The Spa Dr — Search + Shopping WoW Report Revenue +39% · Recovery

Current Week (A): Jul 20–26  ·  Prior Week (B): Jul 13–19  ·  Generated: Jul 28, 2026  ·  Data: CDP (Redash) + Google Ads API + Bing Ads API + Shopify
⚠ Structural week: NB PMax paused Jul 22 · HairMask Shopping paused · Demand Gen paused Jul 27 · Bing pulled via CDP on the same Mon–Sun weeks as Google
Report Navigation
Section 1 of 13

Executive Summary — Jul 20–26 vs Jul 13–19

📊 Headline: A clean recovery week. Blended revenue rose +39% WoW ($3,505 vs $2,522) and orders +40% (56 vs 40) on spend +16%. Blended ROAS improved 3.00x → 3.58x, CPA fell $21.01 → $17.48. The engine was Search Brand: orders 18 → 36 (+100%), revenue +123%, ROAS 3.89x → 6.54x. This confirms last week's read — the Jul 19 zero-order Saturday and the soft CVR were transient, not structural. No bids were changed to produce it.

This was a structural week — three campaigns were switched off: (1) NB PMax paused ~Jul 22 after the 4-week Bleeding-Rule window closed ($1,590 lifetime CDP spend → 7 CDP orders → ROAS 0.32). (2) HairMask Shopping paused, per last week's action item. (3) Demand Gen paused Jul 27 — but see the Demand Gen section: the advertorial funnel was measurably improving right up to the pause. That is the most important open decision this week.

✓ What Improved

  • Search Brand doubled — 36 orders / $2,467 vs 18 / $1,105; ROAS 6.54x, CPA $10.47 (target $13)
  • Shopping PPT CBO improved — 12 orders / $650 (vs 11 / $537), ROAS 2.73x
  • Blended CPA $17.48 — comfortably under the ~$22 blended target; CVR 5.01%
  • Waste cut decisively — PMax and HairMask both off; ~$400/wk of sub-1x spend removed
  • Demand Gen funnel improved every day — LP1 CTR 6.7% (Jul 21) → 17.3% (Jul 26), LP2 0.0% → 2.9% after the Jul 22 audience/ad-group optimization (see its section)

✗ What Needs Attention

  • Demand Gen paused Jul 27 while its funnel metrics were still climbing — decision needs a conscious re-review, not a default
  • CDP still records 0 orders / 0 ATC / 0 IC for Demand Gen while Tableau path data shows real LP2→checkout movement — an attribution gap that must be closed before judging the channel
  • Search NB (Reviews) deteriorating — spend +68% ($130 → $218) for fewer orders (4 → 3); ROAS 1.18 → 0.55
  • Bing revenue −49% ($451 → $232) on AOV $90 → $58.03; ROAS 5.19x → 2.69x
  • Non-brand is now nearly empty — with PMax + DGEN off, the only NB engine left is Reviews at ROAS 0.55

Shopify Validation

MetricJul 20–26 (A)Jul 13–19 (B)WoW
Gross Revenue (Shopify)*$19,803$20,507-3%
Refunds$0$0
Net Revenue (Shopify)*$19,803$20,507-3%
Orders (sample)*250250
AOV (Shopify)$79.21$82.03-3%
CDP 1dc Revenue (paid G+B scope)$3,505$2,522+39%

⚠ *Shopify totals are capped at the first 250 orders/week (API pagination) — both weeks hit the cap, so use for AOV/direction only. Store-wide Shopify revenue was roughly flat (-3.4%) while paid CDP revenue rose +39%. Read: paid took a larger share of a stable store — the recovery is real paid performance, not a store-wide demand surge.

Google Ads + Bing — CDP Account Overview (Jul 20–26)

Spend
$979
▲ +16% WoW ($841)
Orders (CDP 1dc)
56
▲ +40% WoW (40)
Revenue (CDP 1dc)
$3,505
▲ +39% WoW ($2,522)
ROAS (CDP)
3.58x
▲ +19% WoW (3.00x)
CPA (CDP)
$17.48
▼ -17% WoW ($21.01)
CTR
2.36%
▲ +44% WoW (1.64%)
CPC
$0.88
▲ -6% WoW ($0.93)
CVR (CDP)
5.01%
▲ +13% WoW (4.42%)

Scope: Google Search + Google Shopping + Bing Search (excludes Google NB PMax and Demand Gen — both have dedicated sections below — plus Bing Demand Gen and Facebook). PMax $159 / 0 CDP orders and Demand Gen $1,358 / 0 CDP orders are reported separately.

