SKY Labs Experiment Format
This is a controlled A/B test comparing two distinct ad placement strategies across blog content to measure impact on ad revenue and user engagement.
Objective
To determine the optimal ad placement for blog content by measuring:
- Click-through rate (CTR): Percentage of ad impressions that receive clicks
- Revenue per thousand impressions (RPM): Overall revenue efficiency
- User engagement impact: Bounce rate, time-on-page, scroll depth
- Device-specific performance: Mobile vs desktop differences
- Content type variation: How-to articles vs listicles vs deep dives
Hypothesis: Mid-content ads will have higher CTR due to engaged readers, but above-content ads may capture more impressions from quick-scanners.
⚙️ Setup
Test Articles
20 blog posts (10 per variant)
Total Pageviews
156,892 (6 weeks)
Ad Type
Display ads (responsive)
Split
50/50 A/B test
The Two Placements:
Placement: First ad appears immediately below title, before article text begins. Second ad at bottom after content.
- ✓ Highest viewability (100% above fold)
- ✗ Potential user annoyance
Placement: First ad appears after ~40% of content (natural break). Second ad near end.
- ✓ Engaged reader context
- ✗ Lower viewability (requires scrolling)
📊 Results Overview
Click Heatmap by Scroll Depth
Key insight: Mid-content ads (40-60% scroll depth) received 41% of all clicks despite only 24% of total impressions.
What Worked
- Mid-content ads: 36% higher CTR and 44% higher RPM than above-content placements
- Natural breakpoints: Ads placed between sections (after H2, before next section) performed best
- Mobile mid-content: CTR on mobile was 2.8% (vs 1.9% desktop) for mid-content — mobile users more tolerant
- Second ad near end: Captured users who finished articles (lower CTR but high relevance)
What Didn't Work
Assuming "above the fold" always wins — it doesn't for engaged content.
- Above-content ads: Higher impressions but lower intent — users hadn't committed to reading yet
- Multiple ads above fold: Tested 2 ads above content on some pages — CTR dropped 52% (ad blindness)
- Automatic mid-roll without break: Ads that interrupted sentences had 63% lower CTR
- Square ads in mobile: Performed worse than responsive banners (user feedback)
Key Learning
For blog content, mid-content ads outperform above-content ads by 36-44% because they reach users who are already engaged with the content. The key is placement at natural reading breaks.
- Engagement threshold: Users need to commit ~30-40% of article before ad receptivity peaks
- Mobile opportunity: Mobile users are 47% more likely to click mid-content ads than desktop users
- Content adjacency: Ads next to relevant content sections (e.g., tools, recommendations) perform best
- Ad density: One mid-content ad + one end-of-content ad optimized RPM without annoying users
Action Taken
Based on 6-week A/B test:
- Removed above-content ads from all blog posts (moved to mid-content)
- Implemented smart placement: Ads appear after H2 headings (natural breaks)
- Mobile optimization: Larger tap targets, simplified ad units for mobile
- Dynamic insertion: Longer articles (>1500 words) get second mid-content ad
- User experience guard: Max 2 ads per article, never interrupt sentences
Result: 3 months post-implementation: Overall RPM increased 38%, user engagement (time-on-page) unchanged, bounce rate decreased 4% (fewer annoyed users).
✅ Conclusion
This experiment proves that for content sites, ad placement strategy should prioritize user engagement over raw impressions. Mid-content ads captured higher-intent users and generated significantly more revenue without harming user experience.
Validated
Mid-content ads outperform above-content by 36% CTR and 44% RPM on blog content.
Practical Rule
Place primary ad after 30-40% of content at a natural break (after H2, between sections).
Warning
Don't interrupt reading flow. Test placement with real user feedback.
Data Transparency: Full dataset available on request. Statistical significance at 99% confidence for CTR difference. All tests followed AdSense policies.