How an E-Commerce Brand Lost $35K to Performance Max Cannibalization and Broad Match Bleed

$35,000 burned in 30 days. Performance Max was stealing branded conversions at 10x the cost. Broad match was turning the search budget into a money furnace. And the Google Ads↗ dashboard showed everything in green.

This is the hidden danger of modern Google Ads: the platform's automated features are designed to spend your budget efficiently from Google's perspective, not yours. When a mid-market skincare e-commerce brand came to us with a "good but not great" 1.5x ROAS, they assumed they needed better creative or a bigger budget. They didn't. They needed to stop Google from eating their own money.

The Background

The client sold premium skincare products through their Shopify store, averaging $180,000/month in revenue with a healthy mix of organic, email, and paid traffic. They had a well-known brand name in the natural skincare niche with strong organic search volume — approximately 12,000 branded searches per month.

Their Google Ads structure, built by a previous agency, consisted of:

• 1 Performance Max campaign: $500/day, targeting all product categories
• 1 Search campaign: $650/day, containing 4 ad groups with 40+ broad match keywords each
• 1 Shopping campaign: $200/day (standard Shopping, running alongside PMax)

Total daily spend: $1,350/day ($40,500/month). Reported ROAS: 1.5x. The previous agency had called this "solid performance" and recommended increasing budget.

What Went Wrong

Our audit revealed two massive budget leaks that were invisible in the high-level reporting.

Budget Leak 1: Performance Max Brand Cannibalization

Performance Max campaigns, by design, serve ads across all Google surfaces — Search, Shopping, Display, YouTube, Gmail, and Discover. What Google doesn't prominently tell you is that PMax will aggressively bid on your branded search terms, often outbidding your own Search campaigns for traffic that would have converted anyway.

We pulled the PMax insights report and asset group details:

PMax Traffic Source	Spend Share	Conversions	CPA
Search (branded terms)	45%	68% of PMax conversions	$62
Search (non-branded)	15%	12% of PMax conversions	$95
Shopping	25%	15% of PMax conversions	$45
Display/YouTube/Discover	15%	5% of PMax conversions	$180

45% of the PMax budget — $225/day — was going to branded search terms. These are queries like "BrandName skincare," "BrandName vitamin C serum," and "BrandName coupon code." These users were already going to buy. The branded search campaign was capturing them at $6 CPA. PMax was intercepting those same users at $62 CPA — a 10x markup for the same conversion.

Even worse, PMax was winning the auction against the brand's own Search campaign because PMax uses a different bidding mechanism. The Search campaign's branded keywords were being suppressed because PMax was outbidding them in the same auction.

The net effect: the brand was paying $62 for conversions that cost $6 through their own branded search campaign. PMax was claiming credit for 68% of its conversions from users who were going to convert regardless.

Budget Leak 2: Broad Match Keyword Hemorrhage

The Search campaign had 40+ broad match keywords per ad group — a structure that gives Google maximum freedom to match your ads to queries that are, at best, tangentially related to your products.

We downloaded the search terms report for the past 30 days. Out of 2,847 unique search terms that triggered ads:

• 12% were directly relevant (product names, ingredient names, purchase-intent queries)
• 28% were marginally relevant (general skincare queries with no purchase intent)
• 60% were completely irrelevant or competitor terms

Examples of irrelevant search terms that consumed budget:

Search Term	Match Type	Cost	Conversions
"skincare routine for teens"	Broad	$340	0
"how to make face cream at home"	Broad	$280	0
"dermatologist near me"	Broad	$195	0
"is retinol bad for you"	Broad	$175	0
"[Competitor Brand] review"	Broad	$420	0
"free skincare samples"	Broad	$310	0
"skincare for men"	Broad (they only sold women's products)	$260	0

The broad match keywords "natural skincare" and "vitamin C serum" alone had matched to over 400 irrelevant queries, consuming $8,500/month with a 0.1% conversion rate.

There were zero negative keywords in the entire account. None. Not a single one. The previous agency had never implemented a negative keyword strategy.

Root Cause Analysis

The structural failures traced back to three root causes:

1. Performance Max treated as "set and forget": PMax was launched with default settings and never analyzed at the asset-group or search-term level. Google's recommendation to "let the algorithm learn" was interpreted as "never look at the data."

2. Broad match without negative keywords is an open wallet: Broad match can be effective when paired with robust negative keyword lists that prevent irrelevant matching. Without negatives, broad match is essentially telling Google to spend your money on whatever it wants.

3. No search term monitoring process: The previous agency reviewed campaign-level metrics monthly but never downloaded or analyzed search term reports. Without this feedback loop, irrelevant queries accumulated unchecked.

The Fix

We implemented a complete restructuring over 14 days, with results visible within 48 hours of the first changes.

