The account bought $793.94 across two campaigns and got 1,172 installs at $0.677, against $631.70 and 1,128 at $0.560 the day before. Booked revenue was $1,401.05; within-day was $620.48, down 30.9%. Our records read 0.98 of Meta's booked figure. This is the first day the five-cell split test ran, and the first day the report covers two campaigns. Definitions, the report clock and the three revenue measures are in the appendix.
Delivery rows are the delivering campaigns summed, which is exactly how section 5 defines the account, so the two agree. The retired earner's booking tail (7 installs on 08-05 and 1 on 08-06, both on $0.00 of spend) is excluded here and named in section 5, because counting it would put installs into a CPI denominator against spend that was never incurred.
| 2026-08-05 | 2026-08-06 | move | |
|---|---|---|---|
| spend | $631.70 | $793.94 | +25.7% |
| impressions | 53,818 | 70,411 | +30.8% |
| CPM | $11.74 | $11.28 | −3.9% |
| clicks | 4,939 | 5,403 | +9.4% |
| CTR | 9.18% | 7.67% | −16.4% |
| installs (Meta) | 1,121 | 1,171 | +4.5% |
| CPI | $0.564 | $0.678 | +20.3% |
| booked (Meta) | $1,427.51 (2.26) | $1,401.05 (1.76) | −1.9% |
| within-day | $898.32 (1.42) | $620.48 (0.78) | −30.9% |
| cohort @24h | $1,258.59 (1.99) | $702.17 (0.88) | at 58% coverage — completed figure in Caveats |
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.
| app | ARPU (d0) | ARPU (d3) | ARPU (d7) | ARPPU (d0) | ARPPU (d3) | ARPPU (d7) | installs |
|---|---|---|---|---|---|---|---|
| Vloom | $0.5073 | $0.8195 | $0.9360 | $11.23 | $15.85 | $17.13 | 1,373 |
Cost per install rose 20.3% on a quarter more spend, and within-day ROAS fell below 1 for the first time in the series.
CTR fell in 19 of the 20 hours both days delivered in, p = 0.000. A count of hours cannot be a mix artifact.
Payer rate fell with it. Matched hours at a 6-hour horizon, install → pay 5.35% → 3.28%, p = 0.016, on 62 and 36 payers.
The campaign mix barely moved, worldwide 84.9% → 85.2% of spend and German 15.1% → 14.8%, so almost none of the above is composition.
Delivery columns are Meta's instrument on both sides of every ratio; payer rate, ARPU and ARPPU are ours on both sides. The two install counts are never crossed. booked sits on the payment hour; within-day and cohort@24h sit on the install hour.
The full 24-row table is in _runs/2026-08-06_gen7_and_candidates/measures_out.txt; the shape of the day is in the charts below.
Hours both days delivered in, on the UTC clock: 00–08 and 13–23. 2026-08-06 additionally delivered in 09, 10, 11, 12. Only the shared hours are compared.
| metric | hours 2026-08-06 ran higher | hours it ran lower | sign test |
|---|---|---|---|
| CPM | 6 of 20 | 14 of 20 | p = 0.115 |
| CTR | 1 of 20 | 19 of 20 | p = 0.000 |
| click→install | 8 of 20 | 12 of 20 | p = 0.503 |
| CPI | 14 of 20 | 6 of 20 | p = 0.115 |
A sign test throws magnitude away and buys weight-free direction with it.
CTR is the only one of the four that separates, and it does so completely. CPM and CPI point the same way without clearing 0.05.
(i) The portfolio by country, 2026-08-06. On the report clock. Meta serves no hour x country grid, so each country cut is raked from its own Meta day onto UTC; our own columns were already there.
