Funnel Analysis - Finding Where Users Drop Off

This lesson on Funnel Analysis - Finding Where Users Drop Off is hands-on and example-driven. You will build and interpret funnel analysis charts in event analytics platforms like Amplitude to diagnose onboarding drop-offs. You will sequence ordered user milestones and configure conversion windows to isolate exact UX friction points.

What You'll Be Able To Do

• Construct a multi-step onboarding funnel by ordering sequential event milestones.
• Configure conversion window timeframes to evaluate drop-off under different completion constraints.
• Calculate step-by-step drop-off percentages and overall funnel conversion rates.
• Identify the exact onboarding stage responsible for the highest user abandonment.

Detailed Concept Walkthrough

1. Event-Based Funnel Sequencing

A funnel maps a chronological path of user actions to measure retention across successive milestones in a product flow.

• Mechanism: Funnel tracking requires users to execute an entry event (e.g., 'Welcome Screen') and subsequent defined actions in an exact chronological sequence.
• Under the Hood: Analytics engines evaluate user event streams, matching user IDs against ordered event timestamps to verify that Step N occurred strictly after Step N-1.
• Best Practice: Select discrete, high-intent milestones—such as Welcome, Registration, Playlist Follow, and Song Play—to establish a clean onboarding baseline.

-- SQL representation of a 4-step onboarding funnel sequence
WITH funnel_events AS (
  SELECT 
    user_id,
    event_name,
    event_time,
    ROW_NUMBER() OVER(PARTITION BY user_id, event_name ORDER BY event_time ASC) as step_instance
  FROM analytics_events
  WHERE event_name IN ('Welcome Page', 'Sign Up', 'Follow Playlist', 'Play Song')
)
SELECT event_name, COUNT(DISTINCT user_id) AS user_count
FROM funnel_events
GROUP BY event_name;

Key Takeaway: Funnel charts measure conversion integrity by requiring strict chronological completion of sequential product milestones.

2. Conversion Window Dynamics

The conversion window enforces a time ceiling within which a user must complete all funnel steps after entering Step 1.

• Mechanism: Setting a window (e.g., 1 day versus 7 days) defines the maximum elapsed time between the initial trigger event and the final target action.
• Under the Hood: If a user completes Step 2 after the window duration has elapsed from Step 1, the analytics engine marks that user as dropped off at Step 1.
• Best Practice: Align conversion windows with natural user behavior: use short windows (e.g., hours or 1 day) for single-session onboarding and longer windows (e.g., 30 days) for complex multi-session flows.

-- Filtering funnel conversions within a 1-day conversion window
SELECT 
  s1.user_id,
  s1.event_time AS step1_time,
  s2.event_time AS step2_time
FROM user_step_1 s1
LEFT JOIN user_step_2 s2 
  ON s1.user_id = s2.user_id 
  AND s2.event_time BETWEEN s1.event_time AND s1.event_time + INTERVAL '1 day';

Key Takeaway: Shorter conversion windows expose immediate single-session friction, while longer windows capture delayed multi-session completion.

3. Drop-Off and Step Conversion Interpretation

Analyzing conversion charts involves measuring both overall conversion from Step 1 to N and the relative drop-off between adjacent stages.

• Mechanism: Visual bar charts represent total retained users at each step, displaying relative step-to-step drop percentages alongside total funnel progression.
• Under the Hood: Drop-off rate at Step k is calculated as 1 - (Users at Step k / Users at Step k-1), exposing where user drop-off is highest.
• Best Practice: Prioritize UX optimization on the single transition step exhibiting the largest percentage drop rather than diffusing effort across healthy stages.

-- Calculating stage-by-stage drop-off percentages
SELECT 
  step_name,
  current_step_users,
  LAG(current_step_users) OVER (ORDER BY step_order) AS previous_step_users,
  ROUND(100.0 * (1.0 - (current_step_users::float / LAG(current_step_users) OVER (ORDER BY step_order))), 2) AS drop_off_pct
FROM funnel_step_summary;

Key Takeaway: Target product interventions at the specific funnel step that produces the steepest drop-off percentage.

Topics Covered in Funnel Analysis - Finding Where Users Drop Off

1. Onboarding Friction Problem (0:06 - 0:40) — Identifies the business challenge of locating user abandonment during app setup.
2. Sequencing Funnel Steps (0:41 - 1:04) — Demonstrates selecting and ordering AmpliTunes onboarding events in Amplitude.
3. Configuring Conversion Windows (1:05 - 1:48) — Explains how adjusting the conversion timeframe modifies completion calculations.
4. Interpreting Drop-Off Rates (1:52 - 2:19) — Analyzes the resulting bar chart to pinpoint the step with the highest drop-off.
5. Funnel Optimization Wrap-Up (2:20 - 2:37) — Summarizes how funnel charts guide targeted onboarding UX improvements.

Stats & Product Analytics for Analysts Cheat Sheet

• Funnel Step Definition — Sequences chronological events to track user flow progression

  funnel = ['Welcome Page', 'Sign Up', 'Follow Playlist', 'Play Song']
  

• Conversion Window Setting — Sets maximum elapsed duration allowed between first and final steps

  conversion_window = timedelta(days=1)
  

• Step Drop-off Calculation — Calculates percentage of users lost between two consecutive steps

  drop_off_rate = 1.0 - (users_step_2 / users_step_1)
  

• Overall Conversion Rate — Calculates percentage of initial users who complete final milestone

  overall_conv = (users_step_final / users_step_1) * 100
  

Comparison Table

Common Pitfalls

• Mistake: Using overly long conversion windows for simple onboarding flows. Avoid: Set a short window such as 1 day to measure single-session completion friction.
• Mistake: Focusing only on overall conversion rate rather than step-by-step drops. Avoid: Inspect intermediate step percentages to locate the exact point of drop-off.
• Mistake: Out-of-order event instrumentation in the funnel definition. Avoid: Verify that step events follow the true required user journey sequence chronologically.

FAQs

• What happens if a user completes step 3 before step 2 in an onboarding funnel? The user is not counted as completing step 3 until they have completed all preceding steps in sequential order.
• Why does increasing the conversion window increase the funnel conversion rate? A longer window grants users more time to return and complete subsequent steps, capturing delayed conversions.
• Does funnel analysis show why users drop off? No, funnels identify where drop-offs happen; root cause analysis requires further qualitative review or cohort segmentation.