WoW Comparison — Google + Bing Search + Shopping (CDP)

MetricJul 20–26 (A)TargetJul 13–19 (B)WoWStatus
Revenue (CDP)$3,505$2,522+39%🟢 Up +39% WoW
Orders (CDP)5640+40%🟢 Up +40% WoW
ROAS (CDP)3.58x~2.6x3.00x+19%🟢 Above floor, up WoW
Spend$979$841+16%🟢 Scaled with revenue
CPA (CDP)$17.48$22$21.01-17%🟢 Below target
AOV (CDP)$62.60$57$63.04-1%🟢 Above target
CVR (CDP)5.01%~4.1%4.42%+13%🟢 Above floor
CTR2.36%1.2%1.64%+44%🟢 Above target
CPC$0.88$1.20$0.93-6%🟢 Below target

Blended — All PPC Channels (Google + Bing · CDP · 7-week trend)

All Channels CombinedGoogle Ads + Microsoft Ads · CDP · incl. NB PMax + Demand Gen
All PPC (incl. NB PMax + Demand Gen) · Jun 8–14 – Jul 20–26
▮ Orders (left)— CVR% (right)
▮ Revenue$ (left)— CPA$ (right)
▮ Spend$ (left)— ROAS (right)
▮ Clicks (left)— CTR% (right)
▮ AOV$ (left)— ROAS (right)
▮ CPC$ (left)— CTR% (right)

These blended trend charts include NB PMax (purple) and Demand Gen — Discovery (green) alongside Search/Shopping/Bing — switch to By Source to isolate each. Because both spent at near-zero recorded CDP ROAS, they pull the blended Total spend up and ROAS down versus the core-paid KPI cards above (which scope to Search + Shopping + Bing). Demand Gen appears from Jul 13–19 (launched Jul 17, paused Jul 27); PMax from Jun 22–28 (paused Jul 22).

Google Ads — Campaign Breakdown

🔵 Google Ads
Search BrandCPC_Core · Products · Modifiers
Search Brand · Jun 8–14 – Jul 20–26
▮ Orders (left)— CVR% (right)
▮ Revenue$ (left)— CPA$ (right)
▮ Spend$ (left)— ROAS (right)
▮ Clicks (left)— CTR% (right)
▮ AOV$ (left)— ROAS (right)
▮ CPC$ (left)— CTR% (right)
Search Brand · Search Impression Share × Impressions (platform, 7-week)
▮ Impressions (left)— Search IS% (right)

IS held flat at 90.3% (vs 90.4% last week) while brand-search impressions jumped +39% (1,381 → 1,921) — the highest of the 7 weeks. Read: branded demand came back, and we captured it at unchanged visibility. That is exactly the mirror image of last week (impressions down, IS flat) and confirms the Search-Brand swing is demand-driven, not auction-driven. No bid action required.

CampaignA SpendA OrdA RevCPA AB SpendB OrdB RevCPA BWoW
SR_US_BR_SpaDr_CPC_Core$30431$2,163$9.80$23214$899$16.61+121% ord · +141% rev
SR_US_BR_SpaDr_Products$675$304$13.44$464$205$11.56+25% ord · +48% rev
SR_US_BR_SpaDr_Modifiers$60$0$60$00 ord · 0 rev
Search Brand — orders doubled on returning branded demand; last week's dip was transient.

Combined Search Brand (CDP): 36 orders / $2,467 on $377 vs 18 / $1,105 last week. ROAS 3.89x → 6.54x, CPA $10.47 (target $13), AOV $68.53 (+12%). Both CVR and AOV recovered — the two things that broke last week.

Daily read (CDP orders, Jul 20–26): 8, 6, 4, 3, 3, 8, 4. No zero-order day this week — including Saturday Jul 25, which posted the week's best day (8 orders / $591). That is the cleanest possible evidence that the Jul 19 zero-order Saturday was a tracking artifact rather than lost demand, and it means part of last week's reported −40% was never real.

CPC_Core (do not touch): 31 orders / $2,163 at CPA $9.80, ROAS 7.12x. Platform IS 95.3% (4.7% lost to rank, 0% to budget) — brand defense intact at higher volume. Hold bids.

Products: 5 orders / $304, ROAS 4.52x at IS 75.3% with 24.7% lost to rank — the clearest remaining scale lever in the account now that PMax and DGEN are off. These are QS-10 brand-product terms on Manual CPC, so a measured bid-up applies directly. Modifiers: 0 orders on $6 again — chronic; consolidate into CPC_Core.

Search Non-BrandSR_US_NB_SpaDr_Reviews
Search NB · Jun 8–14 – Jul 20–26
▮ Orders (left)— CVR% (right)
▮ Revenue$ (left)— CPA$ (right)
▮ Spend$ (left)— ROAS (right)
▮ Clicks (left)— CTR% (right)
▮ AOV$ (left)— ROAS (right)
▮ CPC$ (left)— CTR% (right)
CampaignA SpendA OrdA RevCPA AB SpendB OrdB RevCPA BWoW
SR_US_NB_SpaDr_Reviews$2183$120$72.56$1304$154$32.51-25% ord · -22% rev
Search NB (Reviews) — spend up 68% for fewer orders. The one channel moving the wrong way.

