Google Ads waste matrix

• Dead money — spend that produced zero revenue and zero conversions, on enough clicks that zero is a signal rather than noise. This is waste in any business at any margin. It needs no assumption and it is the only number here we call waste.
• Underwater — spend that earned back LESS than it cost (ROAS below 1.0). This also needs no assumption: revenue below cost loses money at any margin, because even a 100% margin merely breaks even.
• What we cannot know: their margins. Google reports revenue, never profit. So underwater is a floor, not the truth. A business on a 40% margin needs 2.5x ROAS just to stand still, which means plenty of spend between 1.0x and 2.5x is also losing money. We can't size that without their numbers, so we don't pretend to. Ask the client for a margin and the real figure will be considerably larger.
• Underperforming phrases are never listed as negatable. A core category word can sit below break-even and still earn well, and it almost always appears inside clear winners too — negating it would be self-harm. The n-gram row therefore only claims phrases whose entire cohort earned nothing.
• Display, Demand Gen and Video are never judged on last-click revenue — upper-funnel is supposed to look bad on it. They are only checked for junk inventory.
• PMax placements can never carry a dollar figure. The API exposes impressions only, so they get a rule-scored suspect list with a 1,000-impression floor.

Client 1 · PMax placements: visible and excludable, but never priceable
43,095 placements served 111,049 impressions (6 mobile apps). Google gives names, types and target URLs but attaches NO spend, so PMax placement waste cannot be quantified through the API — by Google's design, not a query bug. So we rank suspects by rule instead. See the suspect list.

Client 2 · Upper-funnel campaigns held out of the waste total (Mike's rule)
$12153 in Display/Demand Gen/Video. These must NOT be judged on last-click conversions. 0 of 1 show zero conversions, zero all_conversions AND zero view-through — those are worth a look, but they are reported here, not counted as recoverable waste.

Client 2 · Demand Gen / Display placements NOT judged on last-click (for reference only)
If we had condemned upper-funnel placements on last-click revenue, this row would read $9542 instead of $0. That number is meaningless — Discover and Gmail placements are SUPPOSED to show low last-click revenue. Recorded here so nobody re-introduces the mistake.

Client 2 · PMax placements: visible and excludable, but never priceable
118,233 placements served 355,325 impressions (591 mobile apps). Google gives names, types and target URLs but attaches NO spend, so PMax placement waste cannot be quantified through the API — by Google's design, not a query bug. So we rank suspects by rule instead. See the suspect list.

Client 3 · Upper-funnel campaigns held out of the waste total (Mike's rule)
$36194 in Display/Demand Gen/Video. These must NOT be judged on last-click conversions. 0 of 1 show zero conversions, zero all_conversions AND zero view-through — those are worth a look, but they are reported here, not counted as recoverable waste.

Client 3 · Demand Gen / Display placements NOT judged on last-click (for reference only)
If we had condemned upper-funnel placements on last-click revenue, this row would read $29104 instead of $0. That number is meaningless — Discover and Gmail placements are SUPPOSED to show low last-click revenue. Recorded here so nobody re-introduces the mistake.

Client 3 · PMax placements: visible and excludable, but never priceable
204,454 placements served 1,820,270 impressions (1300 mobile apps). Google gives names, types and target URLs but attaches NO spend, so PMax placement waste cannot be quantified through the API — by Google's design, not a query bug. So we rank suspects by rule instead. See the suspect list.

Client 4 · PMax placements: visible and excludable, but never priceable
27,913 placements served 198,027 impressions (9 mobile apps). Google gives names, types and target URLs but attaches NO spend, so PMax placement waste cannot be quantified through the API — by Google's design, not a query bug. So we rank suspects by rule instead. See the suspect list.

Client 5 · Landing-page crawl was blocked by the site
37 of 38 top URLs returned 403/429/timeout to our crawler. These pages are almost certainly fine for real users. This detector is N/A for this account — it needs a real browser (Playwright) to check properly.

Client 6 · PMax placements: visible and excludable, but never priceable
54,005 placements served 114,160 impressions (14 mobile apps). Google gives names, types and target URLs but attaches NO spend, so PMax placement waste cannot be quantified through the API — by Google's design, not a query bug. So we rank suspects by rule instead. See the suspect list.

Client 7 · Upper-funnel campaigns held out of the waste total (Mike's rule)
$813 in Display/Demand Gen/Video. These must NOT be judged on last-click conversions. 1 of 3 show zero conversions, zero all_conversions AND zero view-through — those are worth a look, but they are reported here, not counted as recoverable waste.

