Lesson: ROAS guardrails were applied before creative refresh in Session 2. This was wrong — creative must improve before constraining PMax. All 3 guardrails were rolled back.
1
Fix Tracking
Dedup, YouTube, USA tag
2
Creative Refresh
PMax UK first, then DE/FR
3
Reactivate Revenue
DE Ritalin, ADHD fix
4
New Campaigns
UK/DE/FR Search
5
Guardrails
ROAS floors last
All Issues — Full Detail
Critical — Fix immediately
C1 Conversion double-counting
5 primary purchase actions + duplicate "Google Shopping App Purchase" catch-all. Every CPA/ROAS metric is inflated. Smart Bidding optimises against wrong signals.
Fix: Change "Google Shopping App Purchase" from Primary to Secondary. Verify each market-specific action fires independently. Monitor 7 days post-change.
C2 YouTube follow-on views still enabled
YouTube video views counted as conversions, distorting Smart Bidding. Currently secondary but still enabled — should be disabled entirely.
Fix: Disable "YouTube follow-on views" conversion action entirely in Google Ads UI (not just set to secondary).
C3 SERP USA Brand — broken conversion tracking
£344/mo spend, 88% impression share, only 5.1 conversions. ROAS: 0.40x. Brand search should convert at 10-20% CTR. This is a tracking failure, not a demand problem.
Fix: Test US checkout tag with Google Tag Assistant on US VPN. Check "US Purchase" conversion action attribution window. Fix or pause campaign.
High — Fix this week
H1 PMax UK — 24-month stale creative
2 asset groups from Jan/Apr 2024. ROAS collapsed from 3.53x (launch) to 1.46x (current) as spend scaled 6x with zero asset refresh. Single biggest fixable issue in the account.
Fix: Full creative refresh: 15 headlines, 5 long headlines, 5 descriptions, 15+ images. Use SCENE framework for product-in-scene shots. Allow 14-day learning period after upload.
H2 Seasonal/promotional assets still running
"New Year Sale | code FOCUS10" still active in late March/April. Damages brand credibility and wastes impression share on expired promotions.
Fix: Remove all seasonal asset groups and replace with evergreen messaging. Check all PMax campaigns for stale promo copy.
H3 SERP UK ADHD — structurally broken
15.3% impression share (84.7% lost). £832/mo at 0.97x ROAS. New ad group created in Session 2 has keywords but no RSA ads attached — cannot serve.
Fix: Add RSA ads to new ad group. Add medication-term negative keywords (Ritalin, Adderall, Concerta, etc.). Verify landing page alignment. Target: recover 84.7% lost IS.
H4 PMax DE Ritalin stuck at £0.01/day
Historically best campaign: 2.43x ROAS, £14.84 CPA across £36,625 lifetime spend (2,468 conversions). ENABLED but zero impressions — likely blocked on product feed issue.
Fix: Check Merchant Center feed attachment. If feed exists, increase budget to £200/day. If no feed linked, configure it first. Potential: +£13,000/mo revenue.
H5 RecoverUp conversion pollution
Conversion actions from a separate brand (RecoverUp) exist in the Brainzyme account. Potentially contaminating conversion data and Smart Bidding signals.
Fix: Disable all RecoverUp-branded conversion actions. They belong in a separate Google Ads account.
Medium — Optimise this month
M1 CTV/Connected TV waste
CTV placements in PMax likely consuming budget with low conversion attribution.
Fix: Pull PMax placement reports. Exclude CTV if CPA is 2x+ above campaign average.
M2 SERP DE Brand — 54% impression share
Losing 46% of brand impressions in Germany. Competitors may be bidding on Brainzyme brand terms.
Fix: Run auction insights. Increase budget or bid if budget-limited. Improve ad relevance if rank-limited.
M3 80+ paused campaigns — account bloat
Navigation difficulty, accidental re-enablement risk, messy change history.
Fix: Archive campaigns paused 90+ days with zero conversions. Do in batches.
M4 No ad schedule adjustments
Bidding equally across all hours despite likely conversion rate variance by time of day.
Fix: Pull 90-day hourly data. Suppress bids 1am-6am if CVR significantly below average. Est. impact: 5-10% CPA improvement.
M5 No device bid adjustments
Mobile vs desktop CPA likely differs but no adjustments applied.
Fix: Pull device performance. Apply bid adjustments based on CPA variance. Est. impact: 5-15% CPA improvement.
All projections use: CPA = Spend / Conversions, ROAS = Revenue / Spend, 30.4 days/month. These are estimates, not guarantees. Actual results depend on market conditions, creative quality, and learning phase duration.
Fixing dedup does not change actual revenue. It reveals the real numbers Smart Bidding should learn from. Expect reported conversions to drop 15-25% but actual business results stay the same.