SR_US_NB_SpaDr_Reviews (CDP): 3 orders / $120 on $218 — spend +67% while orders fell 4 → 3 and revenue −22%. ROAS 1.18 → 0.55x, CVR 5.80% → 2.86%. CPA $72.56 is still inside the $90 target, so this is not yet Critical Waste — but the direction is wrong and it is now the account's only non-brand engine.

Why it matters more this week: with NB PMax and Demand Gen both paused, Reviews carries 100% of non-brand acquisition. It runs Target CPA on a $140/day budget at IS 74.3% (25.7% lost to rank, 0% to budget) — rank-limited, not budget-limited. Do not scale it while CVR is falling; fix the landing-page routing first (the reviews cluster lands on /pages/reviews rather than the product page), then re-test.

Shopping BrandPPT_ACQ_CBO · HairMask_NightCream (paused)
Shopping Brand · Jun 8–14 – Jul 20–26
▮ Orders (left)— CVR% (right)
▮ Revenue$ (left)— CPA$ (right)
▮ Spend$ (left)— ROAS (right)
▮ Clicks (left)— CTR% (right)
▮ AOV$ (left)— ROAS (right)
▮ CPC$ (left)— CTR% (right)
CampaignA SpendA OrdA RevCPA AB SpendB OrdB RevCPA BWoW
SHP_TSD_US_PPT_ACQ_VOLUME_CBO_WK45$23812$650$19.86$18711$537$17.01+9% ord · +21% rev
SHP_TSD_US_BR_HairMask_NightCream_PURCHASE PAUSED$601$36$59.72$1522$276$76.14-50% ord · -87% rev
Shopping Brand — the headline decline is HairMask winding down; the core engine actually improved.

Segment (CDP): 13 orders / $686 on $298, ROAS 2.30x (vs 2.39x). Orders were flat and revenue −16% — but the composition changed completely.

1) SHP_TSD_US_PPT_ACQ_VOLUME_CBO_WK45 (do not touch) improved: 12 orders / $650 at ROAS 2.73x and CPA $19.86 — up from 11 orders / $537. On its own it is running at roughly the segment's historical ~2.7x.

2) SHP_TSD_US_BR_HairMask_NightCream_PURCHASE — now PAUSED. It spent $60 for 1 order before being switched off, after $152 / 2 orders (CPA $76) the week before. This closes last week's action item. Its removal is why segment revenue reads down while the campaign that matters went up — a quality-of-spend improvement, not a performance loss.

Bing Ads — Search Brand (Jul 20–26 vs Jul 13–19, CDP)

⚠ Bing metrics are pulled from CDP (Redash q2635, source_category = Bing Search) on the same Mon–Sun weeks as Google. Bing Demand Gen / Audience is currently inactive ($0) and is excluded.

🟠 Microsoft Ads (Bing)
Bing Search BrandCPC_Core · Modifiers · Products
Bing Blended · Jun 8–14 – Jul 20–26
▮ Orders (left)— CVR% (right)
▮ Revenue$ (left)— CPA$ (right)
▮ Spend$ (left)— ROAS (right)
▮ Clicks (left)— CTR% (right)
▮ AOV$ (left)— ROAS (right)
▮ CPC$ (left)— CTR% (right)
CampaignA SpendA OrdA RevCPA AB SpendB OrdB RevCPA BWoW
SR_US_BR_SpaDr_CPC_Core$582$169$28.98$755$451$15.03-60% ord · -63% rev
SR_US_BR_SpaDr_Products$171$36$17.41$60$0new · new
SR_US_BR_SpaDr_Modifiers$111$27$10.86$60$0new · new
Bing Search Brand — softer week on AOV, but the pattern is normal low-volume variance. Do not touch.

Segment (CDP): 4 orders / $232 on $86, ROAS 2.69x, CPA $21.55. The revenue drop (−48%) is almost entirely AOV: $90.12 → $58.03. Orders only moved 5 → 4.

Read carefully before acting: at 4–5 orders/week a single high-value basket swings ROAS by 2x. Last week CPC_Core alone posted a $90 AOV; this week the orders spread across all three campaigns (Products and Modifiers each converted 1, after 0 last week) at lower ticket. Spend was flat ($87 → $86), CPA $21.55 is still far inside the $13–$22 brand range on a blended basis. No bid or budget change — this is variance, not decay. Revisit if ROAS stays under 3x for two more weeks.

