Skip to content
Faisal Mahmud TusherDigital marketing, SEO & analytics
with Faisal Tusher

Glossary term

Smart Bidding

Google Ads' automated bid strategies that set each auction's bid from conversion signals.
Paid Mediaplatform featureUpdated

Definition

Google Ads Help describes Smart Bidding as a subset of automated bid strategies that use machine learning to optimise for conversions or conversion value in every auction, a capability it calls auction-time bidding. Four strategies belong to the family:

StrategyGoalOptional constraint
Maximise conversionsAs many conversions as the budget allowsA target cost per conversion (target CPA)
Maximise conversion valueAs much conversion value as the budget allowsA target return on ad spend (target ROAS)
Target CPAConversions at about a chosen average costSet as the target on maximise conversions
Target ROASConversion value at about a chosen returnSet as the target on maximise conversion value

In the current interface the two target strategies are set as options on their maximise counterparts rather than chosen separately, and older accounts may still show the standalone names. Other automated strategies, such as maximise clicks and target impression share, are not Smart Bidding because they do not optimise towards conversions. The family as a whole is one kind of Bid Strategy.

The signals the system uses at auction time include device, location, time, the search query, audience membership, browser and operating system, among others. Advertisers cannot see the bid or set those signals by hand; they see outcomes and set the goal.

In practice

Three conditions decide whether Smart Bidding works. The conversions it optimises towards must be tracked completely, because a conversion the model never sees is one it will not pursue. The conversion actions counted in the campaign's goal must be ones the business values, because every included action is treated as worth chasing. And for value-based strategies, the values passed must reflect what the business earns. The Smart Bidding strategies guide covers choosing between the four, setting the target number and changing it without restarting learning.

Data volume still matters. A campaign with many recent conversions predicts better than one with few, and a campaign whose conversions are rare holds its target loosely. Where volume is thin, the usual options are to optimise towards a more frequent action that correlates with the real goal, or to pool campaigns under one portfolio strategy. Conversions recorded in Google Analytics can be imported as conversion actions, which the GA4 and Google Ads integration tutorial walks through.

Illustrative example — not a real result

A dental lab narrows what counts as a conversion

Bellhaven Dental Lab ran target CPA at £40 on a conversion action that counted every contact form submission. Over a quarter the campaign delivered 300 conversions at £38 each, but the lab's inbox showed that 170 were job applications and supplier pitches, leaving 130 case enquiries at an effective £88 each. The strategy had found the cheapest form fills it could, and job seekers were cheap. The lab split the form, made case enquiries the only conversion action in the goal, and reset the target to £90. The next quarter delivered 150 case enquiries at £78 each, so the cost per real lead fell even though the reported CPA rose.

How it interacts with match types and structure

Smart Bidding sees the actual search in each auction, which changes how Keyword Match Types behave. Under manual bidding, a broad keyword buys every related search at roughly one price; under Smart Bidding the system can bid high on a related search that looks likely to convert and low on one that does not. That is why Google recommends broad match with Smart Bidding, and why the keyword match types guide treats the two decisions together. Account structure matters for the same reason: fewer campaigns sharing a goal give each strategy more conversions to learn from.

What it cannot see

The model optimises the conversions and values it is told about, and nothing else. It does not know about offline outcomes, such as a lead becoming a customer weeks later, unless those outcomes are imported. It does not know the margin on a sale unless margin is built into the conversion values. And it cannot tell whether a conversion would have happened anyway: a strategy that wins many conversions from people already searching for the brand looks excellent while adding little. Questions of incrementality have to be answered outside the bidding system. The wider pattern, automation that optimises a stated objective with signals the advertiser cannot inspect, is discussed in AI in ad platforms.

Bid Strategy is the general concept; Smart Bidding is Google's conversion-based family of it. CPA and ROAS are the two targets it can pursue, and the choice depends on whether conversions differ in value, as target CPA versus target ROAS explains. Keyword Match Types decide which searches enter the auctions it bids in. The Google Ads bidding hub explains the auction mechanics both sit on.

Common confusions

  • Automated as intelligent about the business. The system optimises the goal it is given; a wrong conversion action is pursued as faithfully as a right one.
  • A target as a limit. Targets are averages across auctions, so individual conversions can cost far more or less.
  • Smart Bidding as every automated strategy. Maximise clicks and target impression share are automated but not conversion-based.
  • Reported results as incremental. Conversions the strategy wins may have happened without the ad.
  • Learning as a fixed period. Large changes to targets, budgets or conversion actions send the strategy back into learning, and data volume decides how long it takes.

More terms to look up

Where this term is used

Sources

  1. About Smart Bidding — Google Ads Help

Related