Monthly campaign report — August 2026 to date, against July 2026

1Overview

Rates first, because they are the part that compares. The level rows are printed because the report has to reconcile, and their change column is empty by construction.

Cohort value — ARPU and ARPPU at d0, d3 and d7

Keyed on the INSTALL day and measured from each user's own install stamp: dN is their first (N+1) x 24 h. ARPU is Adjust's own series, the one the partner's dashboard reports; ARPPU is our payment ledger, because Adjust exposes no cumulative distinct-payer count. A horizon the window has not lived through yet is left empty rather than written as zero.

Portfolio spend $9,239.32. B1 95.2%, the shared Nomad Node accounts 4.8% on five days of the fifteen; sections 3 onward are B1 alone.

Campaign split inside B1: Worldwide 83.8% over thirteen days, the retired earner 13.5% over three, the German buy 2.7% — never simultaneously live.

Impressions and installs both got dearer, and the unweighted hour count agrees: CPM higher in 20 of 24 hours, CPI in 22 of 24.

More people clicked and fewer of them installed: CTR +3.8%, click→install −13.8%, and the unweighted count agrees at 18 of 24 hours lower.

The sealed return barely moved; the booked one nearly doubled on age alone. July's nine days were the account's first, with nothing behind them.

The shared buy leads B1 on both comparable measures and trails on booked, which cannot compare — section 2.

2The two ad accounts, side by side

Both accounts are ours: B1, act 1738541847343652, campaigns bailingxia_; and the two shared Nomad Node accounts, acts 835434076171203 and 3267486106793669, whose campaigns for this app are 0811_ and 0814_*. Those two accounts also carry other products and none of that is counted here. Spend comes from each account's own export; DNU, the daily new users, and the three revenue measures are our own records taken across everything each account's campaigns carried, which is why B1's row here sits a little above the ad-id-scoped totals in the sections below.

The shared buy is five days inside fifteen, so its 4.8% is a start date. Its returns compare; its dollar total does not.

The shared buy is ramping: $31.77 of a $903.50 portfolio day on 08-11, $166.42 of $663.72 on 08-15. The window figures above under-read it.

Rank the two on within-day and cohort@24h, never on booked. Booked puts B1 ahead 2.28 to 2.09; the measures that cannot inherit reverse it.

The shared side buys new users at 43% of B1's price and each is worth less on landing. Cost is the larger effect.

Both accounts' revenue lands in one record, so only the portfolio row is account-blind (DECISION_LOG.md #64).

The shared accounts' export carries no Ad ID, campaign name, link clicks, leads or country, so sections 3 to 7 are B1 alone.

3Hourly, B1, 2026-08-01 to 2026-08-15

B1 alone, on the report clock. The axis is hour of day pooled across the fifteen days: each bucket holds fifteen clock hours, and the chart has 24 of them. Read it for diurnal shape and never as a sequence. Delivery columns are Meta's instrument on both sides of every ratio; payer rate, ARPU and ARPPU are ours on both sides, and the two install counts are never crossed. ours is our registrations in that bucket and cov is the bucket's cohort coverage at H=24.

⚠ Hour 14 is almost entirely one hour of one day, the rejected spill build on 2026-08-03 (Caveats).

Pooling fifteen days puts 21 to 39 within-day payers in every bucket, above the 16-payer bar (DECISION_LOG.md #37); no daily's hourly table can.

The cheap hours and the paying hours are the same block — 00 to 09 — and the account spends more outside it.

That coincidence is descriptive and not yet an instruction: these buckets pool four different buys, the earner's three days through the single-arm regime from 08-11.

4Country, 2026-08-01 to 2026-08-22

5Aggregate, both periods, B1

B1 alone, pooled over the 24 hours both periods delivered in, on the UTC clock. Meta's instrument on both sides of every ratio.

• click → install 22.66% → 19.54% (-13.8%, p = 0.000) — fell.
• CTR 5.80% → 6.02% (+3.8%, p = 0.000) — rose.

Pooling across hours weights each hour by what it delivered. The hourly pairing is the unweighted read of the same two periods; where they disagree, the mix moved.

Both reads agree: CPM up, CTR up, click→install down, CPI up, so the movement is not a mix effect at the hour grain.

The mix-neutral account line, $14.44 → $8.24 held to July's mix at $11.03, is about two campaigns and not the account (Caveats).

The three measures side by side

Only the bracketed returns compare. The dollar figures are fifteen days against nine.

The three measures disagree; read within-day. Booked's 1.27 to 2.28 is barely buying, and the cohort row is a floor 3% short of coverage.

0.95 to 0.97 is not a movement at this account's daily variance: August's own days run 0.57 to 1.44 and July's 0.39 to 1.71.

Booked revenue by how long its payer had been installed

Over one day this is a split by install date. Over a period it is a split by lag, because the same calendar day is a different age for every payer in it. The top row is the within-day total seen from the other side, which is the identity that makes the split worth drawing.

Booked revenue arriving from its own day fell 75.0% to 42.3%, mostly on the account's age. Within-day survives it, counting only the same-day row.