Google Demand Gen — Discovery-Only (Special Focus)

🟢 Google Ads · Demand Gen (Discovery)

DGEN_US_NB_TSD_HairSerum_Discovery · ID 24044975032 · launched Jul 17 · PAUSED Jul 27, 2026 · Target CPA (tCPA $130) · $200/day · advertorial funnel → offer.thespadr.com. Discovery-surface only (the "Discover arbitrage" play — click-billed, VTC optimization OFF). Business metrics = CDP (source_category = Google Discovery); delivery = Google Ads platform; on-site funnel = Tableau LP-path analysis. All charts in this section are per-day — the campaign has only ~10 days of life, and the story is entirely in the day-over-day trajectory.

Demand Gen · Hair Serum DiscoveryJul 20–26 · CDP + platform + LP funnel

Overview — Week A (Jul 20–26)

Spend (Wk A)
$1,358
▲ +72% WoW · $200/day
Orders (CDP 1dc)
0
ATC 0 · IC 0
Revenue (CDP 1dc)
$0
ROAS 0.00
CPA (CDP)
0 orders · tCPA $130
Impressions
39,399
▲ +39% WoW
Clicks
3,430
▲ +244% WoW
CTR
8.71%
▲ from 3.52% (Wk B)
Avg CPC
$0.40
▼ from $0.79 (Wk B)

✓ The optimizations worked — the on-site funnel improved every step of the way

On Jul 22 we concentrated the campaign on AG1_InMarket_HairCare and paused AG2_Affinity_CleanBeauty (the broad affinity audience that had been buying most of the volume). Three things moved immediately and together from that day forward:

  • Ad CTR jumped 6.1% → 10.7% (Jul 21 → Jul 26) — better-matched audience, same creative
  • CPC fell $0.46 → $0.38 — cheaper clicks despite the tighter targeting
  • LP1 (advertorial → offer page) more than doubled: 5.6% → 15.6% average, peaking at 17.3%
  • LP2 (offer page → checkout) went from a hard 0.0% to 2.1% average — the first time this funnel step registered at all

That is a clean, directionally correct optimization result: we bought better traffic, it engaged the advertorial more, and for the first time it started reaching checkout.

Advertorial Funnel — LP1 & LP2 Click-Through by Day

— CTR LP1 · advertorial → offer page — CTR LP2 · offer page → checkout

Source: Tableau LP-path analysis (user-path based), Jul 17–26. CTR LP1 = users whose path contains LP2 or who initiated checkout ÷ all users. CTR LP2 = users whose path contains LP2 and (path contains LP3 or checkout started) ÷ users whose path contains LP2. CTR LP3 has no recorded data in the period and is omitted. The dashed line marks Jul 22, the day the ad-group/audience change took effect.

DateCTR LP1 (advertorial → offer)CTR LP2 (offer → checkout)CTR LP3
2026-07-170.00%
2026-07-189.25%0.00%
2026-07-195.35%0.00%
2026-07-206.95%0.00%
2026-07-216.65%0.00%
2026-07-2214.55%3.25%
2026-07-2314.45%1.70%
2026-07-2417.00%1.20%
2026-07-2514.70%1.75%
2026-07-2617.30%2.85%

Green rows = after the Jul 22 optimization.

What the funnel gain is actually worth — and why it still wasn't enough at $200/day

The trajectory was right. The level was not yet viable. Applying the measured LP1/LP2 rates to the real daily click volume, Week A delivered roughly 453 users onto the offer page and about 8 checkout starts across the whole week. Even assuming a generous 40% checkout→purchase rate, that is ~3 orders on $1,358 of spend — an implied CPA around $410 against a $130 tCPA target.

So both statements are true at once, and the report should not collapse them: the optimizations produced a real, measurable improvement (LP1 5.6% → 15.6%, LP2 0.0% → 2.1%), and the funnel was still an order of magnitude away from paying for $200/day of Discovery clicks at the moment it was paused. The gap was closing, not closed.

The measurement problem is still open — and it is now the deciding question. CDP records 0 orders, 0 ATC, 0 initiate-checkout for this campaign all week, while the Tableau path data shows ~8 checkout starts. Those cannot both be right. Either CDP is not attributing offer.thespadr.com events to Google Discovery (the tracking gap flagged at build), or the path data is counting sessions the CDP never sees. Until that is reconciled, we do not actually know this campaign's true CPA — we only know its upper bound is bad and its trend was good.

Daily Delivery (platform · Jul 17–26)

▮ Spend$ (left)— Clicks (right)
▮ Impressions (left)— CTR% (right)
▮ Avg CPC$ (left)— CTR% (right)

The arbitrage mechanics improved steadily: after Jul 22 the campaign bought more clicks, cheaper, at a higher CTR — CPC $0.46 → $0.38 while CTR ran 6.1% → 10.7%. Spend stayed pinned at the $200/day cap throughout, so this is a pure efficiency gain, not a budget effect.

