How media planning will evolve in the age of AI
The big agency groups are cognizant that the long-established media planning tools are not fit for purpose, especially in an era increasingly defined by AI. Look to the multi-million investments made by GroupM (M-Platform), Denstu Aegis (Merkle and subsequently M1 platform) and Publicis ('Publicis People Cloud'). These are validations that the tools bestowed onto the talent in agencies needs to evolve. This however is just the tipping point.
The debate rages on about how people will need to co-exist with advancements in analytics or how people will soon be supplanted by fully autonomous AI applications. This for me is missing the point. In its current guise, advanced analytics requires smart creatives. Whether these are the planners of today or indeed the planners of tomorrow is a debate reserved for another time. What is key though is understanding and recognizing how people can support these advancements we refer to.
So how will tools continue to evolve?
We believe, simply in three key areas: data management; processing; and activation.
Solving some of the fundamental data management challenges will enable smart creative talent to leverage data, models, algos et al in more meaningful ways. These challenges are really around automating as much as possible the process of identifying, integrating, ingesting and storing (ETL) new data sets which will draw new insights which will be applied to solving brand challenges. This capability is inaccessible to most, better tools will overcome this we feel, thus bringing planners closer to the process.
To derive smart analytical outputs requires people who have an appreciation of business and consumer context. These people are the ones who should be influencing the specificity of how models are designed and deployed; namely feature selection and feature engineering. While an algorithm will help determine the features which are of most predictive importance, determining in the first place the features which should be brought to the table should be the hunting ground for planners in the future.
Better processing tools will support planners being more responsible for unlocking insights and indeed applying them. From simpler querying, to simple feature engineering to easy simulation of models – these worlds need not be confined to technical users. There is arguably more scalable value to be derived from these capability areas when this is more actionable by non-technical users. And as soon as this is enabled, these tools will likely be pointed at larger brand challenges versus the myopic nature of programmatic campaign execution.
Lastly, enhanced activation tools will also ensure the loop is being closed. Whilst traditionally the domain of buyers, we all must appreciate we’re in the results business. Having a better appreciation of how results are created will only bring more value to end clients and start to shape the planners of the future. If AI can be better leveraged for automated activation; segment creation; optimization etc etc then arguably planners have a much greater role to play in data driven campaign lifecycles; end to end versus purely at the inception.
AI has a huge opportunity to help make everything we do smarter, faster and more scalable. This will likely create more onus on business leaders to adopt it faster through easier to use tool sets. It will also challenge the notion and roles of our current talent and force them all to get closer to the end application of data, it will enhance the outputs of everything we do. And to bring back to the AI theme, those that resist will be exterminated.
Paul Silver was debating the above as part of a panel session at Advertising Week New York. Follow him on LinkedIn here