Performance Max campaigns are genuinely useful. They're also a black box that will quietly cannibalize your best-performing campaigns if you don't actively manage what they can and can't touch. Both things are true, and most takes on PMAX seem to land on one side or the other without acknowledging both.
After running PMAX across dozens of DTC brands in varying categories, spend levels, and account structures, here's where we've landed: PMAX works well in specific conditions, and it creates problems in specific conditions. The conditions matter more than the product itself.
What PMAX actually is (and what it's doing)
Performance Max is Google's fully automated campaign type that serves across all Google inventory: Search, Shopping, Display, YouTube, Gmail, and Maps. You provide asset groups, signals, and budget. Google decides where to show ads, who to show them to, and in what format. The algorithm learns from your conversion data and optimizes toward your stated goal.
The appeal is obvious. One campaign, all surfaces, automated optimization. The catch is that "automated optimization" means Google will route budget toward whatever it thinks will convert at the lowest cost. In practice, this often means brand search traffic. Branded keywords have high conversion rates, low competition, and clean attribution. PMAX loves them. If you're not actively managing this, your PMAX campaign may be spending heavily on traffic that would have converted organically anyway.
PMAX doesn't tell you where it's spending. You can see asset performance and some audience signals, but the channel-level budget allocation is not fully transparent. That opacity is the core challenge.
When PMAX works well
When you have strong conversion volume. The algorithm needs data to optimize. If you're doing fewer than 30-50 conversions per month through Google, PMAX won't have enough signal to learn from and performance will be erratic. With higher volume, it has the data to make meaningful decisions and the results can be genuinely strong.
When you have a large, well-structured product catalog. For brands with hundreds of SKUs, PMAX can find the right product for the right intent signal in ways that manually managed Shopping campaigns struggle to do at scale. This is where the automation earns its keep.
When you want incremental reach, not just conversion efficiency. PMAX can surface your brand on YouTube, Display, and Discovery in ways that traditional Shopping or Search campaigns won't. If incremental reach is the goal alongside conversion, PMAX has more tools to achieve that than a standard campaign type.
When PMAX creates problems
When PMAX cannibalizes branded traffic on Google Ads. This is the most common issue. PMAX will bid on your branded keywords, pull traffic from your existing Brand Search campaigns, and report the conversion as a PMAX win. Your Brand Search ROAS looks like it's declining. PMAX ROAS looks incredible. In reality, the same customer converted, just through a more expensive path. Use brand exclusions to protect your branded campaigns from PMAX overlap.
When it competes with high-performing Shopping campaigns. If you have product-specific Shopping campaigns that are already optimized and performing well, PMAX may redirect budget from them toward less efficient inventory. Run experiments and use campaign priority settings to understand whether PMAX is additive or just redistributing existing demand.
When your asset quality is low. PMAX generates ad combinations from your asset groups. If the creative assets are mediocre, the system will run mediocre creative across all Google inventory at scale. Poor creative in PMAX is harder to diagnose than in a standard campaign because you don't always know which asset combination is running where.
How to structure PMAX so it doesn't run away
Brand exclusions are non-negotiable. Add your brand terms as negative keywords at the account level or use audience exclusions to ensure PMAX is finding genuinely new customers, not converting existing branded intent at a premium.
Run separate campaigns for your core performers. Don't consolidate everything into PMAX. Keep your best-performing Shopping and Search campaigns running in parallel, monitor for overlap, and use PMAX as additive reach rather than a replacement for controlled campaigns.
Use audience signals, not audience targeting. PMAX treats audience inputs as signals, not constraints. Feed it your customer lists, purchasers, high-LTV segments. The algorithm uses these to find similar profiles. Don't expect it to restrict delivery to only those audiences, because that's not how it works.
Watch the true contribution, not just the ROAS. PMAX ROAS is usually flattering because it captures easy conversions. The question is whether PMAX is driving net-new revenue or redistributing existing demand at higher cost. Run a campaign-off test for a defined period if you need clarity on true channel contribution.
Build your account structure around what PMAX will do by default
PMAX is not a set-it-and-forget-it solution, and it's not something to avoid. It's a powerful tool with specific failure modes that require active management. The brands that get the most from it understand what the algorithm will optimize toward by default, then build their account structure to prevent the outcomes they don't want while letting the automation do what it's actually good at. That requires knowing what it's actually good at, which most brands skipped. For a look at YouTube-first Google strategy done right, see how we drove results for CordaRoy's on YouTube Ads.
More from The Brief
Running PMAX but not sure it's incremental?
We audit Google Ads accounts regularly and can help you understand whether your PMAX campaigns are generating new revenue or just redistributing existing demand.