DateImprClicksCTRAvg CPCSpendOrders (CDP)
2026-07-1712,598560.44%$1.71$960
2026-07-189,4946006.32%$0.79$4730
2026-07-196,2383405.45%$0.65$2230
2026-07-207,9435036.33%$0.40$2020
2026-07-217,1484336.06%$0.46$1990
2026-07-225,0074709.39%$0.41$1950
2026-07-234,86849710.21%$0.38$1890
2026-07-244,77251110.71%$0.37$1880
2026-07-254,94351110.34%$0.38$1930
2026-07-264,71850510.70%$0.38$1930

Green rows = after the Jul 22 optimization. Jul 17–19 fall in Week B; Jul 20–26 are Week A.

Ad Group Performance (Jul 20–26)

Ad GroupStatusImprClicksCTRCostPlat Conv
AG1_InMarket_HairCareENABLED24,5582,53710.33%$97214
AG2_Affinity_CleanBeautyPAUSED14,3978856.15%$3814
AG3_CustomIntent_ThickerHairENABLED44481.80%$50
AG6_RiskLaneREMOVED000.00%$00
Ad group read — the deliberate concentration onto InMarket is what moved the funnel.

AG1_InMarket_HairCare is now the campaign — 2,537 clicks on $972 (72% of spend), at CTR 10.3% — versus AG2_Affinity_CleanBeauty at 6.1% before it was paused. In-market intent beat broad affinity on both click efficiency and, critically, on what happened after the click. AG3_CustomIntent_ThickerHair still barely serves (444 impressions, $5.32) — its custom-intent signal is too narrow to be eligible; it has never been a real test. AG6_RiskLane remains REMOVED per the build plan.

Ad Format Performance (Jul 20–26)

FormatImprClicksCTRCostPlat Conv
Multi-asset (image)39,3993,4308.71%$1,35818
Video responsive000.00%$00

The video responsive ads served zero impressions this week — the campaign ran entirely on multi-asset image ads at CTR 8.71%. Last week's recommendation to trim video is effectively already implemented by delivery. For a click-billed Discovery arbitrage, that is the correct state.

Benchmark — TSD Discovery vs. Shapermint Demand Gen (métricas promedio)

Comparación pedida por Juanma contra YT_US_DG_GMAIL_10210_EBRA_PPT_25W48_BROAD_IMG_META_TEST (Shapermint USA, acct 1037305439, Demand Gen, tCPA, $1,800/día). Solo métricas promedio — son campañas con presupuesto, marca y objetivo distintos, así que los volúmenes NO son comparables y no se muestran como resultado. Misma ventana (Jul 20–26) para ambas.

La comparación es válida en la mecánica de entrega: ambas corren prácticamente 100% en Discover (99% TSD · 100% SHM), o sea compiten en la misma subasta y el mismo formato (imagen). Escala: TSD $1,358 vs SHM $12,661 en la semana (~9.3x).

Métrica promedioTSD · DiscoverySHM · DG GmailΔ TSD vs SHMLectura
CPM — precio del alcance$34.47$29.71+16%TSD paga la subasta más cara
CTR — resonancia del creativo8.71%4.86%+79%TSD casi duplica a SHM
CPC — costo del click$0.40$0.61-35%TSD compra el click más barato
CVR plataforma — click → conv.0.53%1.12%-53%⚠ solo direccional — ver nota

Plataforma Google Ads, Jul 20–26. CPM/CTR/CPC son promedios puros y no dependen del presupuesto. La fila de CVR NO es comparable como negocio: cada campaña optimiza a acciones de conversión distintas (SHM = compra en su propio funnel; TSD = tCPA sobre el funnel advertorial) y, además, el CDP de TSD marca 0 órdenes esta semana — o sea su CVR real no está medida. Se incluye únicamente para mostrar dónde está la diferencia.

Solo ventana post-optimización (Jul 22–26)

Métrica promedioTSD · DiscoverySHM · DG GmailΔ TSD vs SHM
CPM$39.41$28.99+36%
CTR10.26%4.90%+109%
CPC$0.38$0.59-35%
CVR plataforma0.48%1.02%-53%

Aislando los días posteriores al cambio de audiencia del Jul 22, la brecha de CTR se ensancha: TSD llega a 10.26% vs 4.90% de SHM.

Lectura del benchmark — ganamos el click, perdemos lo que viene después

El arbitraje de Discover funcionó mejor en TSD que en SHM. TSD paga un CPM 16% más caro, pero convierte ese alcance en clicks a casi el doble de eficiencia (8.71% vs 4.86%), y el resultado neto es un click 35% más barato ($0.40 vs $0.61). Esa es exactamente la mecánica que buscábamos: en un modelo cobrado por click, un CTR alto compra impresiones gratis. El creativo editorial/advertorial de TSD está rindiendo por encima del benchmark de una cuenta mucho más grande y madura.

