Monday, 2nd September 2019 / Posted in Marketing

Google’s AI: Can it really outperform humans when it comes to PPC?.

AI is increasingly being used – and fine-tuned – to help search engines like Google to give online advertisers a more targeted approach to PPC. But are automated campaigns really better than those created by real humans? Our Digital Marketing Managers, Tom R and Tom W, take a closer look…


Q: How is Google using AI in 2019?

Tom R: Google has increased the use and effectiveness of AI to improve the performance of ad campaigns in particular. The introduction of features such as responsive search ads (which use AI to distinguish which ad variation would perform best in relation to what specific users are searching) is just one relatively new way that Google is harnessing AI to improve ad performance. The search engine is also using thousands of signals from users to determine when and where to show adverts to the most relevant audiences. This means that advertisers can now create – and run – smart campaigns that Google’s AI then decides how and where (such as through search, display, shopping or a Gmail advert) the user will be served the ad content for optimal results.
 

Q: How does this help target ads?

Tom R: Well, instead of gathering user behaviour, interests or location based on a specific device, Google can now follow someone based on the account they’re signed into across multiple devices. For example, if I’m logged into my Google account on my phone, laptop, desktop computer, Google Home or using Gmail or YouTube (the list goes on), Google will build up 100s or even 1000s of signals about my locations, what I’ve been searching for and browsing.
 

Q: How does this impact on PPC campaigns?

Tom R: The information that Google gathers across all your devices and platforms is used to determine whether or not you’re suitable to be served an ad in the display network – and if you’re ready to make an online purchase. This automated system gives online advertisers and digital marketers more information about targeted users, which, in turn, lets them create more relevant online adverts for their target audience.  
 

Q: What about AI’s impact on SEO?

Tom W: Google’s search algorithm used to be entirely coded by humans. They would come up with theories about how to improve the search results, run a test and, if the results were positive, implement the changes in the algorithm. The search engine now uses a machine-learning algorithm called RankBrain to improve the results they provide for each search query. RankBrain looks at what each user searches, and breaks it down to understand what the user is searching for. AI is then used to improve the algorithm to provide better results for the next user who searches that query. 
 

Q: How does AI affect the role of a digital marketer or advertiser?

Tom R: Google’s use of AI is pretty solid when it comes to PPC audience targeting. By using these automated bidding methods, digital marketers are able to make better use of their time. If we’ve tested the strategy and are happy that it’s the most appropriate method, then instead of manually increasing or decreasing the max cost per click of our bids, and measuring which keywords are performing, we’re able to dedicate more time the strategy of an ad campaign.

Tom W: When it comes to SEO, the most important factors were always backlinks and content. But now, Google’s RankBrain is increasingly determining if your site ends up in the SERPS. Ironically, although it’s a machine-learning algorithm, it actually increases the influence of human users on search results. That’s because, as torque.io explains, AI can use direct feedback from how users interact with your content to judge its quality. For that reason, you need to focus less on pleasing the machines and more on actually swaying people to click on your stuff. In that sense, our jobs have naturally focused more on user-pleasing content in order to keep rankings high.
 

Q: How does AI affect the role of a digital marketer or advertiser?

Tom R: Google’s use of AI is pretty solid when it comes to PPC audience targeting. By using these automated bidding methods, digital marketers are able to make better use of their time. If we’ve tested the strategy and are happy that it’s the most appropriate method, then instead of manually increasing or decreasing the max cost per click of our bids, and measuring which keywords are performing, we’re able to dedicate more time the strategy of an ad campaign.
 

Q: We’ve heard a lot about the pros, but what about the cons of using AI for PPC?

Tom R: The downside of relying on Google's AI – or any other platforms that use automated strategies, such as Facebook – is that you’re allowing them to control when and where your ad is served, and how much you pay per click. We have one client, for example, that we’ve implemented a number of manual bidding strategy campaigns to target our ads to the top of the search page – but not the very top. This is saving £1 to £3 per click but still delivering conversions at the same conversion rate. If we left it to AI, we may not be able to maintain this control – and ultimately saving per click.
 

Q: So should advertisers be relying solely on automated bidding strategies at this stage?

Tom R: Although Google has been pushing agencies to move towards more automated bidding strategies, as a Digital Marketing Manager, it is not as simple as changing the bidding strategy and sitting back to watch the results improve. We’ve always found that automated bidding will work better for certain campaigns, at certain times of the year in certain locations than others. But you have to be able to adapt, react and switch it up to generate the best results. There have been a few instances where Google has advised us to use automated bidding strategies, but Tom W and I were sceptical as it would have incurred a higher cost per click. Not surprisingly, our Google account strategist emailed a few weeks later to say that other agencies had reported that automated bidding strategies have not been performing well, therefore proving our scepticism towards changing every campaign to automated bidding.
 

Q: What’s the main takeaway for anyone looking to using automated bidding strategies?

Tom R: Although AI has hugely helped us identify more targeted audiences we can’t fully rely on its automated bidding strategies to get the best ROI yet. I’d always recommend testing different bidding strategies for different campaigns, regardless of automated bidding suggestions. When you find one that is working for your client, stick with it until you feel a change is needed.


To find out more about how our Digital Marketing experts Tom R and Tom W could optimise your next PPC or online marketing campaign, click here to drop us a line.
 
To find out more about Google’s use of AI and RankBrain, visit:
https://ai.google
https://torquemag.io/2018/07/google-rankbrain-seo/