Google Pushes Machine Learning Into More Ad Tools

Google will offer more machine-learning products in the coming months to help advertisers create personalized search ads, bid on YouTube ads to achieve brand lift and boost shopping and local campaigns.

The company unveiled the new products Tuesday at its Google Marketing Live show, including what it said is the biggest change to how search ads are made since AdWords debuted 18 years ago.

Optimizing search creative

Rather than manually developing, testing and optimizing text ads for search, so-called Responsive search ads will use machine learning to immediately determine the best-performing combination among 15 headlines and four description lines supplied by the advertiser. It will continuously optimize based on what appears to be most relevant to the user.

If two people both searched for “best air purifier,” for example, they might see different creative based on signals used in bidding, such as device type. Google claims machine-learning creative optimization can increase clicks by 15%.

YouTube buys based on lift

Over the past few years, Google has been promoting YouTube’s ability to deliver brand lift and positioning it as a destination for premium brand buys, particularly following its significant brand-safety challenges last year.

Now the company’s Maximize lift product offers a bidding strategy to marketers who want to drive higher brand lift on YouTube instead of other KPIs, such as conversions or view-through rate. With an assist from machine learning, Maximize lift tweaks marketers’ video-ad bids in real time to improve lift in awareness, ad recall, consideration and favorability.

Driving foot traffic

Google also is helping marketers use machine learning to increase store visits. Google previously offered local ad formats, but now Local campaigns apply machine learning to a business’s location and creative to enhance ads to drive customers into stores.

And marketers using Google’s Smart Shopping campaigns, which was introduced earlier this year, will now be able to also use the tool to optimize for two additional business goals: store visits and new customers. Within a few weeks, marketers will be able to create and manage Smart Shopping campaigns directly from Shopify.

Responsive ads and Local campaigns will be available within the next several months. Maximize lift is in beta and will launch globally later this year.

Other product news from Tuesday:

Mobile landing page speed scores: Starting Tuesday, businesses using Google Ads can gauge how fast or slow their landing pages load on mobile devices. The scores will be use a 10-point scale based on several factors, such as the relationship between page speed and conversion rate. The data will be updated daily so businesses can track how landing page optimizations are working over time.

Cross-device reporting and remarketing in Google Analytics: The reports will combine data from people who visit a site multiple times using different devices so businesses can get a consolidated look at how users behave on their sites, regardless of the device they're using. Reports will only display aggregated data from users who have agreed to share it; individual user data will not be shared.

Google Measurement Partners: The new program has 23 new and existing partnerships across areas such as viewability, brand lift and brand safety. Partners include Integral Ad Science and Double Verify for viewability, verification and brand safety, and Nielsen and IRI for sales lift and marketing-mix modeling.

Hotel campaigns: The additional campaign type allows travel and hotel advertisers to use machine learning to maximize bookings.

2 Comments

  1. Retail Agency

    I wonder how these machine learning products will help small to medium sized agencies serving other small to medium sized businesses.

    Or, do agencies just get cut out of the loop? Antitrust issues if agencies go away?

    Reply
  2. Informative post, Tilde. Machine learning can be great for the advertising industry. There’s so much data available and this can be used to bridge the gap between a customer's need and a brand's positioning or business goals. Since this can get overwhelming at times, it’s important to have a strategic plan for achieving the primary goals when using machine learning.

    Reply

Add a comment

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>