La diferencia real no está en el medio, está en el destino. SHM convierte el click a 1.12% y TSD a 0.53% (plataforma). Con SHM el click aterriza en un funnel de producto ya optimizado y con años de señal; con TSD aterriza en un advertorial nuevo, en un host nuevo, cuyo paso a checkout apenas empezó a registrar (2.1% LP2). Dicho de otra forma: el problema de TSD nunca fue el costo del tráfico — fue qué pasa después del click, que es la misma conclusión a la que llegan los datos de LP1/LP2 arriba.

Por qué SHM consigue un CPM más barato: corre ~9.3x el gasto diario ($1.800 vs $200) sobre una cuenta con mucho más histórico de conversión en Discover. Más volumen y más señal = mejor precio de subasta y un tCPA con datos para optimizar. No es una ventaja de creativo, es de escala y madurez — por eso el CPM no debería leerse como "SHM lo hace mejor".

Demand Gen Verdict — the pause is defensible, but do not treat it as a conclusion
  • #1 — Close the CDP attribution gap before re-launching anything on this funnel. CDP shows 0 ATC / 0 IC / 0 orders while the LP path data shows ~453 users on the offer page and ~8 checkout starts. Until offer.thespadr.com events resolve to Google Discovery in CDP, every ROAS/CPA number for this play is unmeasurable — and we will make the same blind decision again next time.
  • #2 — Record what the test actually proved. Concentrating on in-market intent lifted LP1 from 5.6% to 15.6% and unlocked LP2 from 0.0% to 2.1%. That is a reusable finding for any future advertorial funnel — the audience lever works; the creative/offer lever is the untested one.
  • #3 — If re-enabling, do it at $50–75/day, not $200. The funnel needed roughly a 5–10× improvement in checkout throughput to reach the $130 tCPA. At a quarter of the budget the same learning accrues at a quarter of the risk, and the LP2 step (currently ~2.1%) is the metric to optimize — not CTR, which is already healthy.
  • #4 — The real bottleneck is now the offer page, not the ads. 8.71% ad CTR and $0.40 CPC are good arbitrage numbers. 15.6% of clicks reach the offer page but only ~2.1% of those start checkout. Any next iteration should test the offer page itself (price framing, form length, mobile speed), not more traffic.

Google Non-Brand Performance Max — Final Read (Campaign Paused)

🟣 Google Ads · Performance Max

PMAX_US_NB_TSD_Serum_HairSerum · ID 23967212829 · launched Jun 30 · PAUSED ~Jul 22, 2026 · Maximize Conversion Value (no tROAS floor was ever set) · $60/day. Last delivery Jul 22. Business metrics = CDP (source_category = Google Max); delivery = Google Ads platform. This section closes the keep/kill question opened last week.

PMax NB · Serum + Hair SerumJul 20–26 (partial — 3 days) · CDP + platform

Overview — CDP (business) vs Platform (delivery)

Spend (Wk A)
$159
▼ -61% — 3 days only
Orders (CDP 1dc)
0
ATC 0 · IC 0
Revenue (CDP 1dc)
$0
ROAS 0.00
Lifetime CDP ROAS
0.32x
vs 0.50 target
Impressions
5,004
platform
Clicks
321
CTR 6.41%
Avg CPC
$0.50
platform
Ad Strength
AVERAGE
both asset groups
Bleeding-Rule outcome: the 4-week window opened Jun 30 and the campaign was paused on Jul 22, three weeks in. Lifetime CDP: $1,590 spend → 7 CDP orders → $510 revenue → ROAS 0.32x, against a 0.50 target and a $120 CPA target (actual lifetime CPA $227.19). It never reached the target in any single week. The pause is consistent with the Bleeding Rule and with last week's recommendation — with the caveat that the recommended intermediate step (setting a tROAS floor of ~0.4–0.5) was never tested, so we do not know whether a capped PMax could have worked.

7-Week Trend (CDP)

NB PMax · Jun 8–14 – Jul 20–26
▮ Orders (left)— CVR% (right)
▮ Revenue$ (left)— CPA$ (right)
▮ Spend$ (left)— ROAS (right)
▮ Clicks (left)— CTR% (right)
▮ AOV$ (left)— ROAS (right)
▮ CPC$ (left)— CTR% (right)

CDP (source_category = Google Max). PMax ran Jun 27 – Jul 22; earlier and later weeks are zero. Orders/ROAS are the real CDP numbers (platform all-conversions overcount badly — see below).

