LTV per Download

The Pour Over App · value of a single install · cohort-survival model with power-law retention extrapolation.

LTV per Download (all-traffic) Q3 Rocks · Paid cohort →

Sponsor: May 18 – Jul 12 (8 wks) · Upgrade: trial-era scope (§2)

Method v2 (Jul 14) — trial funnel live. The Jul 9 paywall update made the $48 annual sub trial-first (3-day free trial, first starts Jul 8, RevenueCat) and repriced monthly to $9.99 ($5 legacy). Annual direct purchases collapsed (18→5→0/wk) as trials took over, so the old blended upgrades ÷ new users window rate is stale. Upgrade LTV now uses recent trial-era rates: trial starts × new-install share × trial conversion → annual, plus recent monthly/lifetime rates. 85% of trial starts come from July-installed users (RevenueCat install-month segment) and 75% of upgraders convert on install day (GA4 days-to-upgrade), so per-download attribution is legit. Trial → paid conversion locked at 58% (first resolved cohort: 19 starts → 11 converted, 7 expired, 1 pending). Sponsor window extended to May 18 – Jul 12; Jul 6–12 ran zero app sponsor placements (0 windowed clicks), dragging Sponsor LTV $0.85 → $0.75. Caveat: one resolved trial cohort; annual renewal (78%) unproven; 16% monthly churn measured on the $5 cohort.

Total LTV / download

$0.00

Sponsor + Upgrade value of one install

Sponsor LTV

Newsletter sponsor clicks over lifetime

Upgrade LTV

$0.00

Annual + monthly + lifetime conversions

Trial funnel

—

start rate × trial→paid conversion

Recent upgrade rate

—

Any-paid per new user · active scope in §2 (incl. expected trial conversions)

Expected active days

Area under retention curve, D0 → horizon (5 yr)

0 Week-over-week · rolling-window LTV blended value of the whole window each refresh

Refreshed	Window	Wks	Total LTV	Δ WoW	% Δ

Rolling cumulative window (May 18 → present), 5-yr horizon + real monthly churn. Each point is the whole window's blended LTV, not the single newest week — so movement reflects mix shift, not one week's performance. Method break at Jul 14: the 3-day annual trial launched Jul 6 made pre-trial blended rates stale, so Jul 14 onward uses trial-era rates (method v2: trial starts × 85% new-install share × 58% conversion). The Jul 14 point is the Jul 6–13 trial-era scope ($5.85); the hotter last-4-day scope (Jul 11–14, $10.42) is the page default in §2 but too short/partial for this series.

1 Window inputs May 18 – Jul 12 · sponsor side

Active users (window)avg WAU — stock, not summed · (8,254 + 7,554 + 5,958 + 6,297 + 6,122 + 6,209 + 6,002 + 6,339) ÷ 8

Sponsor clicks (window)Edition-windowed app clicks · 441 + 326 + 297 + 186 + 409 + 434 + 58 + 0 (Jul 6–12: zero app placements ran)

New users (window)Downloads / first opens · 2,904 + 1,068 + 1,018 + 1,311 + 994 + 1,104 + 1,014 (Jun 29–Jul 5)

Annual upgrades — windowpre-trial blended comparison only · 14 + 10 + 13 + 13 + 11 + 18 + 14

Monthly upgrades — windowpre-trial blended comparison only · 0 + 0 + 2 + 7 + 10 + 12 + 13

Lifetime upgrades — windowpre-trial blended comparison only · 0 + 2 + 1 + 1 + 0 + 0 + 0

Days in window56 = May 18 – Jul 12 (8 weeks) · validate $2.26 with one week (days=7) + horizon=365

Lifetime horizon (days)how far the retention curve is integrated · 1825 = 5 yr · 365 = 1 yr

2 Trial funnel · RevenueCat feeds Upgrade LTV

New users — scopeGA4 newUsers · Jul 11–14 = 582 (137+155+172+118, Jul 14 partial) · Jul 6–13 = 1,301

Trial starts — scopeRevenueCat trials_new · Jul 11–14 = 10+10+11+19 = 50 · Jul 6–13 = 40 · 3-day trial on annual (Jul 9 paywall update, first starts Jul 8)

New-install share of trials85% — RevenueCat install-month segment: 50 of 59 starts are Jul '26 installs; base-driven 15% excluded from per-download credit

Trial → paid conversion58% — LOCKED assumption (Matt, Jul 14). Basis: first resolved cohort, 19 starts → 11 converted, 7 expired, 1 pending (61% excl. pending). Revisit as cohorts resolve

Direct annual purchases — scopeRevenueCat direct P1Y subs (no trial) · Jul 11–14 = 0 · Jul 6–13 = 5 · trial now gates annual

Monthly upgrades — scopeRevenueCat direct P1M subs · Jul 11–14 = 4 · Jul 6–13 = 20

Lifetime upgrades — scopeRevenueCat non-subscription purchases · zero in both scopes

3 Retention anchors Firebase / GA4

D1 retentionlive Q2 cohort (Apr 1–Jul 7, n=18,682) · % active 1 day after install

D30 retentionlive Q2 cohort · % active 30 days after install

2 Pricing & renewal

Sponsor CPC ($/click)

Annual sub price$48 — current US store price per RevenueCat offering (was $49)

Monthly sub price ($/mo)$9.99 since the Jul 9 paywall update ($5 legacy subs keep their price) · new downloads pay $9.99 · churned on its own rate below

Monthly churn (per month)16% — RevenueCat: 9 of 55 active monthly subs set-to-cancel (Jul 9 2026, n=55). Measured on the $5 cohort — watch whether $9.99 churns hotter. Monthly LTV = price ÷ churn

Lifetime sub price

Annual renewal rate78% — internal Stripe: TPO premium newsletter subs churn ~2%/mo → ~78.5% annual retention. RevenueCat snapshot: ~83% of annual subs set-to-renew, so 78% is conservative

4 Derived parameters

Decay exponent kR(d) = D1 · d^(−k), fit from D1 & D30

—

Sponsor rev / active user / day(clicks ÷ AU ÷ days) × CPC

—

Annual sub LTVgeometric: price ÷ (1 − renewal)

—

Monthly sub LTVgeometric: price ÷ monthly churn

—

Lifetime sub LTV

—

Trial start ratetrial starts × new-install share ÷ recent new users

—

Effective annual conversiondirect annual rate + trial start rate × trial CVR

—

Monthly conversion raterecent monthly upgrades ÷ recent new users

—

Lifetime conversion raterecent lifetime ÷ recent new users

—

Window conversion (pre-trial)ann / mon / life · May 18 – Jul 5 blended, comparison only

—

7 Diagnostics verify vs dashboard

Check	Model	Expected
D7 retention (model)	—	12.4% live Q2
Active days · year 1 (D0–D365)	—	~12.7
Active days · tail (D366→horizon)	—	power-law tail
Sponsor rev / WAU / wk	—	= (clk÷AU)×CPC
Upgrade LTV · pre-trial blended method	—	window rates, comparison
Total LTV · pre-trial method	—	$3.28 Jul 9 refresh (@$49)
Total LTV per download	—	trial-era method v2