google-ads-pmax-retail-optimizer

PMax Retail Optimizer

Specific optimizations for Performance Max campaigns with product feeds (PMax Retail). Focused on e-commerce accounts using Merchant Center feeds.

Quick Diagnostic

PMAX RETAIL HEALTH CHECK
==========================

1. Zombie Products?
   -> >20% products without impressions (30d) --> ACTION NEEDED

2. Listing Group Filters set?
   -> Only UNIT_INCLUDED (all) --> SUBOPTIMAL

3. Custom Labels active in feed?
   -> No custom labels --> OPPORTUNITY MISSED

4. Asset group spend distribution?
   -> >60% budget to 1 asset group --> REBALANCE

5. Product ROAS spread?
   -> Top 20% products = 80% revenue --> SEGMENT

1. Zombie Inventory Detection

Data Retrieval

Zombie products = products with 0 impressions in the last 30 days.

-- Products WITH impressions (active products)
SELECT segments.product_title, segments.product_brand,
    segments.product_type_l1, segments.product_item_id,
    metrics.impressions, metrics.clicks,
    metrics.cost_micros, metrics.conversions
FROM shopping_performance_view
WHERE segments.date DURING LAST_30_DAYS
ORDER BY metrics.impressions ASC
LIMIT 100

-- Total shopping performance (for zombie % calculation)
SELECT segments.product_title, metrics.impressions,
    metrics.clicks, metrics.cost_micros
FROM shopping_performance_view
WHERE segments.date DURING LAST_30_DAYS
AND metrics.impressions = 0
LIMIT 200

Zombie Analysis Framework

ZOMBIE PRODUCT CLASSIFICATION
==============================

TYPE 1: Feed Quality Zombies
+-- Cause: Disapprovals, poor data quality
+-- Detection: Products with warnings/disapprovals in MC
+-- Solution: Fix feed issues (title, image, price)
+-- Impact: High -- these products CANNOT run

TYPE 2: Competition Zombies
+-- Cause: Too much competition, bids too low
+-- Detection: Product exists in feed, no disapprovals, 0 impressions
+-- Solution: Higher bids OR separate asset group with more budget
+-- Impact: Medium -- bid increase can help

TYPE 3: Low Demand Zombies
+-- Cause: Little search volume for this product
+-- Detection: Niche product, few search terms
+-- Solution: Accept OR improve titles with more popular terms
+-- Impact: Low -- sometimes there's just no demand

TYPE 4: Structural Zombies
+-- Cause: Product not in active listing group filter
+-- Detection: Product in feed but outside asset group scope
+-- Solution: Expand listing group filters
+-- Impact: High -- configuration error

Zombie Solution Matrix

Zombie Rate	Severity	Action
<10%	Normal	No action, focus on performers
10-20%	Attention	Check feed quality, optimize titles
20-40%	Problem	Separate asset group for zombies, feed audit
40-60%	Serious	Full feed review, restructure campaigns
>60%	Critical	Feed fundamentally broken or insufficient budget

Action: Separate Asset Group for Zombie Products

Via google_ads_mutate:

operations=[
    {"assetGroupOperation": {"create": {
        "resourceName": "customers/{CID}/assetGroups/-10",
        "campaign": "customers/{CID}/campaigns/{PMAX_CAMPAIGN_ID}",
        "name": "Asset Group - Revive Zombies"
    }}},
    {"assetGroupListingGroupFilterOperation": {"create": {
        "assetGroup": "customers/{CID}/assetGroups/-10",
        "type": "UNIT_INCLUDED",
        "vertical": "SHOPPING",
        "caseValue": {"productCustomAttribute": {
            "index": "INDEX1",
            "value": "zombie"
        }}
    }}}
]

Requires: Custom Label 1 = "zombie" in the feed for products with 0 impressions.