| cc | spend | share | impressions | CPM | registrations (ours) | cost/reg | within-day | wROAS | payers |
|---|---|---|---|---|---|---|---|---|---|
| IN | $160.56 | 20.2% | 25,669 | $6.26 | 344 | $0.467 | $93.59 | 0.58 | 36 |
| DE | $98.57 | 12.4% | 2,162 | $45.60 | 42 | $2.347 | $48.38 | 0.49 | 6 |
| US | $72.78 | 9.2% | 1,733 | $42.01 | 60 | $1.213 | $181.75 | 2.50 | 16 |
| TR | $26.92 | 3.4% | 1,457 | $18.47 | 44 | $0.612 | $5.89 | 0.22 | 2 |
| ID | $25.12 | 3.2% | 3,903 | $6.44 | 82 | $0.306 | $0.00 | 0.00 | 2 |
| MX | $22.94 | 2.9% | 3,110 | $7.38 | 53 | $0.433 | $5.71 | 0.25 | 3 |
| GB | $21.58 | 2.7% | 696 | $31.02 | 29 | $0.744 | $6.04 | 0.28 | 2 |
| AR | $20.79 | 2.6% | 1,759 | $11.81 | 53 | $0.392 | $9.98 | 0.48 | 2 |
| PH | $16.80 | 2.1% | 4,352 | $3.86 | 54 | $0.311 | $5.63 | 0.34 | 5 |
| AT | $15.30 | 1.9% | 524 | $29.20 | 8 | $1.912 | $21.19 | 1.39 | 3 |
| FR | $15.11 | 1.9% | 550 | $27.45 | 19 | $0.795 | $0.00 | 0.00 | 1 |
| CA | $14.28 | 1.8% | 410 | $34.81 | 16 | $0.893 | $4.99 | 0.35 | 1 |
183 further countries are not listed, $283.17 between them (35.7% of this block).
Reconciliation, what is measured and what is fitted. Every per-country cell is an allocation across hours; every total is measured. B1 (UTC-7) $208.10 from Meta 2026-08-05 and $585.84 from Meta 2026-08-06, a measured UTC-day total of $793.94.
Absent from the table above, by name: Nomad Node, KBM1, HR1. Their spend is in every day-grain table; only their country split is missing.
(ii) The same cut per account, 2026-08-06.
B1
| cc | spend | share | impressions | CPM | registrations (ours) | cost/reg | within-day | wROAS | payers |
|---|---|---|---|---|---|---|---|---|---|
| IN | $160.56 | 20.2% | 25,669 | $6.26 | 344 | $0.467 | $93.59 | 0.58 | 36 |
| DE | $98.57 | 12.4% | 2,162 | $45.60 | 42 | $2.347 | $48.38 | 0.49 | 6 |
| US | $72.78 | 9.2% | 1,733 | $42.01 | 60 | $1.213 | $181.75 | 2.50 | 16 |
| TR | $26.92 | 3.4% | 1,457 | $18.47 | 44 | $0.612 | $5.89 | 0.22 | 2 |
| ID | $25.12 | 3.2% | 3,903 | $6.44 | 82 | $0.306 | $0.00 | 0.00 | 2 |
| MX | $22.94 | 2.9% | 3,110 | $7.38 | 53 | $0.433 | $5.71 | 0.25 | 3 |
| GB | $21.58 | 2.7% | 696 | $31.02 | 29 | $0.744 | $6.04 | 0.28 | 2 |
| AR | $20.79 | 2.6% | 1,759 | $11.81 | 53 | $0.392 | $9.98 | 0.48 | 2 |
| PH | $16.80 | 2.1% | 4,352 | $3.86 | 54 | $0.311 | $5.63 | 0.34 | 5 |
| AT | $15.30 | 1.9% | 524 | $29.20 | 8 | $1.912 | $21.19 | 1.39 | 3 |
| FR | $15.11 | 1.9% | 550 | $27.45 | 19 | $0.795 | $0.00 | 0.00 | 1 |
| CA | $14.28 | 1.8% | 410 | $34.81 | 16 | $0.893 | $4.99 | 0.35 | 1 |
183 further countries are not listed, $283.17 between them (35.7% of this block).
Nomad Node — no country block: it delivered nothing on this day, a measured zero.
KBM1 — no country block: it delivered nothing on this day, a measured zero.
HR1 — no country block: it delivered nothing on this day, a measured zero.
The prior day, 2026-08-05, the same two tables.
