How to Track Onboarding Drop-Off: Find Where Users Quit and Why

The three metrics you need to separate

See the onboarding drop-off rate and funnel completion rate glossary entries for precise definitions and formulas.

Why it's a leading indicator

Users who experience friction in onboarding carry that friction forward. The chain is direct:

Drop-off → Activation rate → Retention → Revenue

Fixing onboarding early compounds everything downstream. If you retain 10% more users in week 1, that compounds every month.

How to measure it correctly

You need every step instrumented - not just start and end. Without step-level data you can calculate a total completion rate but you can't find where the problem is.

Event schema

Use one event name with a step property. Don't create one event per step - it makes queries complex and schema sprawl when steps change.

onboarding_started
  → { source, plan, device_type }

onboarding_step_completed
  → { step, time_spent_secs, attempt_count }

  step values:
    "email_verified"
    "profile_created"
    "integration_connected"
    "first_action_taken"
    "teammate_invited"

onboarding_completed
  → { total_time_secs, steps_completed }

activation_event          ← your product's aha moment

onboarding_step_abandoned ← optional, fire on page exit
  → { step, time_spent_secs }

Choose the right denominator

For step-level analysis, use users who entered that step - not total signups. A user who never reached step 4 shouldn't count as dropped off at step 4.

What a funnel actually reveals

Here's a real-looking example. The 47% drop-off at "Data source connected" is the answer - not the step with the second-highest rate, not the one a teammate thinks is the problem.

Setting up your funnel

Why users leave - what the data doesn't tell you

Funnel data tells you where users leave. These signals tell you why:

• Time spent per step - unusually long dwell time = confusion
• Error events - instrument validation failures and API timeouts as discrete events
• Session replays - one replay of a session that abandoned at step 3 is worth more than a week of data analysis
• Device split - 70% drop-off on mobile and 20% on desktop at the same step is a layout problem, not a copy problem
• Source channel - paid ads users often have different expectations than organic

Patterns to look for in replays

Diagnose before you fix

Most teams ship a redesign before validating their data. A surprising number of "drop-off problems" are tracking problems.

Run this triage in order - stop at the first positive:

Common causes at a glance

How to prioritize what to fix

Make each step end with visible progress. Users who see a result are more likely to continue:

• A preview of what their data will look like after connection
• A checklist that updates in real time as steps complete
• An inline confirmation showing the next step becoming available
• Social proof ("3 of 5 teammates have connected Slack")

Tools that cover different layers

Weekly review checklist

Run this every week. Pick one step to improve per sprint.

FAQ

What's a good onboarding drop-off rate?

There's no universal benchmark. A B2C app might expect 80% completion on 3 steps; a complex B2B SaaS might consider 40% on 7 technical steps strong.

The better question: is your completion rate improving week over week? And how does your activation rate correlate with completion? 40% complete + 95% retain beats 80% complete + 40% retain.

How many steps should you track?

Track every step a user must complete to reach your activation event. Don't track every UI click.

• Too few: you have high total drop-off but can't find where
• Too many: 12+ steps and most show under 5% drop-off individually

How do you confirm which step is the real problem?

Find the step with the highest volume of drops (not just the highest rate), then watch 10–20 session replays of users who reached that step but didn't complete it. Consistent pattern = UX problem. Inconsistent = expectation or value timing issue.