Focus Monke — launch ops, ASO research & ads plan

Launch Ops — Focus Monke v1.0 is live (launched 2026-08-07). Every tile carries its own “as of” stamp and a source-latency label — honesty over fake-real-time. The metrics file is republished by launch_poller.py on a ~5-minute deploy cadence; this page re-checks it every 60 s (next check in — s ).

App Store — live status & stars per storefront —

Revenue — RevenueCat —

RevenueCat tiles use each metric’s own reporting period; customer/activity totals are not the post-launch real-user cohort below.

Apple Ads — spend vs cap —

Acquisitions — every real user since launch —

first seen (Pacific)	market	city	user	keyword	campaign	keyword-day CPI (est.)	app sessions	focus done ✓	focus min	status

Ratings & written reviews —

New rating activity

Written reviews —

Review funnel — ask → yes-tap → rating — —

Last 30h (≈ tap→review landing lag)

Triangulation — yes-taps vs observed Apple rating changes

event	when	detail	status / pair

Review arrivals — when each written review surfaced

written (created)	first seen by poller	market	stars	likely trigger	rating showed up

Return trips — time outside the app after tapping review

PostHog — today’s usage —

Ops health —

Reddit Ads — spend vs budget —

Competition here is observed, not asserted. Every “comp” number is the live App Store top-10 for that exact term and storefront (free iTunes Search API), joined to real rating counts. med5 = median rating count of the top 5; weak = 100 − incumbent strength; opp = volume × weak / 100. AppTweak’s own difficulty is shown only as a secondary column and renders “—” when it came back 0/NULL, which it did on ~63% of rows — it is unmeasured, not “no competition”. Flags: W 10k+ incumbent · T relevance trap (top-5 mostly off-genre) · S stale shelf · O open shelf (med5<100, no whale) · P shelf purchased from AppTweak in wave 2, not iTunes ground truth (same scoring function, different observation — these are cells the free crawler could not reach).

Opportunity board

shelf-verified only hide traps

mkt	tier	term	vol	diff	med5	top ratings	weak	flags	cluster	OPP

Keyword × market heat matrix cell = volume · tint = incumbent weakness (green = open shelf) · blank = not measured

SERP explorer

#	app	developer	ratings	genre	updated

Competitor keyword inventories

keyword	their rank	vol

Credit ledger aggregated from api_log — every credit accounted

day	step	calls	credits

Week-1 Apple Ads opening book — plan v1.1 (2026-08-06), wave-2 validated. Source of truth: docs/APPLE-ADS-PLAN.md §3. Every volume / med5 below was re-queried from aso.sqlite when this page was built, and the build fails hard on any mismatch with the plan — cells verified this build, so these numbers cannot drift from the DB. med5 = median rating count of the observed top-5 (real shelf strength; small = winnable shelf). Flags: P shelf purchased from AppTweak, not iTunes ground truth — never read it as ground truth · O open shelf (med5<100, no whale) · W 10k+ incumbent in the top-10 · T top-5 partly off-genre · S stale shelf · ? shelf unmeasured. All keywords are exact match; Search Match runs only inside the Discovery campaigns with every exact negatived.

Daily budget — proposed v1.2, and yours to play with

Total daily cap

Proposed v1.2 = $235/day conservative · aggressive quick-set = 2× = $470 · expected actual spend at proposed caps is $95-165/day — thin-geo campaigns will not fill their caps, and that is the design, not a bug. v1.2 adds MONKE-X ($50) on top of v1.1's $185; no existing campaign was cut to fund it.

Scenarios save to this browser (localStorage) — use Export for safekeeping. The Proposed v1.2 preset itself is read-only in code and cannot be overwritten.