(i) The portfolio by country, 2026-08-05. On the report clock. Meta serves no hour x country grid, so each country cut is raked from its own Meta day onto UTC; our own columns were already there.
| cc | spend | share | impressions | CPM | registrations (ours) | cost/reg | within-day | wROAS | payers |
|---|---|---|---|---|---|---|---|---|---|
| IN | $137.30 | 21.7% | 19,972 | $6.87 | 319 | $0.430 | $128.56 | 0.94 | 37 |
| DE | $71.30 | 11.3% | 1,441 | $49.47 | 35 | $2.037 | $21.85 | 0.31 | 2 |
| US | $54.50 | 8.6% | 1,258 | $43.33 | 50 | $1.090 | $288.78 | 5.30 | 15 |
| MX | $22.10 | 3.5% | 2,454 | $9.00 | 56 | $0.395 | $32.41 | 1.47 | 3 |
| ID | $21.22 | 3.4% | 3,014 | $7.04 | 73 | $0.291 | $43.68 | 2.06 | 2 |
| TR | $20.05 | 3.2% | 1,181 | $16.97 | 41 | $0.489 | $0.00 | 0.00 | 1 |
| GB | $19.08 | 3.0% | 534 | $35.72 | 20 | $0.954 | $0.00 | 0.00 | 1 |
| AR | $15.18 | 2.4% | 1,248 | $12.16 | 50 | $0.304 | $66.93 | 4.41 | 4 |
| MY | $13.07 | 2.1% | 1,546 | $8.46 | 27 | $0.484 | $5.38 | 0.41 | 3 |
| PH | $13.03 | 2.1% | 3,110 | $4.19 | 53 | $0.246 | $31.51 | 2.42 | 5 |
| FR | $10.05 | 1.6% | 299 | $33.62 | 11 | $0.913 | $0.00 | 0.00 | 0 |
| CA | $9.67 | 1.5% | 253 | $38.19 | 11 | $0.879 | $20.88 | 2.16 | 3 |
175 further countries are not listed, $225.17 between them (35.6% of this block).
Reconciliation, what is measured and what is fitted. Every per-country cell is an allocation across hours; every total is measured. B1 (UTC-7) $181.12 from Meta 2026-08-04 and $450.58 from Meta 2026-08-05, a measured UTC-day total of $631.70.
Absent from the table above, by name: Nomad Node, KBM1, HR1. Their spend is in every day-grain table; only their country split is missing.
(ii) The same cut per account, 2026-08-05.
B1
| cc | spend | share | impressions | CPM | registrations (ours) | cost/reg | within-day | wROAS | payers |
|---|---|---|---|---|---|---|---|---|---|
| IN | $137.30 | 21.7% | 19,972 | $6.87 | 319 | $0.430 | $128.56 | 0.94 | 37 |
| DE | $71.30 | 11.3% | 1,441 | $49.47 | 35 | $2.037 | $21.85 | 0.31 | 2 |
| US | $54.50 | 8.6% | 1,258 | $43.33 | 50 | $1.090 | $288.78 | 5.30 | 15 |
| MX | $22.10 | 3.5% | 2,454 | $9.00 | 56 | $0.395 | $32.41 | 1.47 | 3 |
| ID | $21.22 | 3.4% | 3,014 | $7.04 | 73 | $0.291 | $43.68 | 2.06 | 2 |
| TR | $20.05 | 3.2% | 1,181 | $16.97 | 41 | $0.489 | $0.00 | 0.00 | 1 |
| GB | $19.08 | 3.0% | 534 | $35.72 | 20 | $0.954 | $0.00 | 0.00 | 1 |
| AR | $15.18 | 2.4% | 1,248 | $12.16 | 50 | $0.304 | $66.93 | 4.41 | 4 |
| MY | $13.07 | 2.1% | 1,546 | $8.46 | 27 | $0.484 | $5.38 | 0.41 | 3 |
| PH | $13.03 | 2.1% | 3,110 | $4.19 | 53 | $0.246 | $31.51 | 2.42 | 5 |
| FR | $10.05 | 1.6% | 299 | $33.62 | 11 | $0.913 | $0.00 | 0.00 | 0 |
| CA | $9.67 | 1.5% | 253 | $38.19 | 11 | $0.879 | $20.88 | 2.16 | 3 |
175 further countries are not listed, $225.17 between them (35.6% of this block).
Nomad Node — no country block: it delivered nothing on this day, a measured zero.
