Ghost in the ad machine
ZenithOptimedia’s Grant Millar talks stack-storytelling and AI
WITH the rise of the algorithm in the advertising industry, it is becoming easier to tailor messaging around a consumer’s individual preferences. But as traditional segmentations are consigned to history and analytics continues its inexorable march, advertisers face pressure to defend their methods from criticism around privacy.
Grant Millar moved from Dentsu Aegis to become UK chief executive of ZenithOptimedia in January 2015, where he is responsible for its digital performance marketing, branded content and analytics sections, working for clients like RBS, Costa Coffee and Harvey Nichols. He tells City A.M. why targeted advertising is less intrusive and why tech will never replace common sense.
You call yourselves “the ROI agency”, so are there any sure-fire ways to convert interactions into purchases?
You need the best people. Our team comprises people from a broad range of specialisms including more than 40 data analysts, PhD grade machine-learning experts with econometrics capabilities. They can unravel all the information about a campaign to assess the individual impact of each strategy to discern a ROI for a client’s ad spend for TV, radio and so on.
But gauging ROI on social platforms, for example, depends on the client and how closely their activity on social media is connected to the point of purchase, or whichever performance indicator (KPI) they choose. Take car manufacturing, for example. There is a huge journey from a person’s engagement with a car campaign on social media to the purchase of a vehicle. There are so many variables for our analysts to process that a full-blown econometrics analysis would be needed. But if their KPI is a test-drive, that is a journey we can measure much more closely.
Clients are no longer asking us to come up with a spreadsheet of TV, radio and other paid media placements, but create a “brand experience map” which depicts the consumer journey, with all the owned and paid media activities around the journey, showing how those activities work in concert to deliver the right messaging as the consumer comes closer to the point of purchase.
What is stack-storytelling?
This is a very hot topic at the moment. With so many people second-screening, it makes sense for media companies to leverage data from those individual connected platforms so consumers receive a consistent and targeted set of messages across all their devices. Stack-storytelling is generally done through platforms where you log-in like YouTube, Google and Google+, using platforms like Google DoubleClick.
You might start the day on Gmail at home, where you are shown a video from Toyota. If you engage with it, you might be shown the next piece of that campaign on YouTube at work.
Does large-scale data gathering risk making advertising a bit too personal?
Understandably, the public has some concerns. But quite apart from the ethical implications of brand “spying”, we want to reach people when they are most likely to benefit from the message we’re putting in front of them.
We did a campaign for Costa Coffee which used first party data from customers’ loyalty cards to give them messaging that was tailored to their previous buying preferences. We could tell cappuccino drinkers that a new roast is available, and serve them the relevant ads. You can argue that messaging like that is less invasive than blindly sending messages to potential consumers that might not be so relevant.
Tech has swept the industry so quickly. Will humans be managed out?
Programmatic and analytics are opening doors, but tech needs to be overseen by a human eye.
With regards to the viewability of an ad, we want a new currency for measuring whether a communication has actually been engaged by a human being. “Cost per human” should replace the existing metrics for charging clients for digital views. I also think it’s important that a human eye is cast over any communication produced by tools and systems before it is targeted at a consumer.
I don’t think the point will come when AI is capable of human logic, understanding or empathy. Indeed, Demis Hassabi of Google’s DeepMind unit has said that such developments are a long, long way off. A degree of common sense will always be critical to the services we provide for our clients.