Smart Bidding — Google's suite of automated bid strategies powered by machine learning — has transformed how professional advertisers manage Google Ads campaigns. Understanding when, how, and why to use Smart Bidding (and when to avoid it) is essential knowledge for anyone managing significant Google Ads budgets.
How Smart Bidding Works
Smart Bidding uses machine learning to set bids for each individual auction based on dozens of signals that are impossible for human bidders to process in real time: device type, location, time of day, search query, audience membership, browser type, and hundreds of other contextual signals.
The key insight: Google's auction is not just about your bid — it's about the expected value of each impression. Smart Bidding optimizes not just for the cost of clicks, but for the predicted conversion probability of each specific user and context.
Smart Bidding Strategies Explained
Target CPA (tCPA): Tell Google what you want to pay per conversion. The algorithm adjusts bids to achieve that average CPA across your campaign. Requires at least 30–50 conversions per month for reliable optimization.
Target ROAS (tROAS): Tell Google your target return on ad spend. The algorithm adjusts bids to maximize conversion value at your target ROAS. Better for eCommerce with varying product values than flat CPA campaigns.
Maximize Conversions: No CPA target — Google maximizes conversion volume within your budget. Useful for campaigns building conversion data or when volume is prioritized over efficiency.
Maximize Conversion Value: Optimizes for total conversion value rather than conversion count. Ideal for eCommerce where purchase values vary.
Enhanced CPC (eCPC): A hybrid approach — you maintain manual bids but allow Google to adjust them up or down based on conversion probability signals. A gentler step toward full automation.
When Smart Bidding Works and When It Doesn't
Smart Bidding works best when: Conversion volume is sufficient (30+ conversions per month per campaign). Conversion tracking is accurate and complete. Budget is not severely constrained. The learning phase is allowed to complete without frequent changes.
Smart Bidding struggles when: Conversion volume is low and the algorithm lacks sufficient training data. Conversion tracking is missing events or has attribution issues. Budgets are too tight to explore beyond the most obvious converting audiences. External events (promotions, seasonality) create sudden performance shifts the algorithm hasn't experienced.
Smart Bidding Best Practices
Set realistic targets: Setting a tCPA or tROAS target that's significantly better than your current performance will cause the algorithm to underdeliver as it searches for impossible efficiency.
Give the learning phase time: Smart Bidding campaigns require a learning phase (typically 1–4 weeks) during which performance may be inconsistent. Avoid making significant changes during this period.
Use bid adjustments strategically: Some bid adjustments are compatible with Smart Bidding and can guide the algorithm without overriding it. Device and audience adjustments can be applied on top of Smart Bidding targets.
Monitor and adjust targets quarterly: As your business, competitive landscape, and costs change, revisit and update your Smart Bidding targets to reflect current conditions.
Transitioning from Manual to Smart Bidding
Jumping immediately from manual CPC to tCPA or tROAS can be disruptive. A gradual transition approach manages risk:
Start with Enhanced CPC: The least disruptive automated bidding option — Google adjusts your manual bids rather than fully taking over.
Move to Maximize Conversions: Remove the explicit CPA constraint and let Google optimize for volume while you establish baseline performance data.
Set Smart Bidding targets based on observed performance: Once you have performance baselines from Maximize Conversions, set tCPA or tROAS targets at or slightly worse than current performance, then gradually tighten toward your business goals.