KBM1 — no country block: it delivered nothing on this day, a measured zero.
HR1 — no country block: it delivered nothing on this day, a measured zero.
Pooled over the 20 hours both days delivered in, on the UTC clock. Meta's instrument on both sides.
| spend | impressions | CPM | clicks | CTR | installs | click→install | CPI | |
|---|---|---|---|---|---|---|---|---|
| 2026-08-05 | $631.69 | 53,814 | 11.74 | 4,939 | 9.18% | 1,117 | 22.62% | $0.566 |
| 2026-08-06 | $660.34 | 60,124 | 10.98 | 4,643 | 7.72% | 992 | 21.37% | $0.666 |
Pooling across hours is a mix comparison, weighted by delivery, so these rows can move with no hour changing. Pooled and hour-by-hour agree on CTR.
The three measures, both days:
| measure | 08-05 | 08-06 |
|---|---|---|
| booked | $1,427.51 (2.26) | $1,401.05 (1.76) |
| within-day | $898.32 (1.42) | $620.48 (0.78) |
| cohort @24h | $1,258.59 (1.99), 100% covered | $702.17 (0.88), 58% covered |
Cohort@24h is 58% covered on this cut, 08-05's 100%. The completed figure is in Caveats; day-over-day reads above run on within-day.
Booked revenue by the day its payer installed, whose top row is the within-day figure:
| installed | booked on 08-06 | share |
|---|---|---|
| 2026-08-06 | $620.48 | 44.3% |
| 2026-08-05 | $437.59 | 31.2% |
| 2026-08-04 | $84.12 | 6.0% |
| 2026-08-03 | $88.97 | 6.4% |
| 2026-08-02 | $81.64 | 5.8% |
| 2026-08-01 | $66.74 | 4.8% |
Booked held up (−1.9%) while within-day fell 30.9%, and the split says why: 55.7% of what the account booked on 08-06 came from people who installed earlier. On 08-05 that inherited share was 37.1%. A day whose own cohort earns less looks flat on Meta's number and poor on ours; both are true of different populations.
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 a direction and never a magnitude.
| campaign | spend | share | impressions | impr share | CPM | clicks | CTR | cost/click | installs | CPI | within-day | wROAS | payers |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Worldwide | $676.53 | 85.2% | 67,749 | 96.2% | 9.99 | 5,218 | 7.70% | $0.130 | 1,117 | $0.606 | $556.65 | 0.82 | 49 |
| German (DE/AT) | $117.41 | 14.8% | 2,662 | 3.8% | 44.11 | 185 | 6.95% | $0.635 | 54 | $2.174 | $63.83 | 0.54 | 7 |
Booking with no delivery, and excluded from every rate above: Earner (retired 08-03) 1 install (Meta), 13 registrations (ours), $0.00 within-day. Counting them would put installs into the CPI denominator against spend that never happened.
Account CPM $11.74 → $11.28 (−3.9%); held to 2026-08-05's campaign mix, $11.15 (−5.0%). The mix is worth about a point, because the split barely moved.
A German impression costs 4.4× a worldwide one and a German click 4.9×, on 2,662 impressions in its last full day before the switch-off.
Worldwide — every section below this line is Worldwide's: the day, the hours, the countries, the campaigns and the assets, counted on Worldwide's own ledger and its own accounts. Use the buttons above to switch.
Six ad sets, five holding one delivering ad. ..._d100_spill holds a replacement pair of which one delivers, so the ad-set row is the ad row on this day and rows are labelled by asset. E2/mugshot was switched off on 08-04 and appears with $0.00 spend against a conversion-hour tail.
