The PMax Cannibalization Test: What Happens When Paid Stops Eating SEO

◆ The Mission

A major retailer with a large paid search investment wanted to answer a question that most brands avoid asking: Is PMax taking credit for conversions that organic search would have won anyway?

The hypothesis was that PMax campaigns were aggressively bidding on brand terms — intercepting clicks that SEO would have captured at zero marginal cost — and inflating paid ROAS while suppressing organic performance metrics.

The team designed a controlled brand exclusion test to find out.

---

◆ The Setup

Prior to the test, PMax campaigns ran without brand exclusions alongside the organic program. The combined SEO + paid ROAS during this period was $6.82 — a solid number, but one that masked how the channel mix was actually performing.

When the brand exclusion test launched, PMax was forced to run exclusively on non-brand terms. The SEO program would absorb whatever demand PMax had previously intercepted on brand queries.

---

◆ What the Data Showed

▸ SEO Clicks Jumped Immediately

Within the brand exclusion window, SEO clicks on the top 50 tracked terms increased 125–175% per day. Of that increase, 92% came from brand terms — confirming exactly what was suspected: PMax had been suppressing organic brand visibility, not supplementing it.

Two-thirds of all click increases during the exclusion period came directly from brand terms that PMax had previously been capturing.

▸ Overall Visits Went Up, Not Down

The most critical test: would forcing PMax to non-brand terms cause a traffic shortfall?

It didn't. Overall visits (SEO + PMax combined) increased 40–50% after brand exclusions were activated. PMax didn't disappear — it just stopped competing with organic on terms the brand already owned.

• 66% of the SEO visit increase was directly attributable to brand click growth (confirmed via GSC)
• Clicks to store location pages increased +51% (approximately 42k additional clicks), representing roughly 1/3 of total SEO gains

▸ Revenue Attribution Shifted — Total Revenue Didn't

PMax revenue during the exclusion period declined -58% (-~$62k per week on average). This sounds alarming until you look at what happened to SEO revenue simultaneously.

SEO revenue increased +50% (+~$45k per week on average) during the same window. On peak weeks, SEO revenue was up 85–90%.

Internal store revenue data confirmed no drastic changes in actual sales. The revenue hadn't disappeared — it had moved from PMax attribution to SEO attribution. PMax had been taking credit for conversions that organic was driving.

---

◆ The ROAS Story

The most revealing comparison came from isolating the brand exclusion window:

Period	Combined ROAS	Context
Prior period (PLA/LIA + SEO)	$6.82	Lower spend, no brand exclusions
Test period (PMax + SEO)	$5.24	Nearly double the spend, brand cannibalizing
Brand exclusion window only	$10.03 vs $4.66	Prior period vs PMax without exclusions

The prior period was $1.58 more profitable per dollar spent overall. During the brand exclusion comparison window specifically, the gap widened to $5.37 per dollar — largely because PMax spend had nearly doubled while ROAS compressed.

A key context note: higher PMax spend was a contributing factor to ROAS compression. The prior period ran on roughly 50% of the budget (under $20k/week average vs $35–40k/week during the test period). Spending more on a channel that was cannibalizing organic made the combined performance look worse on both sides.

---

◆ The Takeaway

PMax didn't generate the brand demand — it intercepted it. The exclusion test didn't hurt the business. It made the attribution honest.

This case study surfaces a pattern that's common in mature brands with strong organic presence: PMax's default behavior is to target the highest-converting, lowest-risk queries — which are often brand terms that SEO already owns. The result is inflated paid revenue, suppressed organic metrics, and a combined ROAS that appears healthy but is subsidized by cannibalizing zero-cost traffic.

The learnings from this test informed a revised channel strategy: PMax budgets redirected toward non-brand and prospecting use cases, with brand demand protected as an organic asset.

For brands evaluating PMax performance, this test offers a clear framework: run the exclusion, watch what organic does, and let the data decide the channel split.