AI in PPC - are we there yet?

Control - or the lack of it - is an important consideration to any digital marketer when approaching a campaign.

With the rise of AI (or perhaps ‘machine learning’ as it more correctly could be termed as) as an ever increasing feature of the digital landscape, its introduction into the very tools we use to market in this environment has been an inevitability. So is it the time to go ‘all in’ on AI in paid advertising or is there some way to go yet before we hand over the keys to our keen digital friend?

Is broad really better?

What inspired us to really sit down and get down in words what we’ve been discussing for some time around the water cooler was the revelation that ‘broad match’ keywords have become the default for new search campaigns across Google Ads - however, there are dozens of examples we could have picked as the spark to ignite a discussion. The launch of Performance Max campaigns in November 2021? Responsive Search Ads in July 2018? Looking a bit further back we’re actually not too far off the ten year anniversary of the introduction of Smart Bidding even.

A common thread which runs through each of the above is the lessening of the control afforded to campaign managers in favour of the platform itself running the ads. And if you were following the recent antitrust trial brought against Google over the last year or so you might be minded to think that was not such a fantastic thing.

It’s important to emphasise at this point this is definitely not just a ‘Google thing’, it’s a staple of the setup for all major advertising platforms. Only recently we were tasked with an audit of an existing Meta Ads account for a British-based holiday company which had an ad running with the music of Tarantella Napoletana (a classic Italian folk song) overlaying it. A curious choice indeed. The cause of which was the toggle switched on for the AI-driven ‘Music’ enhancements within the Meta dashboard, whereby Meta had deemed the song a good match ‘based on its content’. How, we’re not so sure.

Google knows best?

I think what makes digital marketers such as myself a little cautious when it comes to the rise of the default use of artificial intelligence across paid advertising is how much we’re relying on these particular platforms in understanding the nuances of your business.

We all know you’re the experts in your business but what we believe is crucial in providing CORE’s true value as your digital partner is that we immerse ourselves to the greatest extent possible in your world. Only by understanding your pain points, your USPs, your short term desires and your long term vision can we provide a service truly tailored to your success. Does Google or Meta or Microsoft understand your business like that? Could they ever? A debate for the long term but we believe a ‘no’ in the here and now.

The easy option not always the right option

What we certainly don’t want to pour scorn on is the definite advantages artificial intelligence can bring to the building, maintenance and promotion of your digital presence. We at CORE use it every day to assist us in our working practices and we are dedicated to remaining at the cutting edge of this digital frontier. However, the key word perhaps to dwell on is ‘assist’ - it is certainly not a replacement for all of our working practices.

It is this perception that using AI is the ‘easy’ option in creating and managing digital advertising campaigns that we perhaps baulk at. It is in the combination of using AI - to either provide data-leveraged suggestions (which we then have the final say in utilising) or to speed up our processes - with the knowledge of a business which can only be acquired through a holistic understanding is where the sweet spot lies.

Our top tips

With all of the above in mind we’ve pulled a short list of key points together to consider regarding use of AI in PPC campaigns you may wish to consult:

• No AI-driven campaign should just be ‘set and left’. Many ad providers will tell you they only need your website and budget to create a campaign which will deliver results for you - as they can generate ad text, target keywords and bidding strategy from those basic inputs. Regardless of this claim, we strongly recommend you check in regularly on any campaign and review the data available in order to steer it in a better direction if they have gone off course.
• Negative keywords are your best friend for broad match campaigns. By its very nature ‘broad match’ target keywords will put your ads out there across a wider set of users. This can indeed lead to your ads being seen by relevant users you would not have thought to have targeted as they were unknown to you previously but it also means they can be seen by completely irrelevant users. The only way to stay on top of reducing the latter is by regularly viewing the search terms report within your advertising platform and marking those irrelevant searches as negative keywords to prevent budget being wasted on them in the future.
• Avoid auto apply recommendations in Google Ads. A subject of a previous insights post, allowing Google to automatically apply recommendations to your campaigns is a potentially dangerous move. At a minimum it can lead to inefficient campaign running while at worst prevent your ads from running at all.
• Review the ‘Advantage+ creative enhancements’ in Meta Ads when creating an ad. Aside from the ‘creative’ music choices mentioned earlier in this article, Meta also suggests such features as juggling the order of image/video assets in ads depending on their popularity with users within its ‘creative enhancements’ section. Although this sounds logical, for many advertisers their ad creative has been structured to tell a story through their ads so carefully consider if that story continues to make sense with all the assets jumbled up.
• Ensure you have enough conversion data before accepting Google’s suggested ‘maximise conversions’ bidding strategy. Again, ‘maximise conversions’ sounds like a no brainer in how you want to bid for users. However, the truth is in order for Google to understand how to maximise bidding for conversions they need to understand what a converting user actually looks to you. If they have no (or little) conversion data to fall back on they are going to struggle to understand what these preferred users look like. We recommend campaigns should be at a conversion activity level whereby they are achieving 30 conversions in 30 days prior to switching to a ‘maximise conversions’ strategy.
  

Of course the above is just a flavour of how artificial intelligence is utilised across PPC. If you’d like to have a discussion about whether AI is currently helping or hindering your digital advertising and what can be done to weight it towards the former, we’d be delighted to carry on the conversation. Drop us a message or @ us on our socials.