2. PMax UK creative refresh (H1)
Current
Spend£8,298/mo
Revenue£12,115/mo
ROAS1.46x
CPA£27.51
Asset age24 months
Projected (2.20x ROAS)
Spend£8,298/mo
Revenue£18,256/mo
ROAS2.20x
CPA£18.88
Revenue uplift+£6,141/mo
3. PMax DE Ritalin re-enable (H4)
Historical performance
Lifetime spend£36,625
Lifetime conv2,468
ROAS2.43x
CPA£14.84
Projected (£200/day, 2.20x ROAS)
Monthly spend£6,080
Monthly revenue£13,376
Monthly conv410
Revenue uplift+£13,376/mo (new)
4. USA Brand tracking fix (C3)
Current
Spend£344/mo
Revenue£138/mo
Conv5.1
ROAS0.40x
Projected (5.00x ROAS post-fix)
Spend£344/mo
Revenue£1,720/mo
Conv52
Revenue uplift+£1,582/mo
Combined projected impact
Current state
Monthly spend£27,442
Monthly revenue£44,380
ROAS1.62x
CPA£26.00
Projected state (all fixes, realistic)
Monthly spend£26,080
Monthly revenue£65,479
ROAS2.51x
Revenue uplift+£21,099/mo
Break-even ROAS at ~35% margin: 2.86x
Every £1 above break-even = pure profit. Target 3.50x = £0.64 profit per £1 ad spend.
30/60/90-Day Action Plan
30 Days — Fix critical + high priority
#
Action
Severity
Week
Est. Impact
1
Fix conversion double-counting (C1)
Crit
1
Accurate bidding signals
2
Disable YouTube follow-on views (C2)
Crit
1
Cleaner conversion data
3
Investigate USA Brand tracking (C3)
Crit
1
Recover £344/mo + £1,582 rev
4
Remove RecoverUp conversions (H5)
High
1
Clean data
5
PMax UK creative refresh (H1)
High
2
ROAS 1.46x → 2.20x (+£6,141/mo)
6
Remove seasonal assets (H2)
High
2
Brand credibility
7
Fix SERP UK ADHD structure (H3)
High
3
Recover 84.7% lost IS
8
Investigate PMax DE Ritalin feed (H4)
High
3
+£13,376/mo revenue
60 Days — Medium priority optimisations
#
Action
Severity
Est. Impact
9
Review CTV placements (M1)
Med
Reduce PMax waste
10
Fix SERP DE Brand IS (M2)
Med
Capture 46% more DE brand traffic
11
Archive paused campaigns (M3)
Med
Account hygiene
12
Add ad schedule adjustments (M4)
Med
5-10% CPA improvement
13
Add device bid adjustments (M5)
Med
5-15% CPA improvement
14
PMax DE creative refresh (German)
High
Improve 1.36x DE ROAS
15
PMax FR creative refresh (French)
High
Improve 1.22x FR ROAS
90 Days — Review, best practices, new launches
#
Action
Severity
Est. Impact
16
Full performance review (pre vs post)
Med
Validate all improvements
17
Apply ROAS guardrails (if creative stable)
Med
Lock in gains
18
Standardise naming (L1)
Low
Maintainability
19
Shared negative lists (L2)
Low
Efficiency
20
FR Brand ad strength (L3)
Low
+IS in France
21
Audience signals on Search (L4)
Low
Data for future optimisation
22
Launch SERP UK Problem-Aware Search
High
New UK traffic via Brain Quiz
23
Launch SERP DE Konzentration Search
High
Zero competition in DE
24
Launch SERP FR Anti-Fatigue Search
High
Uncontested FR cluster
Changes Already Made (Sessions 2 & 3)
#
Change
Status
1
Paused 7 Demand Gen campaigns (£5,968/mo waste at 0.21x ROAS)
Created new ADHD ad group with problem-aware keywords
Done
9
Set SERP UK ADHD budget to £30/day
Done
10-12
Set Target ROAS 200% on PMax UK/DE/FR
Rolled back
13-15
Removed Target ROAS → uncapped Max Conv Value (UK/DE/FR)
Done
Still pending (manual / UI required)
Action
Why manual
YouTube follow-on views → disable
Conversion settings in Google Ads UI
Fix PMax DE Ritalin feed
Merchant Center configuration
Create 5 new Search campaigns
create_campaign API tool is broken (T5)
Add RSA ads to SERP UK ADHD new ad group
Not yet attempted via API
Test USA checkout tag
Requires US VPN + Tag Assistant
Final Review List
Instructions: Print this page. Review each item. Tick the checkbox and circle your decision (Approve / Modify / Skip / Discuss). Write notes in the space provided. Then go through each item with Claude one-by-one via voice — say "Item 1" to start, and Claude will execute or discuss based on your decision.
ITEM 1 — CRITICAL
Fix conversion double-counting
Change "Google Shopping App Purchase" from Primary → Secondary. This reveals true CPA (~£32) and ROAS (~1.30x). No revenue change — just accurate data. Smart Bidding will recalibrate over 7 days.
APPROVEMODIFYSKIPDISCUSS
Notes:
ITEM 2 — CRITICAL
Disable YouTube follow-on views conversion
Disable entirely (not secondary). Video views are not purchase signals. Removing them cleans Smart Bidding input.
