Skip to content

Practical Examples & Case Studies

This section contains real-world examples showing how to analyze marketing data, build funnels, investigate problems, and make data-driven decisions. Each case demonstrates specific principles and workflows in action.

Automatic capture of all landing pages without manual configuration. Shows how PostHog tracks every page users land on, not just designated landing pages.

Key Learning: Never need to ask “which pages should I track?” - the system shows everything automatically.


Natural exploration flow from aggregate data to individual user sessions. Click through from landing page metrics to see who visited and watch their session recordings.

Key Learning: Tools should support natural curiosity - clicking on data points should reveal the people behind the numbers.


Deep dive into user profiles to discover properties, events, and user journeys. Learn the funnel structure by exploring real user event streams.

Key Learning: Don’t ask developers “what events do we have?” - just look at user data to discover the structure.


Creating funnels from discovered events and immediately spotting problems. Build funnels with events you discovered yourself, spot 90% drop-off instantly.

Key Learning: The funnel will scream problems at you - 90% drop-off is not subtle.


From aggregate numbers (50,266 drop-offs) to individual investigation and pattern discovery. Every person in the drop-off has complete data you can investigate.

Key Learning: Numbers aren’t mysteries - 50k drop-offs means 50k specific people you can investigate.


Segmentation revealing massive performance differences: 22% vs 0% conversion rates. Pre-quiz pages convert 4.7x better than baseline.

Key Learning: Segmentation turns vague problems (“4.71% conversion”) into actionable insights (“pre-quiz pages convert at 22%”).


Multi-dimensional filtering finds 44% conversion rate - 4.5x better than baseline. Layer filters (landing page + affid + campaign) to isolate winners.

Key Learning: Same landing page performs differently based on traffic source. It’s about who you target, not just what you built.


Same page, different traffic = different results. Proves it’s an audience problem, not a page problem.

Key Learning: Good traffic converts well everywhere. Bad traffic struggles everywhere. Fix traffic first, optimize pages second.


Attributing sales across multiple touchpoints - paid ads, SMS, email retargeting. 76% of sales require retargeting, but all attribute back to original campaign.

Key Learning: Always attribute to first-touch (original campaign), not last-touch. Without that original ad, you wouldn’t have leads to retarget.


From question to answer in 1 minute - understanding event footprints and parameters. Reading parameters reveals SMS retargeting disguised as “Google Ads” sale.

Key Learning: Parameters are your attribution Bible. Surface labels lie, parameters tell the truth.


Real Slack conversation discovering which tracking properties are reliable. referring_domain wins (automatic), utm_source fails (91% null).

Key Learning: Not all tracking properties are equally reliable. Test what exists, use what works, ignore what doesn’t.


  1. Start with workflows that match your needs: Looking to understand funnels? Start with Cases 4-5. Need attribution? Go to Cases 9-11.

  2. Follow the natural progression: Cases 1-3 build foundation skills. Cases 4-8 show analysis techniques. Cases 9-11 demonstrate attribution.

  3. Reference principles: Each case links to specific principles from the Core Principles page.

  4. Apply to your data: These aren’t just examples - they’re workflows you can replicate in your own analytics.

Have a real-world example that demonstrates these principles? Document it using the same format:

  1. Context: What was the starting situation?
  2. Process: What steps did you take?
  3. Discovery: What did you find?
  4. Insight: What did you learn?
  5. Principles: Which principles did this demonstrate?