Client 7 · Demand Gen / Display placements NOT judged on last-click (for reference only)
If we had condemned upper-funnel placements on last-click revenue, this row would read $678 instead of $0. That number is meaningless — Discover and Gmail placements are SUPPOSED to show low last-click revenue. Recorded here so nobody re-introduces the mistake.

Client 7 · PMax placements: visible and excludable, but never priceable
50,100 placements served 333,164 impressions (0 mobile apps). Google gives names, types and target URLs but attaches NO spend, so PMax placement waste cannot be quantified through the API — by Google's design, not a query bug. So we rank suspects by rule instead. See the suspect list.

Client 8 · Upper-funnel campaigns held out of the waste total (Mike's rule)
$6195 in Display/Demand Gen/Video. These must NOT be judged on last-click conversions. 0 of 1 show zero conversions, zero all_conversions AND zero view-through — those are worth a look, but they are reported here, not counted as recoverable waste.

Client 8 · Demand Gen / Display placements NOT judged on last-click (for reference only)
If we had condemned upper-funnel placements on last-click revenue, this row would read $4661 instead of $0. That number is meaningless — Discover and Gmail placements are SUPPOSED to show low last-click revenue. Recorded here so nobody re-introduces the mistake.

Client 8 · PMax placements: visible and excludable, but never priceable
11,382 placements served 534,229 impressions (1351 mobile apps). Google gives names, types and target URLs but attaches NO spend, so PMax placement waste cannot be quantified through the API — by Google's design, not a query bug. So we rank suspects by rule instead. See the suspect list.

Client 9 · Upper-funnel campaigns held out of the waste total (Mike's rule)
$7997 in Display/Demand Gen/Video. These must NOT be judged on last-click conversions. 0 of 1 show zero conversions, zero all_conversions AND zero view-through — those are worth a look, but they are reported here, not counted as recoverable waste.

Client 9 · Demand Gen / Display placements NOT judged on last-click (for reference only)
If we had condemned upper-funnel placements on last-click revenue, this row would read $4148 instead of $39. That number is meaningless — Discover and Gmail placements are SUPPOSED to show low last-click revenue. Recorded here so nobody re-introduces the mistake.

Client 9 · PMax placements: visible and excludable, but never priceable
16,015 placements served 354,947 impressions (0 mobile apps). Google gives names, types and target URLs but attaches NO spend, so PMax placement waste cannot be quantified through the API — by Google's design, not a query bug. So we rank suspects by rule instead. See the suspect list.

Client 10 · Upper-funnel campaigns held out of the waste total (Mike's rule)
$4945 in Display/Demand Gen/Video. These must NOT be judged on last-click conversions. 0 of 1 show zero conversions, zero all_conversions AND zero view-through — those are worth a look, but they are reported here, not counted as recoverable waste.

Client 10 · Demand Gen / Display placements NOT judged on last-click (for reference only)
If we had condemned upper-funnel placements on last-click revenue, this row would read $4497 instead of $1846. That number is meaningless — Discover and Gmail placements are SUPPOSED to show low last-click revenue. Recorded here so nobody re-introduces the mistake.

Client 10 · PMax placements: visible and excludable, but never priceable
35,515 placements served 381,977 impressions (768 mobile apps). Google gives names, types and target URLs but attaches NO spend, so PMax placement waste cannot be quantified through the API — by Google's design, not a query bug. So we rank suspects by rule instead. See the suspect list.

Client 11 · Upper-funnel campaigns held out of the waste total (Mike's rule)
$5177 in Display/Demand Gen/Video. These must NOT be judged on last-click conversions. 0 of 1 show zero conversions, zero all_conversions AND zero view-through — those are worth a look, but they are reported here, not counted as recoverable waste.

Client 11 · Demand Gen / Display placements NOT judged on last-click (for reference only)
If we had condemned upper-funnel placements on last-click revenue, this row would read $3763 instead of $0. That number is meaningless — Discover and Gmail placements are SUPPOSED to show low last-click revenue. Recorded here so nobody re-introduces the mistake.

Client 11 · PMax placements: visible and excludable, but never priceable
29,860 placements served 704,795 impressions (0 mobile apps). Google gives names, types and target URLs but attaches NO spend, so PMax placement waste cannot be quantified through the API — by Google's design, not a query bug. So we rank suspects by rule instead. See the suspect list.