Meta's instrument on both sides of every ratio. Our registrations never enter this table.
| asset | spend | impressions | CPM | clicks | CTR | installs | click→install | CPI |
|---|---|---|---|---|---|---|---|---|
E2/bikini | $144.49 | 16,983 | 8.51 | 1,187 | 6.99% | 240 | 20.22% | $0.602 |
E2/lounge | $142.73 | 10,144 | 14.07 | 869 | 8.57% | 175 | 20.14% | $0.816 |
E2/pool5 | $134.63 | 16,357 | 8.23 | 1,206 | 7.37% | 250 | 20.73% | $0.539 |
E2/spill2 | $131.20 | 9,955 | 13.18 | 1,003 | 10.08% | 216 | 21.54% | $0.607 |
E2/bodysuit | $123.48 | 14,310 | 8.63 | 953 | 6.66% | 235 | 24.66% | $0.525 |
E2/mugshot | $0.00 | 0 | — | 0 | — | 1 | — | $0.000 |
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, 58% covered on this day. Report clock UTC.
| asset | booked | bROAS (Meta) | within-day | wROAS | cohort@H | cROAS | payers | payer% | ARPU | ARPPU | largest payer |
|---|---|---|---|---|---|---|---|---|---|---|---|
E2/bikini | $194.14 | 1.34 (1.04) | $138.01 | 0.96 | $145.13 | 1.00 | 10 | 3.8% | 0.521 | 13.80 | 44.2% |
E2/lounge | $92.38 | 0.65 (1.28) | $92.38 | 0.65 | $113.57 | 0.80 | 10 | 5.3% | 0.491 | 9.24 | 25.5% |
E2/pool5 | $350.56 | 2.60 (2.97) | $121.48 | 0.90 | $121.48 | 0.90 | 13 | 5.0% | 0.467 | 9.34 | 35.4% |
E2/spill2 | $122.74 | 0.94 (1.72) | $98.24 | 0.75 | $146.64 | 1.12 | 8 | 3.5% | 0.427 | 12.28 | 43.3% |
E2/bodysuit | $341.26 | 2.76 (2.44) | $106.54 | 0.86 | $111.53 | 0.90 | 8 | 3.2% | 0.423 | 13.32 | 44.3% |
E2/mugshot | $35.19 | — (—) | $0.00 | — | $0.00 | — | 0 | 0.0% | 0.000 | — | — |
The 16-payer bar (#37): any row above whose payer count is under 16 is a direction and never a magnitude.
⚠ 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.
| asset | spend | India % | US % | other % | US % of its revenue |
|---|---|---|---|---|---|
E2/bikini | $147.80 | 29.3% | 9.2% | 61.6% | 70.9% |
E2/lounge | $156.80 | 18.6% | 12.1% | 69.3% | 5.4% |
E2/pool5 | $124.21 | 23.0% | 13.6% | 63.5% | 58.4% |
E2/spill2 | $133.98 | 16.7% | 11.7% | 71.7% | 8.1% |
E2/bodysuit | $116.09 | 27.5% | 9.1% | 63.3% | 0.0% |
The five-cell test ran its first day. Meta runs A/B windows on the ad account's calendar, "Aug 6, 2026, 12:00 AM – Aug 11, 2026, 12:00 AM Pacific Time", so this block is read on PT over the hours the test actually ran, from the first settled hour after the arms went level.
Three arms went $100 → $150 at 04:05–04:06 PT on 08-06, read from the activity history. The window below is therefore Meta 08-06 hours 04–23 PT, all five arms level at $150/day.
| arm | spend | CPM | CTR | installs | CPI | click→install |
|---|---|---|---|---|---|---|
| bodysuit | $100.23 | 8.39 | 6.75% | 186 | $0.539 | 23.08% |
| pool5 | $107.06 | 8.19 | 6.87% | 182 | $0.588 | 20.27% |
| bikini | $134.20 | 8.74 | 6.63% | 204 | $0.658 | 20.04% |
| spill | $117.26 | 12.69 | 8.42% | 166 | $0.706 | 21.34% |
| lounge | $143.19 | 11.31 | 7.95% | 185 | $0.774 | 18.37% |
Meta's own results page ranks them identically, which is two instruments agreeing.
⚠ The CPI ranking is the spend ranking, Spearman +0.90. With #48's CTR spend elasticity of −0.724, day 1 cannot separate creative from spend rate.
The split does not equalise spend. Realised spend runs $100.23–$143.19, a 1.43× range, and impressions 1.66×, so E2's "delivery forced equal" is too strong.
Nothing separates after correction. The best pair is bodysuit against lounge on click→install, p = 0.0136, one of ten post-hoc tests against a Bonferroni 0.005.
