Speakfloe Onboarding Journey

Customer journey map · Focus: Discovery → Activation · Data as of March 2026

Week-1 Retention
~1.8%
Category benchmark: 25–40%
Onboarding Steps
~50
Target: 5 steps to first conversation
Payment Conversion
0.65%
28 purchases / 4,283 who see paywall
MAU Growth
+173%
8,056 MAU in 28 days

Click any stage card to expand details and recommendations

1
Discovery

User Discovers Speakfloe

Users find Speakfloe through app store search, social ads, referral links, or NGO/institutional partnerships. India (24.6%), Vietnam (19.7%), and LatAm are primary markets.
App Store / Play Store Social Ads Referral Links NGO Partners
🤔
Curious

Pain Points

  • Referral traffic has $0.0004 LTV — essentially zero-value installs
  • NGO traffic may inflate MAU without real engagement intent
  • Users in India/Vietnam are highly price-sensitive ($9.99/week is steep)

Opportunities

  • Kill referral spend, reallocate budget to higher-intent channels
  • Audit NGO cohort quality — separate organic vs. institutional MAU
  • GDP-adjusted pricing ($1/week trial) could dramatically improve checkout rates
✂️ Kill referral spend 🔍 Audit MAU quality
8,056 MAU
+173% growth
$0.0004 referral LTV
↓Unknown % install → open (measure this)
2
Onboarding

50-Step Onboarding Gauntlet

Users hit an extremely long onboarding flow with ~50 tracked steps. Most users fire ob35_assessment_call but only 2% confirm the assessment. This is the primary kill zone — users abandon before ever experiencing the core product.
Profile Setup Level Assessment Goal Selection ob35 Assessment Call ~45 more steps…
😩
Frustrated

This is the #1 Problem

50 steps before the user ever speaks to the AI tutor. By the time they reach the assessment call, motivation is gone. Only 2% confirm — meaning 98% of users who get this far still bounce. The value proposition ("practice English with AI") is buried behind bureaucracy.

Recommended Fix: 60-Second-to-Value

  • Step 1: Name + native language (5 sec)
  • Step 2: "What do you want to practice?" — 3 quick-pick options (10 sec)
  • Step 3: Immediately start a free AI conversation (45 sec)
  • Step 4: After 2-min conversation, show level result + paywall
  • Step 5: Account creation (defer until value is proven)
🔥 P0 — Slash to 5 steps ⏱️ 60-sec to first conversation 📊 Instrument each step
~50 onboarding steps
2% assessment confirm
98% abandon before value
↓~98% drop-off — users never reach the AI tutor
3
Aha Moment

First AI Conversation

The tiny fraction who survive onboarding have their first AI tutoring conversation. This should be the "aha moment" — the point where users feel the product's value. Currently, almost no one gets here fast enough.
AI Conversation UI Voice Input Real-time Feedback Level Detection
😊
Excited

Why This Stage Matters

Users who have a real AI conversation are significantly more likely to convert and retain. The product experience is strong — the problem is almost nobody reaches it. The entire onboarding redesign should be oriented around getting users here in under 60 seconds.

What to Measure

  • Time from app open → first AI utterance
  • Conversation completion rate (did they finish a 2-min session?)
  • Sentiment after first conversation (in-app micro-survey)
  • Correlation between conversation count and Day-7 retention
📐 Measure time-to-first-convo 🎯 This IS the activation event
??? reach first convo
100% crash-free
2.1% DAU/MAU stickiness
↓Low stickiness — users don't return after first session
4
Paywall

Payment Decision Point

4,283 users see the payment page. Only 766 begin checkout (17.9%), and only 28 actually purchase (3.7% of checkout starters). Current pricing: $9.99/week with 7-day free trial, or $29.99/quarter.
Paywall Screen Price Display Free Trial CTA RevenueCat
😬
Hesitant

Conversion Breakdown

  • 4,283 see the paywall
  • 766 begin checkout (17.9% — actually decent)
  • 28 complete purchase (3.7% of checkout — payment friction or price shock)
  • $30.75 avg purchase, $861/mo total revenue

Two Separate Problems

Problem A: 82% don't even start checkout. They haven't experienced enough value yet (see Stage 2/3). Problem B: Of those who do start, 96.3% abandon at payment. $9.99/week ($40/mo) is very high for India/Vietnam/LatAm markets. The GDP-adjusted $1/week test should address this directly.

💰 Test $1/week in target markets ⏰ Move paywall AFTER first conversation 🧪 A/B test trial length
4,283 see paywall
17.9% begin checkout
0.65% overall conversion
↓96.3% abandon during checkout — price shock in target markets
5
Activated

Paying User — Now Retain Them

28 users have purchased. 18 have renewed at least once. Week-1 retention across all users is ~1.8%. The users who do pay need habit-forming loops to stick — or they'll churn at renewal.
Daily Practice Reminder Progress Dashboard Streak System Subscription Renewal
🙂
Cautiously hopeful

Retention Signals

  • 18 of 28 purchasers renewed — 64% renewal rate is actually promising
  • But Week-1 retention at 1.8% means almost nobody hits the renewal window
  • Sonja's suggestion: interview paid users (offer 3-month free incentive) to understand what's working

Retention Levers to Build

  • Daily streak with push notifications (time-of-day optimized per market)
  • Weekly progress email showing improvement metrics
  • Conversation variety — new scenarios daily to prevent repetition fatigue
  • Social proof: "Users like you improved X% in 2 weeks"
📞 Interview the 28 paid users 🔔 Build habit loops 📈 Track Day-1/3/7 retention separately
28 total purchases
64% renewal rate
1.8% Week-1 retention

Onboarding Funnel — Current State

MAU (app opens)
8,056
8,056
See paywall
4,283
53.2%
Begin checkout
766
9.5%
Purchase
28
0.35%
Renew
18
0.22%

Priority Fixes — Sequenced

P0 — This Week

Nuke the 50-Step Onboarding

Reduce to 5 steps. Get users into a free AI conversation within 60 seconds of app open. Defer account creation, assessment, and profile setup until after the user has experienced value. This single change likely moves Week-1 retention from 1.8% to 10%+.

Expected impact: 5–10× improvement in activation rate
P0 — This Week

Move Paywall After First Conversation

Don't show pricing until the user has had their "aha moment." Current flow: onboarding → paywall → value. Flip to: minimal onboarding → value → paywall. Users who experience the product convert at dramatically higher rates.

Expected impact: 2–3× checkout start rate
P1 — Next 2 Weeks

GDP-Adjusted Pricing Test

$9.99/week is $40/month in markets where monthly income is $200–500. Test $1/week in India, Vietnam, Colombia, Peru. Even if conversion rate only doubles, the volume increase in these markets should net positive.

Expected impact: 3–5× conversion in target markets
P1 — Next 2 Weeks

Interview Paid Users

You have 28 people who paid despite every obstacle. They're gold. Sonja's plan: offer 3 months free in exchange for a 20-minute interview. Understand what made them convert and what almost stopped them. Build the onboarding around their answers.

Expected impact: Qualitative data to inform all other fixes
P2 — Month 2

Build Retention Loops

Streaks, push notifications, weekly progress emails, new daily conversation scenarios. Focus on Day-1 and Day-3 retention specifically — if users come back twice in the first week, long-term retention jumps dramatically.

Expected impact: 2–3× Day-7 retention
P2 — Month 2

Instrument the Full Funnel

You have good payment funnel data but gaps in onboarding-to-activation tracking. Add events for: time-to-first-convo, conversation completion, return-within-24h. You can't improve what you can't measure.

Expected impact: Data foundation for all future optimization