None of them pay. Most won't respond to emails.
Here's the data and what I'm missing.
THE NUMBERS
Revenue: 35,238 installs over 3 years, $86 MRR from 20 subscribers, $4,091 lifetime revenue, 0.8% free-to-paid conversion, $905/mo burn (day job subsidized).
Retention: D1 4.2% (industry 40%+), D7 1.1% (industry 25%+), D30 0.9% (industry 15%+), 130% monthly churn.
The Paradox: 9 users reached Level 100+ (10,000+ XP). 6 still active, using it daily. 1 interviewed (PhD + PM): "I'd cry" if app disappeared. Payment status: 0 paying, 4 on expired promos, 1 can't pay (SWIFT blocked).
POWER USER ANALYSIS
Analyzed usage patterns for 3,000+ users by calculating coefficient of variation (CV = std dev / mean of daily completions).
Success archetype (User A):
- CV 0.33 (very consistent daily pattern) - 580 completions/month - 465 days active - Regular professional schedule - High cognitive load (PhD + PM) - Job-to-be-done: "Eliminate decision fatigue"
Failure archetype (User B):
- CV 0.86 (boom-bust cycles) - Still paying subscription - Stopped using 7 months ago - Quote: "Life is too irregular, alarms ring at wrong time"
WHAT I TRIED
Retention fixes (none moved D1 from 4.2%):
- Rebuilt onboarding 4 times (current: v4, takes users through creating first ritual + habit + completing it once) - Added AI coaching (GPT-5, personalized habit advice, agentic tool calls) - Added gamification (quests, XP, levels, streaks) - Built WearOS app (rituals on your wrist) (more for myself as it was half a day of work, but still) - Added analytics dashboard - Daily notifications - Early wins in first session (complete habit in onboarding), streaks
Features built that didn't help:
- AI chat (11.54% conversion when discovered, but only 2.4% of users find it despite being literally in the bottom navigation bar) - Journaling - Habit timers with on-screen overlays - Streak widget with goals and focus on small wins/future progress
Interview attempts:
- Emailed all 9 power users for 15-20 min interviews - Offered 6-12 months free Pro - Got 2 responses after multiple attempts - User C (Level 177, active daily): no response - User D (Level 148): no response - User E (Level 236, ADHD coach, agreed to call): ghosted - 3 others: no response after 2 weeks
Paying customer analysis:
- Queried DB: only 15% of paying customers match "power user archetype" (CV <0.5 or Level 100+) - 4 of 5 known power users DON'T pay, the other pays but doesn't use the app. - Pattern: random trial conversions, not established users - 30% of subscribers have 0 activity (forgot to cancel)
Cancellation reasons (analyzed all churned subscribers):
- 43% "don't use it enough" (avg 63 days before cancel) - 30% cost concerns (avg 33 days before cancel) - 12% found better app (avg 19 days before cancel)
WHAT I THINK I KNOW
1. Product works for people with regular schedules + high cognitive load 2. Fixed-time alarms break when life gets irregular 3. Power users love it but won't pay or engage 4. Random people pay briefly then churn 5. Feature additions don't fix retention (D1 stayed at 4.2% through AI, quests, WearOS, etc.) 6. No monetizable pattern after 3 years
THE QUESTION
Not asking what to do. Asking: What am I blind to?
I have extreme PMF for 9 people who won't monetize. Is this:
- A pricing problem? (tested $1-6/mo PPP) - A positioning problem? (marketed as general habit tracker, should niche down?) - A feature problem? (wrong features locked behind paywall?) - Unfixable? (category too hard to monetize - results take weeks, requires discipline users lack)
Happy to share queries, GA4 exports, whatever helps.
Nek_12•8h ago
Tech stack: Kotlin Multiplatform, Compose Multiplatform, targeting Android/iOS/WearOS. 99% crash-free rate, 4.2* rating. The app works - users just don't stick or pay.
The regularity pattern (CV) analysis was eye-opening: active power users have CV <0.5 (consistent daily routine), churned power users have CV >0.7 (boom-bust). But when I analyzed paying customers, only 15% match this pattern. Most revenue comes from people who try it briefly and churn within 1-2 billing cycles.
If anyone wants to dig into the data, I'm happy to share: - Anonymized GA4 exports - SQL queries for usage pattern analysis - Onboarding flow breakdowns - Retention cohort data
Posting this because after 3 years I'm stuck between "one more pivot" and "shut it down." Hoping someone sees what I'm missing.
leephillips•8h ago
At least one super-famous bestselling author tried to serialize his novel online with voluntary payments, and had to abandon the experiment. Nobody payed.
Nek_12•8h ago
I even ran an A/B test (hard vs soft paywall):
- Started by "recommendation" from popular entrepreneur - 3 months, 3k users - Baseline (soft paywall, can skip): 0.8% conversion - Hard paywall (must activate trial): 0.6% conversion - Revenue tracking broken (Firebase showed $0 both variants) - Killed experiment yesterday.
leephillips•8h ago
They have a bunch of features that I don’t care about and never use (AI, their web browser, ...). I just need their core search product (because Google is useless). Instead of putting features behind a paywall, they offer search free of charge up to a certain number of searches per month (I think). It’s enough to learn if their search is good enough to pay for. I pay to continue to use their core service more frequently (and I guess it’s the same with most of their paying customers).
Your customers seem to find value in your core service but don’t want to pay for the extras. Maybe you could apply the Kagi model?
Nek_12•8h ago
Kagi model applied to Respawn would be: free tier gets 50-100 completions/month. If you're crushing it like the Level 177 user (800/month), you pay to keep using the core.
This makes more sense than what I'm doing. I'm trying to monetize peripherals when the core is what they value.
The risk: does limiting completions kill the habit-building momentum? If someone hits their limit mid-month, breaks their streak, do they just churn instead of convert?
But worth testing. It's at least aligned with value - heavy users pay, light users stay free.
leephillips•8h ago
jasonthorsness•7h ago
Nek_12•7h ago
The Kagi model Lee mentioned above would flip this: free tier is limited (50-100 completions/month), heavy users pay to continue. Then paying makes you smart (you're getting 800 completions/month for $5), not a sucker.
But that requires confidence the core is worth paying for. And based on my conversion data (0.8%), most people try it and decide it's not.
So maybe the real issue is: the core isn't valuable enough to limit. And if I can't confidently say 'this is worth paying for after 100 uses,' then I don't have a product.
jasonthorsness•7h ago