It’s not often issues such as ‘programmatic advertising’ and ‘brand safety’ make the BBC national evening news, however that’s exactly what happened last month thanks to the ongoing furore surrounding brand ads appearing alongside inappropriate online content.
The recent crop of big name brands withdrawing ad budget from Google/YouTube follows on from the controversial ANA report in the US last year, which flagged up the very poor levels of transparency in the programmatic advertising industry and the very high levels of mark-up being applied by media buyers.
When Procter & Gamble’s (P&G) chief brand officer Marc Pritchard recently made a speech characterising the current levels of transparency in the advertising supply chain as “murky at best and fraudulent at worst”, he sounded last orders on the outdated practices of many media buyers.
In his address to the IAB’s Annual Leadership Meeting in January, Pritchard told his audience that P&G had “come to its senses” and would no longer continue to invest in a media supply chain that was complicated, non-transparent, inefficient and, potentially, fraudulent. The global superbrand has given its agencies and media suppliers a year’s notice to clean up their acts – or else.
For many, then, the game is up. Fortunately, for those intent on delivering the value, transparency and brand safety that advertisers such as P&G will increasingly demand, there is a ready-made answer in the form of Machine Learning.
It’s ironic that while much of the media coverage of Machine Learning to date has positioned the technology as an existential threat to media buying firms and the jobs of those employed by them, in light of Marc Pritchard’s comments, Machine Learning may be the one thing that can actually save the beleaguered industry and restore its reputation as a valued partner to the business sector.
In fact, I would argue that the capabilities of Machine Learning point the way forward for media buyers on how to deploy their human resources to best effect. Let me explain why.
Rise of the machine – and the human
The use of programmatic technologies within the delivery of ad campaigns over the past few years, has risen drastically. With this use of automation expected to increase within the next year or so, we can anticipate that it won’t be long before it’s influence crosses paths with the planning of campaigns, whereby Machine Learning will be able to displace a considerable amount of the manual time and effort involved.
The reason? No human planning team of any feasible size can ever hope to replicate the capabilities of machine learning, which can quantitatively evaluate millions of individual customer journeys every hour, producing insights into consumer behaviour at a level and depth that’s never previously been attainable. The insights provided allow the machine to automatically identify and target audiences with relevant ads or messaging and deliver these at a frequency tailored to every stage of the consumer’s individual journey. If media buyers are struggling to resource and execute on planning or delivery, therefore, machine learning can help.
However, for the planning and delivery elements to work successfully, it is vital that they are underpinned by a solid foundation of strong supply chain management. We’ve argued before that trusting machines for planning and delivery of online ad campaigns can free up human resources to dedicate to customer experience and data strategy, but, equally, given the current industry emphasis on brand safety, machine learning can also be applied to support the processes and tasks needed to run in brand safe environments.
Machine learning certainly has a role to play with brand safety, but only providing you have set up a safe environment for the machine to work in. This, in practice, requires manpower. The current generation of media agency managers must decide for themselves where machine learning can be deployed successfully and which elements of service delivery require the human touch and the comprehension and intuition that only human experience can provide.
For example, personalisation of marketing is one area where machine learning can help deliver a finely tuned user in real time. Leveraging multiple data points to iterate creative or landing pages is an effective way to ensure the most relevant and helpful path to purchase.
However, other elements of marketing that should not be automated are the brand story and promise, the vehicle that brand message is delivered in, and the context in which that brand message appears. We know that the art of persuasion is inherently human, and no machine can replicate the emotional level in which we make purchase decisions – that ‘gut feeling’.
The creative or delivery mechanism in which we bring our message to the user must respect their privacy and their sense of security in a non-intrusive way, and the message must be delivered in a context that makes sense so that all elements a brand’s consumer outreach are in perfect accord.
The best way to ensure this is through having a human oversee the emotive aspects of marketing, its creative design, and, crucially, its ultimate placement on media channels that complement the original message. We can rely on machines to help us gather data and make rational decisions, but true persuasive balance is struck when we inject those elements that make us human into the mix and come up with a message that it at once both practical and provocative.
If you are a client of a media agency, like Marc Pritchard of P&G, it’s time to challenge your agencies and vendors to embrace the benefits of machine learning and redeploy human resources to help deliver the brand safety, transparency and value for money you need.