Ok, so we’re not talking about a robot uprising where you surrender your Google Ads account to the machines. We’re talking about Machine Learning and how you can use Google’s data to improve the performance of your Google Shopping Ads and Google Search Ads.
What is Google’s Machine Learning?
Machine Learning (ML) is the use of data and algorithms to improve upon the effectiveness of a task. And Google is a “ML first” company. Google, through their AI, use ML to determine which ad they show, to who and when. Where it has us humans beat is that it can analyse 70 million signals in 100 milliseconds. A little more data than you or I could process…
Google Ad types use ML in various ways to help improve the results from your advertising.
Why should you use it?
When it comes to any kind of decision making, the more data you have, the better your outcome will be. As much as marketers and agencies would like to say they know who best to target, they ultimately don’t have as much data as Google does. There are plenty of opportunities to utilise Google’s data in your ad structure and bidding strategy to help you beat the competition.
Where can you use ML?
You can leverage Google’s ML to get more efficient ad copy through the use of Responsive Search Ads.
To create a Responsive Search Ad you need to provide Google with multiple headlines and descriptions. This is so that Machine Learning can work out the most effective copy. The most successful combination at driving the relevant conversions will be more likely to show.
Better ad copy results in a better expected CTR% which means lower costs per click and better returns from your ad spend.
You can select a number of automated bidding strategies, depending on your goal, to leverage Google’s data.
- Target CPA: Google will try to achieve as many conversions at a specified cost per conversion and budget
- Target ROAS: this will try and achieve the most revenue above a particular return on ad spend. Selecting this strategy may mean that your budget is not fully spent
- Maximise Conversions: this will try and achieve the most conversions for your budget regardless of conversion value
- Maximise Conversion Value: this strategy will try to achieve as much revenue as possible for your particular budget. (You can apply a target ROAS but this may affect how your budget is spent)
- Target Impression Share: Google’s ML will try to optimise your budget to show your ads a specified % of the time
- Viewable CPM: for display and YouTube campaigns, this strategy will spend your budget at a specified cost per thousand impressions
No human would be able to optimise towards these goals with the same data.
Google Smart Shopping campaigns take into account various data points to set bids and attribute budget to ad formats and audiences. Smart Shopping campaigns use the Machine Learning bidding strategy ‘Maximise Conversion Value’. Basically, the changes to the bids and audiences are made to what Google’s Machine Learning calculates will give you the maximum revenue. Google uses data from the following different areas to improve ad performance.
- Location: Google’s Machine Learning looks at what states, cities or suburbs particular products perform best in and optimises towards those.
- Placements: Your Google Smart Shopping campaigns will consider the website placements that perform best. Google Smart Shopping display ads are placed to maximise your conversion value. Whether it be any number of sites on the Google Display Network, YouTube or Gmail.
- Time of day: If your ads are driving more revenue at a particular time or day, your Smart Shopping campaigns will adjust your bids. Your ads will then primarily show at those times. Similarly, your budget will be spread among your product ads, your display remarketing and your display prospecting ads, in real-time, depending on performance.
- Devices: If you have products that perform better on mobile or desktop, your bids will be automatically adjusted to ensure your ads are shown on the correct device.
- Products: Your Smart Shopping campaigns will also consider what products are converting and what products are more likely to bring you greater revenue. Machine Learning will adjust your bids for higher performing products to show those ads more often.
With standard Google Shopping campaigns all of these data points can be adjusted manually. Smart Shopping uses a combination of your historical retailer insights and Google’s user insights through Machine Learning to optimise for you.
Google’s Machine Learning will help you generate the best ads for delivering on your goal in your Google Display campaigns. You can use ad formats such as Responsive Display Ads which, much like Responsive Search Ads, will let Google pick the best performing copy and assets to drive conversions based upon your bid strategy and goal. Machine Learning will select the images, videos and best performing copy to drive more conversions.
Machine Learning acts in the same way in other Display Ad formats such as Discovery Ads.
How does RQMedia use Google’s ML?
Our automated Search Campaigns are built from an integration between our Ad Platform’s software, inventory based websites and Google Ads. We generate ads for every product, service or travel asset at the category and product level. We then add together product or service details with the company’s brand messaging. Depending on inventory size, we can create hundreds, thousands or even millions of ads through our platform; something that no human would be able to do.
We give “the machine“ all the inputs and ad set choices it requires to serve the most appropriate and “likely to convert ad” to a potential buyer actively searching for the exact product or service. This is the perfect match of Google’s big data, buyer behaviour/customer needs and the Dynamic Creative Ad Tech. As a software company first and foremost, we have to accept that sometimes, machines can do things better than us.L