August's rows stop at seven days and do not sum to its booked total: the remainder installed over a week out.

6B1's campaigns, side by side

B1's campaigns alone.

Delivery is Meta's instrument on both sides of every ratio; every revenue measure is ours. Cohort horizon H=24h. Report clock UTC. Rank on CPM, cost per click and CPI only — geography is the treatment here, which is what makes those comparable across campaigns, while revenue still needs 16 payers (DECISION_LOG.md #37, #48) and a campaign under that bar carries only a direction.

Booking with no delivery, and excluded from every rate above: T1 (Jul 23) 0 installs (Meta), 3 registrations (ours), $0.00 within-day; T2 (Jul 23) 0 installs (Meta), 13 registrations (ours), $0.00 within-day; T3 (Jul 23) 0 installs (Meta), 42 registrations (ours), $0.00 within-day; T1 (Jul 25) 0 installs (Meta), 44 registrations (ours), $4.99 within-day; T2 (Jul 25) 1 installs (Meta), 68 registrations (ours), $0.00 within-day; T3 (Jul 25, India/SEA) 5 installs (Meta), 446 registrations (ours), $53.71 within-day. Counting them would put installs into the CPI denominator against spend that never happened.

Account CPM $14.44 → $8.24 (-42.9%). Held to 2026-07-23..2026-07-31's campaign mix it is $11.03 (-23.6%) — the difference between those two is the mix.

The earner's 13.5% is three days of fifteen, never sharing an hour with the worldwide campaign — 11:20 UTC on 08-03 against 12:44 (asset_performance_report_2026-08-03.md).

The earner is the cheapest line in the table, and both campaigns clear the 16-payer bar, so 1.29 and 0.92 read as levels (DECISION_LOG.md #48).

The German buy costs 5.8 times the account's CPM and 3.9 times its CPI, on 0.5% of impressions. Its 9 payers leave 0.47 a direction.

Retired campaigns are still booking, and the largest, T3 (Jul 25, India/SEA), stopped delivering three weeks ago.

The standing charts for this section plot spend share, CPM and CPI by campaign across the run. They are emitted per campaign into the tabs below.

7Inside B1's campaigns

Structure over the period

Twenty assets took spend across the fifteen days, in three campaigns and across several budget regimes: the published dailies name E2-R6 through E2-R10 inside this period alone, and the earner ran its own before them. The tables in this section pool every one of those regimes, and the budget is a geography dial on this account, so no row below is regime-matched and none of them can settle a comparison between two creatives (DECISION_LOG.md #47, #48). What they can do is describe what the account bought and what came back.

The account's shape changed four times inside the period. It opened on the earner's single CBO ad set holding three creatives; on 08-03 a new campaign of five ad sets, one creative each at $100 a day, replaced it; the five-cell split test ran from 08-06; and from 08-11 20:22 UTC the account ran one arm at $450, bodysuit first and pool5 from 08-12 14:18 account time. Meta rejected spill on 08-03 and spill2 on 08-09 at 13:12 UTC, both under the Adult Sexual Solicitation standard, and both rejections are visible in the delivery table as arms whose spend stops.

A. Per asset — delivery

Meta's instrument on both sides of every ratio. Our registrations never enter this table.

E2/spill is here as a record and not as a result: almost all of it is one hour on 08-03, before the build was rejected.

Three arms cost within 1.4 cents of each other per install on CPMs from 8.08 to 11.02; the cheapest-impression arm took the most spend.

E2/lounge is the one arm dearer per install than the pack, $1.043 against $0.758 to $0.772, on the lowest click→install at 16.65%.

B. Per asset — the three revenue measures

Ours on every column. Meta can express only booked, so it rides bracketed beside that column and never as a row. Payers are distinct people, never transactions. Cohort horizon H=24h, 97% covered on this period. Report clock UTC.

The 16-payer bar (#37): any row above whose payer count is under 16 is a direction and never a magnitude.

Seven rows clear the bar, which fifteen days of pooling bought and no daily in this series can.

Clearing the bar makes each row readable and does not rank them: the seven ran in different weeks and budgets, on audiences the auction chose.

E2/pool5 is the largest row here and among the least concentrated: 223 payers, top payer 6.3%, the highest payer rate and ARPPU of the seven.

E2/mugshot is the arm the account was right to cut — the only E2 arm whose delivery and revenue both sit at the bottom.

C. Per asset — country mix

⚠ This table is cut on the META day, while every other table on this page is on the UTC clock. Meta serves no hour × country grid, so per-country delivery exists only on the ad account's own UTC-7 day and cannot be restitched. It is a seven-hour-offset window against tables A and B above; read it for mix, never for level against them.

The last column is the share of that asset's own within-day revenue that came from the United States: where it far exceeds the US spend share, the return is a property of the geography the asset bought.

The three E2 arms with volume earn about twice as much in the United States as they spend there; the earner's arms do not.

The E2 arms and the earner's cannot be compared on revenue: they bought different countries, and country is most of the return here (DECISION_LOG.md #47).