ChatGPT, Midjourney and other open-source AI programs have renewed peoples interest in the potential of AI. Without getting too in-detail about any philosophical, moral, or ethical side of the debates, there is a real excitement around how AI can save people time in answering questions or coming up with creative solutions.
This article explores the use of AI in media and advertising and what it means for marketing strategists. For the purposes of this article, I will be taking AI to mean the same as machine learning, namely the ability of a system to automatically learn and improve from its own experience and other data inputs.
But first, what is strategy about? If you ask anyone, they will tell you it’s all to do with insights (maybe not anyone, but specifically ask a strategist). To get insights though, it involves setting up research questions and deciding which insights are relevant. Additionally, reframing a story due to a client’s change in direction can be a real challenge. But the power of artificial intelligence might be able to help a strategist out.
Can AI understand human behaviour enough to tap into new audience pools?
Human behaviour is the foundation of insights for strategists. This is particularly relevant for marketing strategists in media and advertising, who seek to understand the motivations behind consumers in order to persuade them to purchase products or services.
Brands, particularly in performance media, often prioritize persuading consumers further along in the purchase journey to quickly drive sales, but this can cause audience pools to shrink and result in increased CPAs.
Therefore, the challenge for media strategists is to broaden audience pools to fuel the funnel while still targeting relevant users. However, this is easier said than done.
To grow the audience pools, the behaviour of previous converters can be looked at and mapped to future demand – many digital platforms partially do this to build look-a-likes, targeting users based on similar interests and engagements. AI can enhance this by quickly scraping the social profiles of previous convertors and looking at the time stamps to understand the actual sentiment when those users interact with the brand.
The process would reveal their true behaviour and feelings towards a brand, rather than infer them from page likes or video views. A machine learning algorithm can then map these outputs to identify new audiences with similar traits. This technique is particularly exciting for B2B businesses, helping them broaden their targeting and increase reach, in order to grow conversions, while still focusing on their key audience.
So yes, AI can understand human behaviour. But, only as long as the right actions and sentiments are inputted into it for it to analyse. On it’s own, as it stands, AI would not know what behavioural data to input.
Can AI combine differing data sets to help explain the ‘why’?
Behavioural data is useful for understanding what consumers do, but not why they do it. Market research, on the other hand, can provide insights into the motivation and emotion behind customer behaviour.
Combining these types of data sets can therefore lead to a more robust strategy, but it is challenging and time consuming to map different datasets together to still be scalable when they have different sourcing.
AI can help overcome this difficulty by digesting market research on “why” someone choose to convert as well as tracking behavioural data of those conversions to understand what they did, before joining the dots to provide an answer as to why someone did what they did.
The integration of offline insights and personalized online communication is difficult due to them being in different silos, but AI can break down these barriers with its powerful computing capabilities. By creating a new hybrid segment, leveraging both offline and online data, AI can take an abstract market overview and provide targeted and measurable solutions.
However, while it can use machine learning analytics to bring together behavioural data and market research, AI cannot go out on its own and generate the separate data sets and know it should combine the two together to create a better strategy.
How does AI affect the future of strategy?
At the start of this article, I was looking to see what brands, agencies or individuals had made early use of AI in a future thinking manner. In researching case studies, data sets and articles, two things became clear; firstly, it would be good if an AI could do this, and secondly, there’s not a lot of examples out there yet. But, I’d say, that is exactly the point!
AI has limitations in generating insights and making strategic decisions as it lacks human judgment and cannot determine which data points are most relevant to apply . Although it can be useful for tactical execution or sense-checking media plans, AI cannot fully replace a strategist in generating insights as human judgement is required to find the common thread in the data and tell a story that leads to an actionable conclusion.
While AI may improve and automate some aspects of the strategist’s role, it is unlikely to fully replace them, especially with concerns about data privacy and the depreciation of 3rd party cookies. A good strategist will be aware of this, but a great one should use AI alongside their own processes to enhance the insights they generate and the strategy they deploy – that’s the future.
N.B. all memes were created by me as well as a meme generating website, not an AI (yet…)