Skip to content

category-deep-dives.md

Latest commit

 

History

History
248 lines (186 loc) · 7.39 KB

category-deep-dives.md

File metadata and controls

248 lines (186 loc) · 7.39 KB

Category Deep Dives

Each category has its own economics. Don't apply one playbook to all apps. Data sourced from Adapty 2026 (16K apps, $3B), RevenueCat 2026 (115K apps, $16B), and AppsFlyer 2026 (1.7B installs, 2.9K subscription apps). Marked by Evidence Class.

Health & Fitness

Vendor Aggregate Data unless noted.

The numbers

Metric Value Notes
Annual share of revenue 60.6% Only category where annual dominates
Install LTV (12-mo) $1.21 Highest of any category
Install→trial (NA) 14.5% High-intent users
Trial→paid 35.0–42.2% Top of all categories
First renewal 67.7% Strong recurring
Day 0 conversion share 86.1% + secondary peak Days 4–7
Pricing variance across markets 4.4x DE/JP/CH high; TR/IN/ID low

What works

  • Long onboarding quiz (Noom 77 steps, Flo 70 screens)
  • "Personalized plan" framing (commitment + consistency)
  • Trial on annual only (annual is the dominant LTV product)
  • Geo-pricing (huge willingness-to-pay variance)
  • Authority signals (doctors, scientists, partnerships)

What hurts

  • Short trials (gaming-style ≤4 days under-converts)
  • Generic "Premium" framing (audience expects personalization)
  • Single-region pricing (leaves EU/JP money on table)

Reference apps

Noom, Flo, Calm, Headspace, Strava (see teardowns.md)

Gaming

The numbers

Metric Value Notes
Weekly share of revenue 82% Fast monetization model
Trial length used ≤4 days (73% of apps) Fast value moment
Download→trial (D30) 4.4% Lower base, but high RPI
RPI Day 14 $0.08 Low per-install but high volume

What works

  • Short trials matched to fast aha moment
  • Weekly billing (matches play sessions)
  • In-game currency / consumable IAP alongside subscription
  • Limited-time / event-tied offers (real expiry)

What hurts

  • Long trials (gamers don't extend)
  • Annual-first defaults
  • Heavy onboarding before play (kills install→trial)

Productivity

The numbers

Metric Value Notes
Monthly share 77% Recurring monthly tasks
Trial impact on LTV NEGATIVE: $56.95 direct vs $49.13 trial Counter-intuitive
Trial→paid (mid-range) But trial users worth less

What works

  • Direct-to-paid path (skip trial, save margin)
  • Reverse trial for low-intent
  • Monthly billing as default
  • Feature comparison tables (technical audience)

What hurts

  • Trials (subsidize lower-value users)
  • Aggressive social proof (audience is skeptical)
  • Per-day framing (audience does the math)

Reference apps

Notion, Things, Bear, Todoist (avoid trial heavy-handedness)

Lifestyle

The numbers

Metric Value Notes
Trial impact on LTV Direct buyers ~21% more valuable at 12mo Similar to Productivity
Mixed plan distribution Varies No dominant pattern

What works

  • Direct buyers preferred
  • Test no-trial direct-to-paid
  • Reverse trial as alternative

What hurts

  • Default trial heavy-handedness

Education

The numbers

Metric Value Notes
Trial users 12-mo LTV +50.4% vs direct Strongest trial uplift of any category
Annual median price $44.99 Highest of categories
Discount usage 14.3% (highest of all categories) Audience expects deals

What works

  • Trials strongly recommended
  • Discounts work (unique among categories)
  • Annual-first pricing
  • Authority signals (universities, certifications)

What hurts

  • Gaming-style short trials
  • No discounts at all (audience compares to competitors)

AI Apps

The numbers

Metric Value Notes
Monthly share 59.8% Higher than category-wide
Pricing tolerance $20/mo norm (ChatGPT anchor) Higher than non-AI
Churn rate ~30% faster than category avg Trial cycle, novelty
Trial rate 5.31% Lower than 10.92% average
Annual+trial LTV $66.70 Higher than $49.92 average

What works

  • High pricing tolerance — don't underprice
  • Annual + trial as the LTV anchor (not weekly)
  • Single-action focus (Cal AI: one button = analyze food)
  • Demo video at onboarding (show value in 10 seconds)
  • Geo-tier pricing (ChatGPT Go @ $8 for emerging markets)

What hurts

  • Optimizing for retention over first-month conversion (AI churns fast — accept it)
  • Underpricing relative to ChatGPT $20 baseline
  • Multi-step onboarding without showing AI in action first

Reference apps

ChatGPT, Cal AI, Photoroom, Replika, Character.AI

Photo & Video

The numbers

Metric Value Notes
Trial→paid 22.2% Lowest of categories
APAC refund rate 14.1% Highest regional refund
Mixed plans Varies

What works

  • Hybrid model (subscription + one-time export packs)
  • Sample edits before paywall
  • Watermark removal as primary trigger

What hurts

  • High refund risk in APAC (verify channel quality, payment fraud)
  • Trial without value demonstration

Travel

The numbers

Metric Value Notes
Trial→paid 43.5% Highest of categories
3+ plans 27% of apps Most diverse plan offerings
Annual median price $20 Lowest of categories

What works

  • Multi-tier (basic / premium / VIP)
  • Seasonal pricing
  • Trip-specific value (e.g., per-trip add-ons)

Shopping

The numbers

Metric Value Notes
1-plan paywalls 40% of apps Highest single-plan share
Mixed monetization High Sub + IAP + ads

What works

  • Single membership plan (Amazon Prime model)
  • Annual default
  • Free shipping / discount as benefit (not abstract value)

Business / B2B

The numbers

Metric Value Notes
Download→trial (D30) 9.1% Highest
12-mo RLTV per payer $35.48 Tied for highest
Time to $1K revenue 113 days median Longer sales cycle

What works

  • Sales-assisted enterprise tier
  • Free trial with credit card (filters intent)
  • Annual contract default
  • Per-seat pricing transparency

Geography Cuts

Conversion + RPI by region (RC 2026)

Region D30 Trial D35 Paid D14 RPI
North America 7.1% 2.8% $0.38
Western Europe (mid) (mid) (mid)
IN/SEA 3.0–3.7% 0.7% $0.08

Strategic implication

  • NA: high-intent, premium pricing, mature audience. Test for LTV not just conv.
  • WE: 29–39% pricing premium opportunity per Adapty. Manual per-territory pricing essential.
  • IN/SEA: Android-dominant, lower RLTV, growing fastest. Need geo-tier pricing (ChatGPT Go pattern). Per AppsFlyer 2026: Indian Subcontinent = 49% of net Android paid install growth.

When Benchmarks Don't Apply

  • Your N is small (<200 subs/variant) — vendor benchmarks are not statistical truth for your app
  • Your category is niche — H&F benchmarks don't apply to a chess training app
  • Your audience is specific — B2B SaaS benchmarks don't transfer to consumer mood tracker
  • Recent launch — first 90 days are high-volatility; don't optimize on that
  • Major UA channel shift — paid TikTok traffic ≠ organic Search traffic; benchmarks reflect channel mix

When in doubt: trust principles (placement before paywall, transparency, real social proof) over numbers.