Google has been pushing their Google Ads Performance Max (PMax) campaigns for some time, promising to harness the power of AI and machine learning to optimize ad placement and maximize conversions. However, as the team at three&six has discovered, the reality of PMax is far less rosy than Google would have us believe.

After diving deep into the keyword data from these campaigns, the three&six team has uncovered some troubling truths about PMax. These campaigns are known for their lack of transparency, which makes it difficult for hotel marketers to understand how they operate. This is a major concern as it leaves hotel marketers in the dark about where their advertising dollars are being spent and nearly impossible to make informed decisions about ad spend.

Even more concerning is the fact that while PMax policies suggest it respects keyword targeting and complements existing search campaigns, the small print allows it to take the lead in almost every campaign we analyzed. In campaigns without brand exclusions, a staggering 96.5% of all clicked keywords were brand terms, despite the presence of separate brand campaigns targeting those same keywords. Even in campaigns with brand exclusions, 93.75% still had some level of brand keyword clicks.

This brand keyword cannibalization is not only wasteful, but it also undermines the effectiveness of carefully crafted brand campaigns. When PMax is allowed to run unchecked, it can siphon off valuable brand traffic and revenue, with 98.9% of all generated revenue coming from brand keyword terms in campaigns without brand restrictions.

But the issues with PMax extend beyond just brand keyword targeting. The AI behind PMax seems to struggle with understanding the nuances of the hotel industry, often making questionable decisions when it comes to keyword targeting. From targeting generic restaurant and attraction keywords without any qualification for hotel-related searches, to bidding on competitor hotel names with negative ROAS, to confusing short-term rental properties with generic real estate listings, PMax's AI has demonstrated a clear lack of industry-specific intelligence.

Perhaps most alarming is the case of the North American hotel campaign with worldwide targeting, where PMax decided to allocate 53.9% of all clicks to India, while the entire US accounted for a mere 1.18%. This disproportionate geo-targeting is a clear indication that PMax's AI is far from perfect and requires close monitoring and intervention from human advertisers.

It's not all doom and gloom, however. The three&six team has found that PMax can be controlled to some extent, and Google has recently introduced a way to request negative keywords be added to the account. Early results show that with proper optimization techniques, keyword click wastage can be cut, saving budget and improving ROAS. Additionally, PMax has shown some success in targeting relevant hotel-related non-brand keyword terms and unique amenities.

However, the overall performance of PMax campaigns has been erratic at best, and the instances of wasted ad spend and brand cannibalization are too significant to ignore. Hotel marketers must be cautious and vigilant when utilizing PMax for their advertising, closely monitoring their campaigns and demanding greater transparency from Google.

The true cost of PMax's lack of transparency and questionable AI decision-making is ultimately borne by the hotel advertisers who place their trust and budgets in Google's hands. It is crucial for the industry to hold Google accountable and push for greater visibility and control over these campaigns. Until then, hotel marketers would be wise to approach PMax with a healthy dose of skepticism and question the data to ensure the most effective use of their marketing dollars.

Tristan Heaword
Co-Founder
three&six

View source