A full sample of the FolioKit Monday brief

A complete sample brief — the exact format the generator produces, shown on a four-app portfolio.

Illustrative data. The numbers below are made up to show the format; the structure, rules, and tone are the real generator's. Real numbers from a real app arrive with the first public case study — we don't publish real-looking metrics before then.

Every Monday at 06:00 UTC, FolioKit hands the week's data — events, revenue events, screen views, new users, Apple proceeds, and any statistical anomalies — to an AI analyst and stores the result in your dashboard. The generator has strict rules: be specific and numeric, compute the week-over-week deltas itself, say when the data is too thin for a claim instead of inventing one, exclude internal and test traffic, never invent numbers for pre-launch apps, and stay under 400 words. Here is what one looks like.

Weekly brief — Monday 13 July 2026 · week of 6–12 July Sample

TL;DR

A good week, driven by one release: LuminaClean v2.4 lifted trials and proceeds, and the portfolio grew to $3,982 (+12% WoW). One thing needs your attention before you trust any TicFlow chart — paywall_shown appears to be double-firing since 1.4.2.

Highlights

LuminaClean — proceeds $2,410 (+18% WoW). v2.4 is the cause: trial starts from users who completed onboarding rose from 14.2% to 16.1%, while the skipped-onboarding rate stayed flat at 3.8%. DAU 1,240 (+6%).
Inkfall — proceeds $181 (+38% WoW) off a small base: 9 purchases vs 6 last week. Too thin to call a trend; worth one more week.
BookBinge — steady: $918 (+4% WoW), D7 retention holding at 22%, WAU 940 (+2%). Nothing needs you here.

Concerns

TicFlow instrumentation bug — paywall_shown fired 2,140 times against roughly 1,050 paywall screen views: almost exactly 2×, starting with the 1.4.2 rollout on 8 July. This looks like the event is attached to a view update, not to the paywall's presentation. Until it's fixed, TicFlow's funnel understates trial rate by about half — don't act on any TicFlow conversion number this week.
LuminaClean refunds from Japan — 3 refunds against a zero baseline (z = 3.1). All three were v2.4 purchases refunded within 24 hours. Not conclusive at this volume, but watch it.
TicFlow proceeds — $473 (−3% WoW), the only decline in the portfolio.

Revenue

Portfolio proceeds $3,982 (+12% WoW): $1,730 from new subscriptions and purchases, $2,252 from renewals. 214 trials started (+9%); trial-to-paid conversion 38% on the cohort that resolved this week. Top countries: US 58%, DE 11%, JP 8%. Refunds: 5 totalling $54 — all noted above.

Focus this week

Fix TicFlow's paywall_shown double-fire. It corrupts every downstream read on the app — nothing else about TicFlow is worth deciding until this ships.
Port LuminaClean's v2.4 paywall copy to TicFlow once the events are trustworthy — TicFlow's trial rate trails the portfolio even after correcting for the double-fire.
Watch LuminaClean's Japanese refunds. If two more land this week, review the JP price tier before assuming a product problem.

How this brief gets made

Anchored to your data, not vibes. The generator receives raw weekly aggregates plus detected anomalies (daily z-scores against each app's trailing 28 days) — the Japan-refunds line above is what a woven-in anomaly looks like.
Instrumentation checks are part of the job. Paired events with mismatched counts, sudden per-session volume changes, and release-aligned jumps get flagged as bugs, not celebrated as growth.
Thin data is called thin. At indie scale most weekly numbers are small; the brief says "too thin to call" instead of trend-fitting nine purchases.
Internal traffic excluded. Your own devices and simulators are flagged and removed before the analysis runs.