Step 1: Brand Exclusion from PMax (Day 1)

Google introduced brand exclusion lists for Performance Max in 2023, but many advertisers still don't use them. We created a brand exclusion list containing:

• All brand name variations (including common misspellings)
• Product line names unique to the brand
• Brand + product combinations

This immediately forced PMax to compete only on non-branded queries and Shopping placements, where it could provide genuine incremental value.

Step 2: Search Campaign Restructuring (Days 1-5)

We dismantled the 4 bloated ad groups and rebuilt the search campaign using a modified SKAG (Single Keyword Ad Group) approach:

Branded Campaign ($150/day):

• Exact match branded terms only
• Max CPA target: $10
• Highest priority

High-Intent Non-Branded Campaign ($300/day):

• Product-specific queries: [vitamin c serum], [retinol cream], [hyaluronic acid moisturizer]
• Exact and phrase match only
• Max CPA target: $35

Category Campaign ($150/day):

• Category queries: "natural skincare products," "organic face cream"
• Phrase match with extensive negatives
• Max CPA target: $50

Competitor Campaign ($50/day):

• Competitor brand terms (with proper trademark compliance)
• Exact match only
• Separate budget to prevent bleed

Step 3: Negative Keyword Nuclear Option (Days 1-3)

We built comprehensive negative keyword lists organized by theme:

• DIY/Homemade: homemade, DIY, recipe, make your own, at home (82 terms)
• Medical/Clinical: dermatologist, doctor, prescription, clinic (45 terms)
• Free/Cheap: free, sample, cheap, discount code, coupon (38 terms)
• Irrelevant Demographics: men, teen, baby, pet, dog (32 terms)
• Informational: how to, what is, wiki, definition, reddit (55 terms)
• Competitor Terms (for non-competitor campaigns): 15 competitor brand names

Total: 267 negative keywords deployed across campaign-level and shared negative keyword lists.

Step 4: Search Term Monitoring Script (Day 5)

We deployed a Google Ads script that runs daily and:

• Flags any search term that spent more than $20 without a conversion
• Auto-adds irrelevant search terms to negative keyword lists (with manual review queue)
• Generates a weekly search term health report sent to the team
• Tracks the ratio of relevant vs. irrelevant search terms over time

Step 5: PMax Restructuring (Days 5-10)

With branded terms excluded, we restructured PMax to focus on where it adds genuine value:

• Split into product-category asset groups (instead of one catch-all)
• Added high-quality product images and lifestyle photography
• Set audience signals using first-party customer data (purchasers, add-to-cart)
• Excluded Display and Discover placements (underperforming for this brand)

The Results

Metric	Before	After	Change
ROAS	1.5x	5.2x	+247%
CTR	2.1%	4.8%	+129%
CPC	$1.50	$0.65	-57%
CPA	$75	$18	-76%
Branded CPA	$62 (PMax)	$6 (Search)	-90%
Monthly Wasted Spend	~$21,000	~$1,200	-94%
Monthly Revenue (Ads)	$60,750	$210,600	+247%

The revenue increase wasn't just from better efficiency — by reallocating the $21,000/month that was being wasted on irrelevant queries and brand cannibalization, we funded genuine new customer acquisition that drove incremental revenue.

Key Takeaways

1. Always exclude brand from Performance Max: Unless you're a new brand with zero organic search presence, PMax will cannibalize your branded traffic. Brand exclusion lists should be the first thing you configure when launching PMax.

2. Broad match without negatives is budget arson: Broad match can work well for discovery and scale, but only with comprehensive negative keyword lists. Without them, you're giving Google permission to spend your money on whatever tangentially related query it finds.

3. Green dashboards don't mean healthy campaigns: A 1.5x ROAS looks acceptable at campaign level. But when you discover that half the spend is on irrelevant queries and the other half is cannibalizing your own branded traffic, the "real" ROAS on incremental spend is deeply negative.

4. Search term reports are mandatory reading: Download and review search term reports at least weekly. If you find this tedious, automate it with Google Ads scripts. But never skip it.

5. Structure controls outcomes: The difference between 1.5x and 5.2x ROAS wasn't creative, wasn't bidding strategy, wasn't audience targeting — it was campaign structure. How you organize campaigns, ad groups, match types, and negative keywords determines where your money goes.

Prevention Checklist

• Enable brand exclusion lists for all Performance Max campaigns
• Use exact and phrase match as default; deploy broad match only with negative keyword coverage
• Build negative keyword lists before launching campaigns (minimum 200 terms for e-commerce)
• Review search term reports weekly
• Deploy automated search term monitoring scripts
• Separate branded and non-branded campaigns with independent budgets
• Audit PMax search term insights monthly (limited visibility, but use what's available)
• Set campaign-level and shared negative keyword lists
• Compare branded CPA across PMax vs. Search campaigns quarterly
• Track incremental revenue, not just attributed revenue

---

Explore our media buying services →