.png)
2023 has been a big year for AI, and it can feel as though everyone is doing it but nobody is giving specific examples of exactly what it is that they’re using it for.
Some might be surprised to know that AI has actually been around for a long time - the term first came into usage in the 1950s and that’s when the LISP programming language was created, which is still used for AI research to this day.
At its basic level, Artificial Intelligence is simply the ability of a machine to emulate human thought and behaviour. Machine learning is a subset of AI which uses models and algorithms to learn and improve.
Recently, we’ve seen big strides forward in AI, including;
Genie Goals moved to Google’s machine learning-driven bidding tools (Smart Bidding) in 2019 when this started to outperform other bidding tools in our regular testing. Smart bidding allowed us to bid based on dozens of signals in real time - while our keyword bids were previously automated using our in-house technology, this was a big change to how we adjusted for context attributes like device, demographics, etc.
Google Ads has continued to develop and evolve the use of machine learning to optimise performance and this has been taken beyond bidding to almost every aspect of Google Ads, including creating copy and creative, keyword matching, automated account management, reporting and more. Some of these features are fantastic, others have huge downsides.
Google PMax campaigns mix all of these automations together in a campaign type where you input goals and creative, and machine learning attempts to achieve those objectives within the constraints that you’ve set. Meta has launched a similar offering with Advantage+, a near campaign type for Instagram and Facebook Ads.
Artificial intelligence is only as good as the inputs it is given, and setting up automated campaign types without care and thought can leave you with a campaign type which is extremely good at spending the budget you set, but very poor at giving insights. The insights you lose might be ones which drove valuable business decisions. You can lose visibility and control of what your ads look like, where they’re appearing, who is seeing them, your top search terms, and competitive insights that you previously took for granted.
A lot of thought needs to go into making sure that strategic business priorities are taken into account during setup, initial launch and optimisation. These challenges can be amplified for businesses with strict brand guidelines or a large and varied product range.
While we have found that technological advancements have really helped to drive efficiencies in our marketing spend, a lot of our work now goes into making sure that we’re optimising the inputs and guiding the system to make the right decisions.
There has been a lot of talk about new technologies emerging which can do things like;
Many of these are in their infancy, and a lot of fun to test and play with to see what they can do. It is possible to get results from these tools that are usable, but for the time being they work best with simple, repeatable tasks. For the time being, we wouldn’t recommend entrusting your media planning to ChatGPT. An important consideration is to ensure that any data you enter into systems like these is done in a data and privacy-compliant way, with appropriate consent from the owner of the data. A good option is to test out tools using dummy data.
AI and ML tools from Google and Meta are driving extremely positive results for our clients and there is a definite competitive advantage to making sure you’re using these in the right way. These are clever tools which need experienced professionals to get the best out of them, in terms of driving the performance you want, aligning with your branding guidelines, and getting the insights you need to make decisions about your business and marketing.
There are loads of other exciting new ways that are emerging in which AI can support you in your work. As with any emerging technology, there are risks to being an early adopter. We recommend taking a risk-averse approach to your company’s data, but otherwise to try things out, get good at using them and have fun.