German (DE/AT) — every section below this line is German (DE/AT)'s: the day, the hours, the countries, the campaigns and the assets, counted on German (DE/AT)'s own ledger and its own accounts. Use the buttons above to switch.
Two ad sets, and this is the campaign that needs the ad level: deat_broad_purchase_adv_d75 holds 15 ads with 2 delivering, and deat_ww4_broad_purchase_adv holds 4 delivering concurrently.
Meta's instrument on both sides of every ratio. Our registrations never enter this table.
| asset | spend | impressions | CPM | clicks | CTR | installs | click→install | CPI |
|---|---|---|---|---|---|---|---|---|
DEww/bodysuit | $62.69 | 1,461 | 42.91 | 86 | 5.89% | 30 | 34.88% | $2.090 |
DEww/spill2 | $20.82 | 281 | 74.09 | 34 | 12.10% | 6 | 17.65% | $3.470 |
DE/german_mature | $19.67 | 467 | 42.12 | 38 | 8.14% | 12 | 31.58% | $1.639 |
DEww/bikini | $6.73 | 169 | 39.82 | 12 | 7.10% | 3 | 25.00% | $2.243 |
DEww/pool5 | $5.00 | 251 | 19.92 | 11 | 4.38% | 1 | 9.09% | $5.000 |
DE/german | $2.50 | 33 | 75.76 | 4 | 12.12% | 1 | 25.00% | $2.500 |
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, 58% covered on this day. Report clock UTC.
| asset | booked | bROAS (Meta) | within-day | wROAS | cohort@H | cROAS | payers | payer% | ARPU | ARPPU | largest payer |
|---|---|---|---|---|---|---|---|---|---|---|---|
DEww/bodysuit | $24.75 | 0.39 (0.40) | $24.75 | 0.39 | $24.75 | 0.39 | 3 | 11.5% | 0.952 | 8.25 | 37.6% |
DEww/spill2 | $0.00 | 0.00 (0.00) | $0.00 | 0.00 | $0.00 | 0.00 | 0 | 0.0% | 0.000 | — | — |
DE/german_mature | $33.34 | 1.69 (1.70) | $33.34 | 1.69 | $33.34 | 1.69 | 3 | 21.4% | 2.381 | 11.11 | 65.5% |
DEww/bikini | $5.74 | 0.85 (0.86) | $5.74 | 0.85 | $5.74 | 0.85 | 1 | 33.3% | 1.915 | 5.74 | 100.0% |
DEww/pool5 | $0.00 | 0.00 (0.00) | $0.00 | 0.00 | $0.00 | 0.00 | 0 | 0.0% | 0.000 | — | — |
DE/german | $0.00 | 0.00 (0.00) | $0.00 | 0.00 | $0.00 | 0.00 | 0 | 0.0% | 0.000 | — | — |
The 16-payer bar (#37): any row above whose payer count is under 16 is a direction and never a magnitude.
⚠ 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.
| asset | spend | India % | US % | other % | US % of its revenue |
|---|---|---|---|---|---|
DEww/bodysuit | $59.36 | 0.0% | 0.0% | 100.0% | 0.0% |
DEww/spill2 | $18.13 | 0.0% | 0.0% | 100.0% | — |
DE/german_mature | $18.56 | 0.0% | 0.0% | 100.0% | 0.0% |
DEww/bikini | $6.47 | 0.0% | 0.0% | 100.0% | 0.0% |
DEww/pool5 | $5.78 | 0.0% | 0.0% | 100.0% | — |
DE/german | $1.64 | 0.0% | 0.0% | 100.0% | — |
Seven payers across the whole campaign. No German cell has ever cleared the 16-payer bar, so every figure in table B is a direction.
Within-set concentration recurred. Inside deat_ww4_broad_purchase_adv, bodysuit took 67.7% of the set's spend and pool5 3.8%, so its four worldwide creatives cannot be ranked.
The campaign was switched off at 22:49 PT on 08-06, ending Gen 7 at about 33 hours. The full read is in german_worldwide_swap_2026-08-06.md.