Unit economics — what a tap may cost annual $17.99, 7-day trial · net of Apple 15% + VAT · §1 of the plan

$1.08 max CPA

$0.67 max CPT

net $15.29/yr · bench CPT $1.58 — surgical exacts only

GB / AU / CA

$1.03 max CPA

$0.64 max CPT

net $14.50/yr · bench $1.31 (GB) — habit cluster only, wk-2

EU belt (fi dk no ie cz hu gr)

$0.99 max CPA

$0.61 max CPT

net $14.00/yr · bench $0.51-0.75 — core spend: clears the cap before optimization

$0.99 max CPA

$0.49 max CPT

net $14.00/yr · bench $0.73 — mid-budget; CVR (49% baseline) is the lever

MY / SG

$0.89 max CPA

$0.55 max CPT

net $12.50/yr · bench ~$0.20-0.40 — core spend

$0.46 max CPA

$0.30 max CPT

net $6.50/yr · bench $0.40-0.60 — small conversion test

$0.56 max CPA

$0.36 max CPT

net $7.90/yr · bench $0.40-0.60 — small conversion test

Derivation: RPI (expected year-1 revenue per install) = net × 6% install→trial × 35% trial→paid × 0.9 refund haircut × 1.25 revenue-mix bump × 1.5 paid-intent multiplier. Ben's 50%-recovery rule ⇒ max CPA = 2 × RPI; max CPT = max CPA × tap→install CVR (62% Tier A · 49% JP · 65% BR/MX). Annual prepay makes recovery measurable ~day 14 per cohort. At upside priors (10% trial, 40% t2p) every cap roughly doubles (belt ≈ $1.90) — week-2 cohort data decides which world we're in. Year-2 renewals (~28%) and the organic-rank dividend sit on top of all caps.

Phases, gates & the week-2 book

Week 1 — kill/scale rules, checked daily via the reporting API

keyword spend ≥ 3× its geo CPA cap with 0 downloads → pause
keyword CPA > 2× cap over ≥10 downloads → bid −30% (keep the history, don't delete)
impression share > 90% on a winner → bid −10% (stop overpaying)
campaign CPA ≤ cap over ≥20 new downloads → cap +50-100%
read new downloads, never installs (redownloads inflate)

Week 2 — cost per trial replaces install CPA as the gate

RC keyword-level cohorts (AdServices → RevenueCat) go live; target cost/trial ≤ $16 Tier A ($0.99 cap ÷ 6% trial rate) · ≤ $8 BR/MX; kill anything > 2× target
harvest Discovery → exacts weekly; expectation is asymmetric — US/JP should yield, the belt ~nothing (92% of 384 adjacent belt variants floored; an empty belt harvest is the predicted result)
stage the 3 CPPs (study / blocker / habit) — the best-documented CVR lever (+6-39%)
if install→trial measures 10% instead of 6%, every CPA cap rises ~1.7×

Week 3-4 — scale on measured recovery

cohort with ≥50% spend recovery at day 14 (n ≥ 10 trials) → 2× budget, repeat weekly while it holds
<25% recovery at day 21 → cut to maintenance or zero
only after gates clear: open the aggressive variant (2× caps, US bids to $1.50 on proven keywords)

Named week-2 expansion — TIER2-HABIT-X (ch be nz hk), ~$40/day

Wave 2 gave the tier-2 storefronts their first deliberate measurement: `habit tracker` clears 46-49 volume at difficulty 10-19 in all four, with soft shelves (73-160) — the same signature as the belt cells week 1 already buys. nz is the standout storefront: its `streaks` / `flora` / `focus` / `pomodoro timer` shelves are softer than anything in week 1, so all four fold into the same campaign. Opens in week 2 only if the week-1 belt habit cells clear their gates.

keyword	mkt	vol	med5 shelf	flags	note

Excluded on the shelf test despite clearing volume+difficulty: at (habit tracker med5 1,567) and tw (1,688) — no organic dividend, they wait for measured trial data. at is the mirror image: soft on pomo (`pomodoro` 26/101, `pomodoro timer` 32/233). Fallback week-2 option: ENA-HABIT-X (gb 56/406, au 55/1.0k, ca 54/910) — tier-2 outranks it on shelf softness. KR stays closed: both Korean spellings of pomodoro measure at the volume floor — Korea does not search "pomodoro"; no ko localization, no KR ads.