2. Listing Group Filter Optimization

Analyze Current Filters

-- Listing group filters per asset group
SELECT asset_group.name, asset_group_listing_group_filter.type,
    asset_group_listing_group_filter.listing_source,
    asset_group_listing_group_filter.path
FROM asset_group_listing_group_filter

Filter Strategy Decision Tree

LISTING GROUP FILTER STRATEGY
================================

How many products do you have?
|
+---> <50 products
|   +---> UNIT_INCLUDED per asset group (all)
|       +-- 1-2 asset groups max
|       +-- Segmentation is not needed
|
+---> 50-200 products
|   +---> FILTER PER CATEGORY
|       +-- productType L1 or productBrand
|       +-- 3-5 asset groups
|       +-- Each with own assets/signals
|
+---> 200-1000 products
|   +---> FILTER PER CATEGORY + PERFORMANCE
|       +-- productType + productCustomAttribute (margin/performance)
|       +-- 5-10 asset groups
|       +-- Separate group for bestsellers
|       +-- Separate group for zombies
|
+---> >1000 products
    +---> MULTI-CAMPAIGN + GRANULAR FILTERS
        +-- Campaign per main category
        +-- Asset groups per subcategory + margin
        +-- Custom labels for performance tiers
        +-- Regular label updates in feed

Filter Types

Filter Type	When	Example
UNIT_INCLUDED	All products in scope	Everything
UNIT_EXCLUDED	Exclude specific products	Exclude low margin
SUBDIVISION + UNIT_INCLUDED	Only specific subset	Only brand "Nike"
productBrand	Segment by brand	Nike, Adidas, own brand
productType	Segment by category	Clothing, Shoes, Accessories
productCustomAttribute	Business-specific	Margin tier, season
productItemId	Individual product	Specific SKU
productChannel	Online vs local	ONLINE, LOCAL

Mutate: Setting Listing Group Filters

Brand-based filter via google_ads_mutate:

operations=[
    {"assetGroupListingGroupFilterOperation": {"create": {
        "assetGroup": "customers/{CID}/assetGroups/{AG_ID}",
        "type": "SUBDIVISION",
        "vertical": "SHOPPING",
        "caseValue": {"productBrand": {}}
    }}},
    {"assetGroupListingGroupFilterOperation": {"create": {
        "assetGroup": "customers/{CID}/assetGroups/{AG_ID}",
        "type": "UNIT_INCLUDED",
        "vertical": "SHOPPING",
        "caseValue": {"productBrand": {"value": "Nike"}}
    }}},
    {"assetGroupListingGroupFilterOperation": {"create": {
        "assetGroup": "customers/{CID}/assetGroups/{AG_ID}",
        "type": "UNIT_EXCLUDED",
        "vertical": "SHOPPING",
        "caseValue": {"productBrand": {"value": ""}}
    }}}
]

3. Custom Label Strategy for PMax

5 Custom Labels -- Recommended Usage

CUSTOM LABEL STRATEGY
========================

LABEL 0: MARGIN TIER
+-- high_margin   -> Products with >50% margin
+-- medium_margin -> Products with 20-50% margin
+-- low_margin    -> Products with <20% margin
+-- USE: Separate campaigns per margin tier
    +-- High margin = higher ROAS target, more budget

LABEL 1: PERFORMANCE TIER
+-- bestseller -> Top 20% by revenue (last 90 days)
+-- standard   -> Middle 60%
+-- slow_mover -> Bottom 20%
+-- new        -> Products <30 days in assortment
+-- USE: Separate asset groups per tier
    +-- Bestsellers = most budget + best assets

LABEL 2: SEASON
+-- spring, summer, autumn, winter
+-- evergreen (season-independent)
+-- holiday (christmas, black friday)
+-- USE: Seasonal campaigns activate/pause
    +-- Budget shift to relevant seasonal products

LABEL 3: LIFECYCLE
+-- new_arrival    -> First 30 days
+-- core           -> Standard assortment
+-- on_sale        -> Discounted
+-- clearance      -> Clearance sale
+-- last_chance    -> Almost sold out
+-- USE: Align messaging per lifecycle
    +-- New arrival = "New!", Clearance = "Last items"

LABEL 4: PRICE RANGE
+-- premium     -> >$200
+-- mid_range   -> $50-$200
+-- budget      -> <$50
+-- USE: Audience alignment
    +-- Premium = high-income audiences, Budget = deal seekers

