Why machine learning in advertising still needs real life human input

Jamie Swain

In this article written exclusively for Business Leader, Jamie Swain, Head of PPC at Pixel Kicks, discusses why machine learning in advertising still needs real-life human input.

Google Ads has become an intrinsic part of most marketers’ strategies. Traditionally, it has been a manual process that has been driven by research and insight. However, things are changing as Google is now pushing machine learning more than ever. Recent regular updates to the platform suggest a less hands-on approach for marketers will continue to become the norm.

When setting up new Display or Shopping advertising, Google now selects ‘smart campaigns’ as the default option. A smart campaign does a lot of the advertiser’s work for them, using AI to choose how the ads are targeted and displayed. If you’re optimising a search campaign, you’ll often find the dynamic ads format being pushed towards you by Google, which sources and creates the ad copy without human input.

It’s fair to say that reception has been mixed and it has been noted widely in the PPC community that Google seems to be aggressively pushing a more automated approach. They’ve removed some features that provided more data and transparency, which have historically enabled marketers to be more hands-on. This development is a concern for some, but the key will be to embrace it rather than fight it as users will be pushed to give Google more control over their campaigns.

It’s important to note that automation can be a good thing in some areas of campaign management. Automated bidding strategies can give users a higher conversion rate, and some areas of Google’s automated features also work well as part of a PPC campaign. You can set up automated bidding strategies based on your marketing goals meaning Google’s machine learning will automatically gear the campaign towards getting the most results for your budget.

While it sounds like this could be a positive move, putting too much power into Google’s hands can be detrimental to your advertising goals and budgets. In some cases, a smart campaign can run smoothly, with Google’s algorithms being as accurate and reliable as you need. In many other cases though, it’s not something to rely on too heavily.

Google’s AI won’t have a full, humanlike understanding of every business type and its goals. A client from a niche industry will have a unique insight that’s difficult to replicate through algorithms. As such, automated campaigns may strive to increase results for a specific product or service, while the client may have more specific requirements not being met.

It’s not just the campaign performance that deteriorates though – you could also see your advertising budget take a hit with automation in place. Looking at Search campaigns specifically, some of the AI features available can put your ads in front of users who are searching for terms that Google deems valuable to your business. All of a sudden, you could be paying to have your ads show on search results pages for keywords not at all related to your products & services. Users will click, costing you money, and then immediately exit when they see you’re not what they’re looking for.

These shortfalls mean you should consider working with an ads specialist. Even Google themselves say that marketers are crucial to the success of automated campaigns. A McKinsey & Company report published in 2018 states that the most difficult activities to automate are those related to decision making and creative work, two key areas of any Google Ads campaign.

While Google Ads tries to encourage users to utilise as many of these automation features as possible, it’s better to find a good balance. It’s tempting to avoid paying a marketer to manage your campaign when Google can fully run it for you. However, a marketer can use their knowledge and insight to create an efficient marketing strategy, and while they may still use some of the automated tools, their manual input can provide you with better results. Google Ads’ algorithms can ensure your campaigns are using effective bidding strategies to get you a strong return, while human input provides you with better quality ad copy, research, and planning to meet your requirements.