Google’s Bidding and Budgeting Updates: Our Unvarnished Take

Google announced several new AI-powered bidding and budgeting updates for Search, Shopping, and Performance Max ahead of Google Marketing Live. Key updates include expanding Smart Bidding Exploration into Shopping and PMax campaigns, introducing journey-aware bidding for lead generation campaigns, and launching demand-led budget pacing that automatically adjusts spend based on shifts in consumer demand.

The updates:

• Reinforces Google’s continued shift toward AI-driven automation across bidding and budget management
• Could reduce manual pacing and optimization work during fluctuating demand periods
• Place greater importance on strong conversion signals, attribution, and campaign inputs as automation expands
  
  

What kind of guardrails do you need to put on these updates for them to work most effectively for you?

• Implement offline conversion tracking
  • These updates rely heavily on Google AI understanding the full customer journey, so accurate CRM integrations, offline conversion imports, enhanced conversions, and clear lead quality signals are critical.
  • Ideally, we would want clear differentiation between early-stage conversions (form fills, demos) and qualified conversions (SQLs, Opportunities). Otherwise, the system risks optimizing towards low-quality volume.
    
    
• Budget and pacing guardrails
  • It’s important to still monitor spend even with spend caps in place. Especially in B2B, aggressive pacing on “high-demand days” could overspend on lower-intent traffic if not monitored carefully.
    
    
• ROAS / CPA boundaries
  • tCPA/tROAS guardrails
  • negative keyword management
  • audience exclusions where needed
  
  
  
• Segmentation protections
  • I’d avoid immediately applying these updates universally. Brand, high-intent non-brand, remarketing, and prospecting campaigns should carry different levels of bidding flexibility.

How would you recommend structuring tests for these?

• Start with isolated campaign tests. I recommend choosing campaigns with stable historical performance, high conversion volume, and strong tracking implementation.
  
  
• During testing, it’s also important to avoid testing during seasonality spikes, landing page changes, or any outside changes.
  
  
• I would also recommend evaluating query quality, not just volume. The biggest risk is AI broadening into less relevant searches. I’d monitor:
  • search term themes
  • branded vs non-branded mix
  • conversion quality
  • down-funnel CRM performance
    
    
• It’s also important to give tests enough learning time. These systems need stable learning periods. I’d recommend 2–4 weeks minimum, enough conversion volume to stabilize, and minimal changes throughout the testing period. Otherwise, results become noisy and difficult to attribute.

What are some reasons these might give you skepticism?

• • what users actually searched
  • why spend shifted
  • what drove performance changes
    
    

Expanding into “less obvious queries” sounds promising, but many advertisers have seen broad AI expansion increase lower-quality leads because of: 

• low-intent traffic
• spam
• unqualified form fills
• inflated top-funnel metrics

• • Google AI optimizes toward the signals it’s fed.
  • If CRM feedback is weak, the system may optimize toward cheaper conversions instead of meaningful business outcomes.
    
    

I'd also have real concerns about potential budget volatility, which can be especially challenging for accounts with strict monthly pacing requirements. Demand-led pacing could create uneven daily spend swings that: 

• complicate forecasting
• create pacing anxiety
• overinvest during short-term spikes

It's also hard to predict how effective any beta will be out of the gate. Historically, some Google betas perform very differently across:

• B2B vs. ecommerce
• long sales cycles vs. short sales cycles
• low-volume vs. high-volume accounts

For those reasons, I’d be cautious about assuming universal success from aggregate benchmark numbers.

With all of those caveats in place, I do think advertisers should start testing and getting fluent with the new tools since Google is obviously heading in that direction. Just make sure you have perspective on what could go wrong -- and put guardrails in place to prevent that from the start.

JDM helps growth teams identify the content, search, paid media, and AI visibility gaps that affect pipeline. We turn those gaps into practical campaigns that help challenger brands get found, trusted, and chosen. Schedule a growth strategy call.