What is automated
In the case of Smart Bidding strategies like Target CPA, Target ROAS and Enhanced CPC, Google automatically predicts the likelihood of conversions by looking at auction-time signals including device, location, language, dayparts and more. These predictions feed the automated bids that are used for every unique auction.
What still needs to be done manually
While Google can predict changes in conversion rate and conversion value based on a variety of factors that are widely applicable across a range of advertisers, these systems don’t yet consider unique factors that impact individual advertisers. This means that advertisers should supplement “automated” bid strategies with a management methodology that changes targets based on business-specific conversion factors.
Things like flash sales, coverage in the media, weather, social media buzz and so on can all impact how an ad campaign converts, but these factors may not be apparent to Google’s machine learning, so the advertiser who is aware of these factors must do active bid management. But instead of managing things by changing a max CPC bid, management now entails changing the target.