Funnel

Per-step viewers and completers from PostHog, joined to the cohort lane order. Drop-off shows the ratio of viewers who reach the next step — not viewed − completed, which understates conversion on steps where users complete asynchronously.

Cohort

About to start Already on GLP-1

Window

7d 30d 90d

About to start · last 30 days · choice_selected[step_id=journey_status].choice = 'about_to_start'

Step	Viewers	Completers	View → next
Welcome `welcome`	133	133	76.7% −31
Sign In `sign_in`	102	1	2.0% −100
AI Consent `ai_consent`	2	2	0.0% −2
Name Input `name_input`	0	0	— 0
Name Affirmation `name_affirmation`	0	0	— +148
Journey Status `journey_status`	148	143	96.6% −5
Journey Affirmation `journey_affirmation`	143	137	95.1% −7
Gender Selection `gender_selection`	136	134	98.5% −2
Birthday `birthday_picker`	134	132	98.5% −2
Height `height`	132	132	100% 0
Weight `weight`	132	120	90.9% −12
Weight Goal `weight_goal`	120	118	92.5% −9
Goal Weight `goal_weight`	111	111	100% 0
Weight Pace `weight_pace`	111	109	100% +5
Weight Goal Affirmation `weight_goal_affirmation`	116	115	99.1% −1
Activity Level `activity_level`	115	112	100% +9
Notifications Permission `notifications_permission`	124	124	90.3% −12
HealthKit Benefits `healthkit_benefits`	112	105	0.0% −112
Tracking Style (flag: trackingStyleQuestion) `tracking_style`	0	0	— 0
Tracking Affirmation (flag: trackingStyleQuestion) `tracking_affirmation`	0	0	— +105
Plan Preview `plan_preview`	105	103	1.0% −104
App Store Review (flag: reviewPromptEnabled) `app_store_review`	1	1	100% 0
Completion `completion`	1	0	— —

Notes

View → next is the ratio of users who viewed this step and also viewed the immediately-next step in the cohort lane. The terminal step shows —. Steps where ≥95% of viewers progress render in slate; 80–95% in amber; below 80% in red.
Cohort split is derived from the user's onboarding.choice_selected event on the journey_status step (properties.choice). Users who drop off before that step are unattributed and fall out of both cohorts — their events appear under neither lane.
Numbers are de-duplicated by distinct_id, so a user re-entering the same step within the window counts once.