The cohort figure below was published at 58% coverage and has since completed. At the full 24-hour horizon 2026-08-06's cohort is $715.22, the figure the 2026-08-07 report carries in its prior-day column; what this page shows is a lower-coverage read of the same cohort, not a different number. Against 08-05's completed $1,258.59 the day reads −43.2% — the comparison this page declined. Within-day and booked are unaffected; they were sealed and settled when this page shipped.
The prior day's figures differ slightly from the published 08-05 report. That report was built from an hourly export pulled at 19:23 PT on 08-05, two hours after the UTC day closed; this one uses a settled re-pull. Spend moves $625.17 → $631.70, +1.0%. The direction is the documented one, Meta going on settling for a few hours, and it is why exports are pulled after the day closes.
Cohort@24h is 58% covered on this cut. The day-over-day reads on this page run on within-day; Caveats carries the completed figure.
The country table is a META-day cut and is seven hours offset from every other table here.
The split's one-arm assignment is what removed the disjoint-delivery failure that voided Gen 5.
The day the CTR fell is the day the split test began, and those may not be independent. Before 08-06 the five ad sets shared an audience, and Meta's auction-overlap rule entered the highest-value one of them into each auction. From 08-06 each person is assigned to exactly one arm, so the account lost a max-of-five selection it had been getting for free. That mechanism predicts a CTR fall with no creative getting worse and no audience getting worse. It is a hypothesis, but it means pre-test days are not a clean baseline for test days, and the 15.9% pooled CTR fall should not be read as decay.
The test has four days left and 08-07 is the first day comparable to another in-test day. The German campaign is off, so 08-07 is a single-campaign day again and section 5 will carry one row plus the German booking tail.
python -m ad_ops.measures --config _runs/2026-08-06_gen7_and_candidates/run_config.json --day
2026-08-06 --prev 2026-08-05
python -m ad_ops.day_compare --config … --today 2026-08-06 --prev 2026-08-05 --h 6
python -m ad_ops.cohort_arpu --config … --days 2026-08-02,…,2026-08-06
python -m ad_ops.asset_estimator hourly --meta <08-04,08-05,08-06 hourly> --lum <run>/lum --fx … \
--ads ww_bikini_10s,ww_bodysuit_10s,ww_pool5,ww_spill,ww_lounge \
--campaign bailingxia_meituan_ww_cvr_260802 --horizon 24
python -m ad_ops.report_tables --config … --day 2026-08-06 --prev 2026-08-05 [--campaign-table | --campaign ww|de]
python -m ad_ops.report_charts --config … --day 2026-08-06 --prev 2026-08-05 --days 2026-08-03,…,2026-08-06Exports pulled 2026-08-07 04:07–07:49 PT: ads × hour for Meta days 08-05 and 08-06 (24 buckets each), ads × day and ads × country × day for Jul 1 → Aug 7. Lum Telegram export taken on the laptop 2026-08-07 07:55–08:00 PT: 15,234 registrations, 1,598 payments, 8 unmatched, payment data cut 2026-08-07 14:53 UTC. Both counts rose against the previous cut and unmatched sits at 0.5%, which are the three tells that the export is not truncated.
The report clock is UTC. Adjust reports in UTC, so a UTC-bucketed report lines up with the attribution dashboard. Meta's ad account is fixed at UTC-7, so a UTC day is Meta day D-1 17:00–23:59 plus Meta day D 00:00–16:59, which is why every daily needs two hourly exports. Table C's per-asset country mix cannot follow it: the rake is per ACCOUNT, against that account's own hourly margin, and an asset has no hourly margin of its own.
The three revenue measures, never one, never mixed, always labelled:
| measure | counts | moves after the day closes? |
|---|---|---|
| booked | revenue that arrived inside the day, whatever day its payer installed. Meta's own number is this basis | settles ~4 h after midnight, then fixed |
| within-day | installed that day and paid before that day closed | never. Sealed at midnight |
| cohort at H | that day's installs, counted within H hours of each person's own install. House horizon H=24 | rises until every install has lived H hours. Always state coverage |
All three are read on our own payment records. Meta can express only booked, so it appears bracketed beside that column and never as a row of its own.
A regime is a stretch over which the ad set's daily budget did not change. It matters because the budget is a geography dial on this account: raising it buys wider, cheaper countries, so two creatives measured across a budget change are compared on different audiences.