Label Update Frequency

Label	Update Frequency	Source
Margin Tier	Monthly	ERP / purchasing data
Performance Tier	Weekly	Google Ads conversion data
Season	Per season (4x/year)	Marketing calendar
Lifecycle	Daily	Inventory management
Price Range	On price change	Product feed

4. Budget Allocation Framework

Analyze Asset Group Performance

-- Asset group performance with spend share
SELECT asset_group.name, asset_group.status,
    campaign.name, metrics.impressions, metrics.clicks,
    metrics.cost_micros, metrics.conversions,
    metrics.conversions_value
FROM asset_group
WHERE segments.date DURING LAST_30_DAYS
ORDER BY metrics.cost_micros DESC

Budget Allocation Decision Matrix

BUDGET ALLOCATION FRAMEWORK
============================

STEP 1: Calculate ROAS per asset group
+-- ROAS = conversions_value / (cost_micros / 1000000)
+-- Rank asset groups by ROAS
+-- Identify top, middle, and low performers

STEP 2: Assess spend share
+-- Spend share = asset group cost / campaign cost
+-- Compare with expected distribution
+-- Ideal: High ROAS asset groups get more budget

PROBLEM SIGNALS:
+-- High ROAS asset group <15% spend -> Underfunded
+-- Low ROAS asset group >40% spend -> Budget waste
+-- Asset group with 0 impressions -> Configuration error
+-- 1 asset group >70% spend -> No diversification

STEP 3: Redistribute budget
+-- Option A: Adjust campaign-level budget
|   +-- Split asset groups across multiple campaigns
|       with own budgets (most control)
+-- Option B: Optimize asset group content
|   +-- Improve assets in underperforming groups
|       so Google allocates more budget
+-- Option C: Refine listing group filters
    +-- Target more precisely so relevance increases

Budget Shift Protocol

MAX 20% BUDGET SHIFT PER WEEK
===============================

Week 1: Identify opportunity
+-- Asset Group A: ROAS 5.0x, 15% spend
+-- Asset Group B: ROAS 1.5x, 45% spend
+-- Action: Consider splitting into separate campaigns

Week 2: First shift
+-- Create Campaign 2 for top performers
+-- Move Asset Group A to Campaign 2
+-- Budget Campaign 2: +20% of original budget
+-- Budget Campaign 1: -20%

Week 3: Monitor
+-- Check ROAS both campaigns
+-- Check impressions/conversions distribution
+-- No changes (learning period)

Week 4: Evaluate and continue
+-- Performance improved -> Consider further shift
+-- Performance equal -> Stable, good
+-- Performance worsened -> Roll back
+-- Never shift more than 20% per week

5. Feed + PMax Coexistence Optimization

Feed Quality Impact on PMax

FEED QUALITY -> PMAX PERFORMANCE
====================================

TITLE QUALITY (Highest impact)
+-- Keyword-first titles -> Better search matching
+-- Brand + product + attribute format
+-- Max 150 chars, first 70 chars most important
+-- Example: "Nike Air Max 90 Men's Running Shoes Black Size 10"

IMAGE QUALITY
+-- White background (Shopping requirement)
+-- High resolution (min 800x800, ideally 1200x1200)
+-- Product fills 75-90% of frame
+-- Lifestyle images as supplement (not primary)

PRICE COMPETITIVENESS
+-- Competitive price -> More impressions
+-- Sale price set -> Shopping annotations
+-- Price higher than competitor -> Less delivery
+-- Google compares prices in real-time

AVAILABILITY
+-- Out of stock -> Automatically paused
+-- Limited stock -> Lower priority
+-- Fast sync needed (at least daily)
+-- Feed delays -> Missed sales

GAQL: Product-Level ROAS Analysis

-- Top performing products (for bestseller label)
SELECT segments.product_title, segments.product_brand,
    segments.product_item_id, segments.product_type_l1,
    metrics.impressions, metrics.clicks,
    metrics.cost_micros, metrics.conversions,
    metrics.conversions_value
FROM shopping_performance_view
WHERE segments.date DURING LAST_30_DAYS
AND metrics.cost_micros > 0
ORDER BY metrics.conversions_value DESC
LIMIT 50

