We would essentially generate the right cohorts, suggest sensible metrics, and propose lines of inquiry when someone searches for something like "What's a good experiment to test if onboarding is too long?" so users don't have to define everything upfront. It brings intuition and structure to product thinking, right inside PostHog.
PostHog Assistant
What's a good experiment to test if onboarding is too long?
Let me analyze your current onboarding flow data and suggest experiments to test if it's too long.
Analysis Complete: Based on your onboarding metrics over the past 60 days, I've designed several potential experiments to test if your onboarding process is too long.
Current Onboarding Metrics
Onboarding Completion
47%
Completion Rate
5.2
Avg. Steps Completed
8.7 min
Avg. Completion Time
5.3% decrease compared to previous period
Drop-off Analysis
Highest drop-off points:
Step 4: Team Invitation (32% drop)
Step 6: Integration Setup (24% drop)
Recommended Experiments
Experiment 1: Strategic Step Elimination
Skip Team Invitation During Initial Onboarding
Test removing the team invitation step (highest drop-off point) from initial onboarding and moving it to post-activation.
Hypothesis: Moving team invites out of the critical path will improve completion rates by 25-30% without negatively impacting team collaboration metrics.
Recommended Test Parameters:
Test Group: 25% of new signups
Control: Current onboarding flow
Primary Metric: Onboarding completion rate
Secondary Metrics: Time to first value, 7-day activation rate, 14-day team size
Minimum Sample Size: 1,200 users (600 per variant)
Estimated Runtime: 14 days based on current signup volume
Experiment 2: Progressive Disclosure
Staged Onboarding vs. Complete Setup
Test a minimalist "core-only" onboarding against the full setup process. Focus on getting users to first value faster.
Hypothesis: A shorter initial onboarding focused only on core features will improve activation rates by 35% and increase long-term retention.
Implementation Details:
Reduce onboarding from 7 steps to 3 essential steps
Move advanced configuration to contextual prompts after initial value moment
Introduce gamification elements to encourage completing deferred setup steps
Experiment 3: Personalized Paths
Role-Based Onboarding Optimization
Create multiple onboarding paths tailored to different user personas, based on role selection at signup.
Hypothesis: Personalized onboarding paths will improve completion rates by 40% and increase feature adoption relevance.
Proposed Paths:
Developer Path: Focus on SDK implementation and event validation
Product Manager Path: Focus on insights and experiment setup
Marketing Path: Focus on funnel analysis and attribution
Implementation Complexity: High (requires significant frontend changes)