Network / Channel Split — WoW (Jul 20–26 vs Jul 13–19, platform)

NetworkCostClicksConv (plat)Conv ValuePlat ROAS
ABWoWABABABA vs BWoW
Discover$112$283-60%30773934$200$2011.79 vs 0.71+151%
Search$19$75-75%21602$0$1450.00 vs 1.93-100%
YouTube$29$51-43%126401$0$290.00 vs 0.57-100%
Search Partners$0$00000$0$00.00 vs 0.00
Total (platform)$160$410-61%32181937$200$3761.25 vs 0.92+36%

Week A covers only Jul 20–22 (3 days) before the pause, so absolute values are not comparable to Week B's 7 days — read the mix, not the totals. Conversions/value are platform (per-network CDP isn't available); CDP recorded 0 orders in Week A.

Post-mortem — Discover took the budget and never paid it back.

Discover consumed 70% of the final week's spend ($112 of $160) and, across the campaign's life, was consistently where the money went and the sales did not follow. Search — the one network with any conversion consistency — was squeezed to $19 on 2 clicks in the final days.

The structural cause was never fixed: the campaign ran uncapped Maximize Conversion Value with no tROAS floor for its entire life. With no efficiency constraint, Google optimized toward the cheapest available inventory (Discover at ~$0.38 CPC), which converted at a fraction of Search. The lesson to carry forward: never launch a PMax NB test without a tROAS floor, even a deliberately loose one — it is the only lever that keeps a mixed-network campaign honest.

Platform measurement was actively misleading here. Platform reported 378 "all conversions" in the final 3 days against 0 real CDP orders — the gap is Product-View / ATC / list-view micro-events. Anyone reading the Google Ads UI would have concluded this campaign was working. CDP was the only source that told the truth.

Asset Group Performance (platform · Jul 20–26)

Asset GroupStatusAd StrengthImprClicksCostAvg CPCConv
… - Hair SerumENABLEDAVERAGE4,968321$160$0.503
… - SerumPAUSEDAVERAGE360$00

Both asset groups finished at AVERAGE ad strength — never upgraded to GOOD/STRONG despite being flagged for three consecutive weeks. If this campaign is ever rebuilt, fix assets before spending: more headlines, long headlines, and image/video variety are free relevance, and an AVERAGE asset group handicaps every network it serves on.

Daily Delivery — Final 14 Days (platform)

DateImprClicksSpendConv (plat)
2026-07-091,31265$982
2026-07-101,426124$620
2026-07-111,43598$612
2026-07-121,684113$660
2026-07-131,23060$404
2026-07-142,151155$641
2026-07-151,22079$610
2026-07-162,336175$641
2026-07-172,121143$571
2026-07-181,36084$650
2026-07-192,430123$590
2026-07-201,19562$570
2026-07-211,895122$533
2026-07-221,914137$500

Delivery stops after Jul 22 — the pause date.

Channel Spend & Revenue Mix (Jul 20–26)

Spend Share — $979 total
Spend$979
Search Brand $377 (39%)
Search NB $218 (22%)
Shopping Brand $298 (30%)
Bing Search Brand $86 (9%)
Revenue Share (CDP) — $3,505 total
Revenue$3,505
Search Brand $2,467 (70%)
Search NB $120 (3%)
Shopping Brand $686 (20%)
Bing Search Brand $232 (7%)

Search Brand is back to being the engine — 39% of spend returning 70% of revenue at ROAS 6.54x, versus 34% / 44% last week. Shopping Brand takes 30% of spend for 20% of revenue (ROAS 2.30x). Bing is 9% of spend for 7% of revenue — still efficient at 2.69x despite the soft AOV week. Search NB is the mix problem: 22% of spend for only 3% of revenue.

Excluded from the donuts: Google NB PMax ($159 / $0 CDP revenue — paused, own section), Google Demand Gen ($1,358 / $0 recorded CDP revenue — paused, own section), Bing Demand Gen ($0), and Shopping Non-Brand (deprecated). Efficiency ranking by ROAS: Search Brand 6.54x → Bing SB 2.69x → Shopping Brand 2.30x → Search NB 0.55x.

Auction Insights — Impression Share (Jul 20–26 vs Jul 13–19)

Per-competitor domains require Standard API access (UI only). Below is campaign-level IS with WoW trend and where the loss goes (rank vs budget) — the actionable cut.

CampaignA ISB ISWoWLost RankLost Budget
CPC_Core (Brand)95.3%94.4%+0.9pp4.7%0.0%
Products (Brand)75.3%77.9%-2.6pp24.7%0.0%
Modifiers (Brand)96.0%88.0%+8.0pp4.0%0.0%
NB Reviews74.3%70.5%+3.8pp25.7%0.0%

Shopping Campaigns

CampaignA ISB ISWoWLost RankLost Budget
PPT Brand (Shopping)27.1%28.5%-1.4pp72.9%0.0%
HairMask (Shopping Brand) — now PAUSED10.0%16.1%-6.1pp8.3%85.9%
Read

Brand defense strengthened while volume grew — CPC_Core IS rose to 95.3% (from 94.4%) even as impressions jumped +39%. Holding share on a rising auction is the strongest possible signal that no bid change is needed.