-- Underperformers (for exclusion or feed improvement)
SELECT segments.product_title, segments.product_brand,
    segments.product_item_id,
    metrics.impressions, metrics.clicks,
    metrics.cost_micros, metrics.conversions
FROM shopping_performance_view
WHERE segments.date DURING LAST_30_DAYS
AND metrics.cost_micros > 0
AND metrics.conversions = 0
ORDER BY metrics.cost_micros DESC
LIMIT 50

Product Exclusion Strategy

WHEN TO EXCLUDE PRODUCTS
==========================

EXCLUDE if:
+-- Product has >$50 spend and 0 conversions (30 days)
+-- Product ROAS <25% of average account ROAS
+-- Product is discontinued but still in feed
+-- Product has structurally poor reviews (doesn't convert)
+-- Product market price is too high vs competition

HOW TO EXCLUDE:
+-- Option 1: Remove product from listing group filter
|   +-- assetGroupListingGroupFilterOperation UNIT_EXCLUDED
+-- Option 2: Custom label "exclude" in feed
|   +-- Asset group filter exclude on label
+-- Option 3: Pause product in Merchant Center
|   +-- Stops for all Shopping/PMax campaigns
+-- Option 4: Lower priority via separate asset group
    +-- Less budget, not fully stopped

WHEN NOT TO EXCLUDE:
+-- Product is new (<14 days, insufficient data)
+-- Product is seasonal (wait for the season)
+-- Feed quality is the problem (fix the feed first)
+-- Budget is too low for all products (increase budget)

Check Search Terms & Apply Negative Keywords (v20+)

PMax now supports campaign-level negative keywords (added in v20, 2024).
Use search term data to find irrelevant queries and add negatives.

STEP 1: Pull search terms
google_ads_run_gaql(query="
  SELECT
    campaign.name,
    campaign_search_term_view.search_term,
    metrics.impressions,
    metrics.clicks,
    metrics.cost_micros,
    metrics.conversions
  FROM campaign_search_term_view
  WHERE campaign.advertising_channel_type = 'PERFORMANCE_MAX'
    AND segments.date DURING LAST_30_DAYS
    AND metrics.cost_micros > 0
  ORDER BY metrics.cost_micros DESC
  LIMIT 200
")

STEP 2: Identify irrelevant high-spend terms
+-- Terms with high cost but 0 conversions
+-- Competitor brand names (if not desired)
+-- Category terms outside your scope

STEP 3: Add campaign-level negative keywords via google_ads_mutate
+-- Use campaign_criterion with negative=true on PMax campaign
+-- Start with exact match negatives for clearly irrelevant terms

Optimization Cadence

WEEKLY:
+-- Check zombie product rate
+-- Review asset group spend distribution
+-- Scan search terms for irrelevance (use campaign_search_term_view)
+-- Monitor ROAS trends per asset group

BIWEEKLY:
+-- Update performance tier labels in feed
+-- Replace "Low" performance assets
+-- Evaluate listing group filter effectiveness
+-- Check new product performance

MONTHLY:
+-- Full ROAS analysis per product
+-- Consider budget reallocation
+-- Custom label updates (margin, lifecycle)
+-- Competitor price comparison
+-- Asset group consolidation or splits

QUARTERLY:
+-- Season label updates
+-- Campaign structure evaluation
+-- Feed quality audit
+-- Strategic repositioning

Tools Reference

Action	Tool	Recipe/Query
Shopping performance per product	google_ads_run_gaql	Shopping performance recipe
Asset group performance	google_ads_run_gaql	PMax asset group performance recipe
View asset labels	google_ads_run_gaql	PMax asset group assets recipe
Listing group filters	google_ads_run_gaql	Listing groups recipe
Merchant Center check	google_ads_run_gaql	Merchant Center link recipe
Set listing group filter	google_ads_mutate	assetGroupListingGroupFilterOperation
Create asset group	google_ads_mutate	Recipe #18/#19 (assetGroupOperation)
Change campaign budget	google_ads_mutate	Recipe #3
Create PMax Retail	google_ads_mutate	Recipe #19
Look up accounts	google_ads_list_accounts	--