APPROVEMODIFYSKIPDISCUSS
Notes:
ITEM 3 — CRITICAL
Investigate USA Brand conversion tracking
£344/mo, 88% IS, 5.1 conv. Need to test US checkout tag via Tag Assistant + US VPN. If tag broken, fix it. If tracking is correct and it genuinely converts this poorly, consider pausing.
APPROVE INVESTIGATIONPAUSE NOWDISCUSS
Notes:
ITEM 4 — HIGH
Remove RecoverUp conversion actions
Separate brand's conversion actions exist in Brainzyme account. Should be disabled or moved to own account. Do you recognise these and know their origin?
APPROVEDISCUSSSKIP
Notes:
ITEM 5 — HIGH
PMax UK full creative refresh
Replace all 24-month-old assets. Generate 15 headlines, 5 long headlines, 5 descriptions, 15+ images (product-in-scene via Flux Kontext Max + lifestyle via Nano Banana Pro + text-heavy via Ideogram). Estimated cost: ~£12 for images. Est. impact: ROAS 1.46x → 2.20x (+£6,141/mo revenue).
APPROVEAPPROVE COPY ONLYDISCUSS BRIEFSKIP
Notes:
ITEM 6 — HIGH
Remove seasonal/promotional assets
"New Year Sale | code FOCUS10" still running in April. Replace with evergreen messaging across all PMax campaigns.
APPROVEMODIFYSKIP
Notes:
ITEM 7 — HIGH
Fix SERP UK ADHD structure
Add RSA ads to new ad group (currently has keywords but no ads). Add medication-term negatives. Only 15.3% impression share captured. Needs landing page review.
APPROVEDISCUSS COPYSKIP
Notes:
ITEM 8 — HIGH
Investigate and re-enable PMax DE Ritalin
Best campaign historically (2.43x ROAS, £14.84 CPA). Currently at £0.01/day. Need to check Merchant Center feed. If feed OK, set budget to £200/day. Potential: +£13,376/mo.
APPROVE INVESTIGATIONAPPROVE + SET BUDGETDISCUSS
Notes:
Final Review List (continued)
ITEM 9 — MEDIUM
Review and exclude CTV placements
Connected TV placements in PMax may be consuming budget with low conversion attribution. Need to pull placement report first.
APPROVEDEFER TO 60-DAYSKIP
Notes:
ITEM 10 — MEDIUM
Fix SERP DE Brand impression share (54%)
Losing 46% of DE brand impressions. Run auction insights, increase budget or bid.
APPROVEDEFERSKIP
Notes:
ITEM 11 — MEDIUM
Archive 80+ paused campaigns
Account bloat from old campaigns. Remove those paused 90+ days with zero conversions.
APPROVEDEFERSKIP
Notes:
ITEM 12 — MEDIUM
Add ad schedule bid adjustments
Suppress bids during 1am-6am low-conversion hours. Est. impact: 5-10% CPA improvement.
APPROVEDEFERSKIP
Notes:
ITEM 13 — MEDIUM
Add device bid adjustments
Apply mobile vs desktop CPA-based bid adjustments. Est. impact: 5-15% CPA improvement.
APPROVEDEFERSKIP
Notes:
ITEM 14 — BUDGET
Budget reallocation model
Current allocation: UK £8,741 (32%) / DE £11,718 (43%) / FR £4,014 (15%) / US £2,139 (8%) = £27,442 total (over £20k target).
Proposed: UK £8,500 / DE £7,000 / FR £3,000 / US £500 = £19,000 + £1,000 for new campaigns.
Do you want to reduce to £20k target, or maintain current levels?
CUT TO £20KKEEP CURRENTCUSTOM SPLITDISCUSS
Notes:
ITEM 15 — CREATIVE BRIEF
PMax DE + FR creative refresh (German/French)
After UK refresh, apply same process to DE (German copy) and FR (French copy, "anti-fatigue mentale" angle). Phase 3 copy from strategy dashboard ready. Est. image cost: ~£24 for both markets.
APPROVE AFTER UKAPPROVE ALL NOWDISCUSS
Notes:
ITEM 16 — NEW CAMPAIGNS
Launch 3 new Search campaigns (UI-created, PAUSED)
(a) SERP UK Problem-Aware: "can't concentrate", "brain fog" → Brain Quiz
(b) SERP DE Konzentration: German focus keywords (zero competition)
(c) SERP FR Anti-Fatigue Mentale: uncontested French cluster
All created PAUSED via Google Ads UI (API tool broken). Review before enabling.
APPROVE ALLUK ONLYDEFER TO 90-DAYDISCUSS
Notes:
ITEM 17 — GUARDRAILS
ROAS guardrails — deferred
ROAS floors (150-200%) on PMax campaigns. Currently rolled back. Will only apply after creative refresh has been live 30+ days and shows stable improvement. Do you agree with this sequencing?
AGREE — DEFERAPPLY SOONERDISCUSS
Notes:
ITEM 18 — NAMING + HYGIENE
Low-priority best practices (L1-L4)
Naming conventions, shared negative lists, FR ad strength, audience signals. All low-impact improvements for 90-day window. Approve as a batch or discuss individually?