Products is the live scale lever: IS 75.3% with 24.7% lost to rank and 0% to budget, on QS-10 brand-product terms under Manual CPC at ROAS 4.52x. With PMax and Demand Gen both off, this is the cleanest place to redeploy budget. NB Reviews at 74.3% IS is also rank-limited, but its CVR is falling — fix the LP routing before buying more of it.

Cross-Channel CDP Overview (Google + Bing)

ChannelA SpendA OrdA RevROAS AB SpendROAS BWoW SpendWoW Rev
Google Search (Brand+NB)$59539$2,5874.35x$4143.04x+44%+106%
Google Shopping (Brand)$29813$6862.30x$3392.39x-12%-16%
Bing Search (Brand)$864$2322.69x$875.19x-1%-48%
Google NB PMax (paused Jul 22)$1590$00.00x$4100.17x-61%-100%
Google Demand Gen (paused Jul 27)$1,3580$00.00x$7910.00x+72%
Total (excl PMax + Demand Gen)$97956$3,5053.58x$8413.00x+16%+39%

Google Search carried the week — $2,587 revenue on $595 (ROAS 4.35x vs 3.04x). Shopping held at 2.30x on less spend. Bing softened on AOV but stayed profitable. The two experimental non-brand engines are both off: PMax spent $159 for 0 orders before pausing Jul 22, and Demand Gen spent $1,358 for 0 recorded orders before pausing Jul 27. Together they consumed $1,518 this week — 61% of all paid spend — for no recorded revenue. Removing them is the single biggest efficiency change in the account this month.

Action Items

1
PMax NB — keep/kill decision resolved: PAUSED Jul 22 DONE
Closed at $1,590 lifetime CDP spend → 7 orders → ROAS 0.32x, never reaching the 0.50 target. Consistent with the Bleeding Rule. Note for the record: the recommended tROAS floor was never applied, so "capped PMax" remains untested.
2
HairMask Shopping — paused DONE
Closed last week's action item. It spent $60 for 1 order before switching off. The PPT CBO engine improved to 12 orders / ROAS 2.73x without it.
3
Close the CDP attribution gap on offer.thespadr.com URGENT
CDP shows 0 orders / 0 ATC / 0 IC for Demand Gen all week, while Tableau LP-path data shows ~453 users reaching the offer page and ~8 checkout starts. Both cannot be right. Verify the cross-domain linker, gclid persistence lp1→lp2, and whether offer.thespadr.com events map to source_category "Google Discovery". Without this, no advertorial funnel can be evaluated — and this is the second consecutive week the question has stayed open.
4
Demand Gen — decide consciously whether the pause is final DECISION
The campaign was paused Jul 27 while its funnel metrics were still climbing: LP1 5.6% → 15.6%, LP2 0.0% → 2.1%, CTR 6.1% → 10.7%, CPC $0.46 → $0.38. The trend was right; the level (~8 checkout starts/week on $1,358) was not yet viable. If re-enabling, do it at $50–75/day and optimize the LP2 step, not CTR. If not, document the audience finding (in-market beats affinity) before it is lost.
5
Search Brand Products — measured bid-up on QS-10 terms OPPORTUNITY
IS 75.3% with 24.7% lost to rank, 0% to budget, at ROAS 4.52x and CPA $13.44. Manual CPC means keyword bids apply directly. With both NB engines off, this is the account's cleanest incremental volume — step bids 10–15% and re-measure.
6
Search NB (Reviews) — fix LP routing before any scale REVIEW
Spend +67% for fewer orders; ROAS 1.18 → 0.55x, CVR 5.80% → 2.86%. It is now the only non-brand engine in the account, which makes the deterioration more consequential, not less. Route the hair-serum-reviews ad group to the product page rather than /pages/reviews, then re-measure before touching tCPA.
7
Non-brand coverage is now structurally thin MONITOR
With PMax and Demand Gen both paused, non-brand acquisition rests entirely on SR_US_NB_SpaDr_Reviews at ROAS 0.55x on $218/week. The account is now ~78% brand-defensive spend. That is efficient but not a growth position — plan the next non-brand test deliberately rather than by default.
8
Do not touch: CPC_Core (Google + Bing), PPT_ACQ Shopping MONITOR
Brand IS 95.3%, top terms QS 10, ROAS 7.12x. This week's recovery happened with zero bid changes — the correct action last week was to wait, and it was.

Do Not Touch List

🔒 Do not modify without explicit authorization:
The Spa Dr — WoW Report · Week of Jul 20–26, 2026 · Generated Jul 28, 2026
Data: CDP (Redash query 2635) · Google Ads API (9662795380) · Bing via CDP · Shopify (the-spa-dr.myshopify.com) · LP funnel via Tableau path analysis