How much conversion data does Smart Bidding need to work well?

Google generally recommends a minimum of around 15-30 conversions in the recent past (commonly 30 days) before Smart Bidding strategies like Target CPA have enough signal to bid reliably.

Can Smart Bidding be combined with manual bid adjustments?

Limited combination is possible depending on the specific strategy, but Smart Bidding already factors in many signals (device, location, audience) automatically, which is why Google restricts some manual bid adjustments once certain Smart Bidding strategies are active.

Which Smart Bidding strategy should a new account start with?

Maximize Conversions is commonly used to start, building initial conversion volume and data, before moving to a stricter Target CPA or Target ROAS once enough historical performance has accumulated to set a realistic target.

Does Smart Bidding work the same for Search, Shopping, and Display campaigns?

The underlying predictive bidding mechanism is conceptually similar, but the available strategies and specific signals used can vary somewhat by campaign type.

What happens during the Smart Bidding learning period?

Performance can be more volatile as the system calibrates to the account's data, Google generally advises against making frequent changes during this window, since it can restart or extend the learning process.

What Is Google Smart Bidding?

Google Smart Bidding is Google Ads' family of automated bidding strategies, including Target CPA, Target ROAS, Maximize Conversions, and Maximize Conversion Value, that use machine learning to set bids automatically for each individual auction based on the predicted likelihood and value of a conversion.

How Google Smart Bidding Works

Smart Bidding evaluates a wide range of contextual signals at auction time, device, location, time of day, remarketing list membership, browser, operating system, and many more, and uses predictive modeling to forecast that specific auction's conversion likelihood and, where relevant, expected value. It then sets a bid tailored to that forecast, rather than applying one fixed bid across every instance of a keyword or audience.

Each Smart Bidding strategy optimizes toward a different goal: Maximize Clicks and Maximize Conversions pursue volume within a budget without a strict cost target; Target CPA and Target ROAS pursue a specific cost or return target; and Maximize Conversion Value pursues the highest total conversion value achievable within budget, without a fixed ROAS target attached.

Examples & Use Cases of Google Smart Bidding

An ecommerce account switches from manual CPC bidding to Target ROAS once it accumulates enough conversion value data, and sees Smart Bidding begin favoring auctions predicted to produce larger orders rather than simply the most clicks.
A lead-gen business new to Google Ads starts on Maximize Conversions with no strict cost target, prioritizing building conversion volume and data before layering in a Target CPA constraint once enough history accumulates.
An account compares performance under manual bidding versus Smart Bidding for an identical campaign and budget, finding Smart Bidding achieves a meaningfully lower CPA once past its initial learning period.

How is Google Smart Bidding calculated?

Smart Bidding itself has no formula an advertiser applies, but the targets it can be configured against follow standard formulas, Target CPA = Spend ÷ Conversions, and Target ROAS = (Conversion Value ÷ Spend) × 100, with Google's underlying predictive models handling the actual per-auction bid calculation.

Google Smart Bidding is Google Ads' specific, branded implementation of the broader automated bidding and algorithmic bid management concepts described generally across the industry. Other platforms offer their own equivalent automated bidding systems under different names, but Smart Bidding refers specifically to Google's version and its particular set of available strategies.

How to Interpret Google Smart Bidding

Smart Bidding performance should be read over a rolling window after an initial learning period, typically one to two weeks after activation or a significant change, during which performance can be less stable as the system calibrates. A sustained trend in the target metric after that learning period is the meaningful signal, not day-to-day fluctuation.

Why Google Smart Bidding Matters

Smart Bidding lets an account react to per-auction signals, device, time, audience, location, combined together, at a speed and granularity no manual bidding process could realistically replicate, and is generally considered Google's recommended approach for accounts with sufficient conversion data, reflecting how central automated bidding has become to modern Google Ads management.

Frequently Asked Questions

How much conversion data does Smart Bidding need to work well?: Google generally recommends a minimum of around 15-30 conversions in the recent past (commonly 30 days) before Smart Bidding strategies like Target CPA have enough signal to bid reliably.
Can Smart Bidding be combined with manual bid adjustments?: Limited combination is possible depending on the specific strategy, but Smart Bidding already factors in many signals (device, location, audience) automatically, which is why Google restricts some manual bid adjustments once certain Smart Bidding strategies are active.
Which Smart Bidding strategy should a new account start with?: Maximize Conversions is commonly used to start, building initial conversion volume and data, before moving to a stricter Target CPA or Target ROAS once enough historical performance has accumulated to set a realistic target.
Does Smart Bidding work the same for Search, Shopping, and Display campaigns?: The underlying predictive bidding mechanism is conceptually similar, but the available strategies and specific signals used can vary somewhat by campaign type.
What happens during the Smart Bidding learning period?: Performance can be more volatile as the system calibrates to the account's data, Google generally advises against making frequent changes during this window, since it can restart or extend the learning process.

See NYX in action.

Book a demo