Case 8: Traffic Quality Persistence Analysis

Case 8: Traffic Quality Persistence: Analyzing Multi-Step Conversion

Context

After finding campaign winners at step 1 (landing → achieve_screen), check if quality persists

The Advanced Funnel View

From filtered winner segment:

Step	Event	Persons	Conversion	Drop-off	Time
1	`achieve_screen_loaded`	7,855 (100%)	-	-	-
2	`achieve_screen_viewed`	4,996 (63.6%)	63.6%	2,859 (36.4%)	9s
3	`sex_screen_loaded`	4,517 (57.5%)	90.4%	479 (9.6%)	14s

Overall Funnel Performance

Total conversion rate: 30.85% (much better than baseline ~4-10%)
This is the “mio moi” filtered view showing good traffic

The “Loaded vs Viewed” Bounce Effect

Key Observation at Step 1→2

achieve_screen_loaded (page loaded in browser): 7,855 people
achieve_screen_viewed (user actually engaged with page): 4,996 people (63.6%)
36.4% drop-off between page loading and viewing

What This Means

This is a “micro-bounce” within the same page
User’s browser loaded the page, but they didn’t engage with it
Could be: page closed immediately, tab closed, bounced before interaction

The Critical Insight: Same Page, Different Traffic Sources

The Test

Everyone at achieve_screen sees the EXACT SAME screen
The screen doesn’t change based on which landing page they came from
Yet conversion rates differ based on original traffic source

Findings

When you filter by good campaigns (like campaign 1202363… + affid 1000):

Loaded → Viewed conversion is likely HIGHER than 63.6%
These users engage more with the content

When you filter by bad campaigns:

Loaded → Viewed conversion is likely LOWER than 63.6%
These users bounce even from the second screen

Since the screen is identical for everyone, the difference must be:

Audience quality: Who you’re targeting
User intent: How motivated they are
Expectation match: Whether your ad promised what the page delivers

The Conclusion: It’s an Audience Problem

Evidence

✅ Different campaigns perform differently on same landing page (5% vs 44%)
✅ Good traffic outperforms at step 1 AND step 2 (not just one step)
✅ Same page shows different conversion based on traffic source
✅ Bad traffic performs badly on ALL landing pages

Diagnosis: Primary issue is audience quality/targeting, NOT page design

Why This Matters

If you had assumed “low conversion = bad landing page”

You’d spend weeks redesigning pages
Testing new layouts, copy, colors
Optimizing load times
Results: Minimal improvement (because traffic is bad)

Now you know “low conversion = bad traffic”

Fix targeting on campaigns
Adjust ad creative to attract right users
Pause underperforming campaigns
Scale winning campaigns
Results: 4-5x improvement immediately

Actionable Next Steps

Priority 1: Traffic/Audience Optimization

Audit all campaigns for targeting quality
Compare ad creative between winners (44%) and losers (5%)
Analyze audience characteristics of high-converters
Replicate winning campaign targeting
Pause or retarget low-quality traffic sources

Priority 2: Landing Page Optimization (Secondary)

After fixing traffic, THEN optimize pages
Even bad pages convert at 40%+ with good traffic
But good pages can push good traffic from 44% → 60%+

The Universal Lesson

Good traffic + mediocre page > Bad traffic + perfect page

Always fix traffic quality first. Page optimization is the multiplier, not the foundation.

Principles Applied

Thinking Flow Applied

Advanced Isolation Method ✓
Multi-Step Validation ✓
Hold page constant, vary traffic → Find massive difference ✓
Check subsequent steps → Confirms traffic quality issue ✓

Tool Features Enabling This Analysis

Multi-step funnel with “loaded” vs “viewed” event tracking
Ability to filter entire funnel by campaign/affid
Conversion rates recalculated for filtered segments
Step-by-step drop-off visualization
Time between steps tracking (9s from loaded to viewed)
Person counts at each step (not just percentages)

Previous: Case 7: Campaign Winners
Next: Case 9: Multi-Touch Attribution
Back to Index: All Cases