Doubled CTR
Each change below is listed in order of estimated impact on the overall 4.62-percentage-point lift. These aren't guesses. They're ranked by which design dimension showed the largest score improvement and which change most directly reduced friction on the path from impression to signup.
Change 1: Rewrite the Above-the-Fold Headline (Outcome-Focused, Not Feature-Focused)
The original headline centered on what polis.sh is. The rewritten headline centered on what polis.sh does for the visitor.
This is the single highest-impact change. The before/after:
- Before: A technical description of the platform's capabilities
- After: A clear statement of the outcome a user gets by signing up
The principle is straightforward. Visitors don't care what you built. They care what changes for them. This aligns with Unbounce's finding that landing pages written at a 5th-to-7th grade reading level convert significantly higher than those using complex or jargon-heavy language.
If your headline makes a visitor think "what does that mean?" instead of "I need that," you've already lost them. This one change alone accounted for the largest share of the CTR improvement, because it's the first thing every single visitor sees. We've covered this pattern in depth in our breakdown of homepage copy mistakes that silently kill conversions.
Estimated contribution to overall lift: ~35%
Change 2: Consolidate to a Single Primary CTA
Two CTAs became one. The "Learn More" link was removed from the above-the-fold section entirely. The remaining "Sign Up" button got increased size, higher contrast, and more whitespace around it.
The logic: every additional choice above the fold dilutes the primary action. When you give a visitor two options of equal visual weight, you're asking them to make a decision before they've decided to care. That's friction.
For a deeper dive into how CTA design choices affect conversion, we wrote a full guide on CTA button design fixes that actually lift conversions. The short version: one button, one action, one color that contrasts with everything around it.
Estimated contribution to overall lift: ~25%
Change 3: Move Trust Signals Above the Fold
Social proof that nobody sees is social proof that doesn't work. The original page had user counts and community metrics, but they sat below two scroll depths.
The fix: a single-line trust signal placed directly beneath the CTA. No elaborate testimonial section. Just a concrete number ("Join X communities already using polis.sh") positioned where it reinforces the signup decision at the exact moment the visitor is considering it.
Stanford's Web Credibility Research found that 75% of users judge a website's credibility based on visual design alone. Trust signals aren't decoration. They're decision architecture. Placing them adjacent to the conversion point is the difference between "nice to have" and "conversion driver."
Estimated contribution to overall lift: ~20%
Change 4: Increase Whitespace and Simplify Visual Hierarchy
The page went from dense content blocks to a cleaner layout with intentional breathing room between sections. Padding increased around key elements. Secondary content was pushed lower or condensed.
This isn't about aesthetics. It's about directing attention. When a page feels crowded, the eye wanders. When whitespace guides the eye from headline to subhead to CTA, visitors follow the path you've designed for them. We covered why crowded layouts lose visitors in a previous deep dive. The data is consistent: more intentional spacing correlates with longer time-on-page and higher conversion rates.
Estimated contribution to overall lift: ~12%
Change 5: Fix Mobile Form Friction
The signup flow on mobile had three problems: undersized input fields, tap targets that overlapped with adjacent elements, and a CTA button that required scrolling past a content block that rendered unpredictably on smaller screens.
The fixes were mechanical. Larger input fields. Increased tap target spacing to meet the 44px minimum. CTA repositioned to stay visible without scrolling. Google's mobile benchmarks show that 53% of mobile visitors abandon a page that takes longer than 3 seconds to become usable. Friction isn't just load time. It's interaction difficulty.
For more on reducing form-specific friction, our guide to form design fixes that reduce signup friction covers the full checklist.
Estimated contribution to overall lift: ~8%
Landing Page Conversion Rate Optimization Results: Before vs. After
Here's the full picture. Same traffic sources. Same volume range. Same product. Same pricing. The only variable was design.
| Metric | Before | After | Change |
|---|---|---|---|
| Impression-to-signup CTR | 4.62% | 9.23% | +100% |
| Above-the-fold messaging score | 4/10 | 8/10 | +4 points |
| CTA hierarchy score | 4/10 | 9/10 | +5 points |
| Trust signal score | 5/10 | 8/10 | +3 points |
| Visual hierarchy score | 5/10 | 8/10 | +3 points |
| Mobile experience score | 3/10 | 7/10 | +4 points |
polis.sh's 9.23% post-optimization CTR places it at 2.4x the SaaS industry median of 3.8%. That moves it from "slightly above average" to comfortably within the top quartile.
The most important column in that table is the one that isn't there: nothing else changed. No new traffic channels. No retargeting campaigns. No product updates during the measurement period. This was a design-only lift. CRO-driven redesigns deliver an average 223% ROI, and this case reinforces why: design changes are high-leverage because they compound across every visitor you already have.
How to Improve Landing Page CTR: A 15-Minute Self-Audit
You probably don't have the same page as polis.sh. But you almost certainly have some of the same problems. Here's a 15-minute self-audit based on the five changes above. Open your landing page in one tab and work through these:
1. Read your headline out loud to someone who doesn't know your product. Ask them: "What does this product do for me?" If they can't answer in one sentence, your headline is feature-focused when it should be outcome-focused. Rewrite it to complete the sentence: "This helps you [specific outcome]."
2. Count your CTAs above the fold. If there's more than one, pick the one that matters most and remove or visually de-emphasize everything else. One page, one primary action.
3. Check where your trust signals live. If social proof, user counts, or logos require scrolling to find, move at least one proof point adjacent to your primary CTA. The trust signal should be visible at the exact moment the visitor considers clicking.
4. Squint at your page. Literally. Squint until you can't read the text. Can you still see a clear visual path from top to CTA? If everything blurs into a uniform block, you need more whitespace and stronger visual hierarchy between sections.
5. Try signing up on your phone. Don't just look at the mobile layout. Actually tap the fields, type in them, and hit the button. If anything feels awkward, undersized, or requires precision tapping, you're losing mobile conversions.
This audit won't give you polis.sh's exact results. Every page is different. But these five checks address the five highest-impact design dimensions for impression-to-signup conversion, and they cost nothing but 15 minutes of honest evaluation.
What This Proves (and What It Doesn't)
This case study proves one thing clearly: design-led changes can produce significant, measurable conversion improvements without touching traffic acquisition, pricing, or product.
It doesn't prove that every page will see a 100% lift. polis.sh had specific problems and a baseline that left room for improvement. Your mileage depends on where your page starts and which dimensions are weakest.
What it does suggest is that if your signup conversion rate benchmarks fall near or below the 3.8% SaaS median and you haven't done a structured design audit, you're likely sitting on unrealized conversion potential. Not "maybe." Probably. Only 22% of businesses report being satisfied with their conversion rates, which means 78% know something is off but haven't systematically identified what.
That's the gap SiteCritic is built to close. If you want to see which of the 8 design dimensions your landing page scores lowest on, paste your URL into SiteCritic and get a scored, timestamped critique in under 60 seconds. No designer required. No guessing. Just specific, ranked feedback you can act on today, the same way polis.sh did.