AI Will Disrupt Advertising Media, But Not the Way You Think
Chris Peterson, Managing Partner
Mention artificial intelligence (AI) to someone in advertising and the conversation usually turns to chatbots like Siri or Alexa. The idea is that these virtual assistants will collect audio data to enable better ad targeting. So a married couple arguing at home within earshot of an Echo Dot will start seeing divorce attorney messages on their mobile phones – or something like that.
AI will certainly disrupt advertising, but it has almost nothing to do with chatbots. Where Siri or Alexa’s natural language processing enables human-machine interactions, AI in advertising media management will enable interactions between a machine and a marketplace.
Disruption from AI will force media agencies to become technology companies while turning brand marketing departments into media control centers. Along the way, brand marketers will make smarter, higher return investments across their entire media investment portfolio while media agencies will discover a new path to profitable scale.
This isn’t some wild-eyed vision of the future because it’s already unfolding. In fact, we’re probably more than half way to the point of significant disruption. The next three to five years are the home stretch for realizing true transformation when media agencies and brands will find themselves on one side of the disruption or the other.
To understand where this is headed, first consider a simple definition of AI from Merriam Webster: “AI is a branch of computer science dealing with the simulation of intelligent behavior in computers.”
To “simulate intelligent behavior,” you first need sensors that deliver robust data that can be used to make decisions. A self-driving car, robotic surgeon, or delivery drone can’t operate without a robust set of environmental data from which to simulate behavior. To borrow a phrase from military and emergency response, sensors deliver data that combine to provide very strong “situation awareness.”
Sensors in advertising tend to come in eight categories: target audience, media delivery, business results, brand health, competitive activity, economic indicators, marketplace conditions, and seasonal influences. The ability to tap into data from sensors across all of these categories has dramatically improved over the past several years. Sensors in several categories are now providing near real-time data that form a strong foundation to enable AI-driven media management.
If you think back just two years, it’s easy to recognize the incredible strides made in the formation of situation awareness. Signals, such as views, completions, clicks, and traffic from such digital devices as mobile phones, notebook computers, and TV’s are now connected in device graphs that map a consumer’s interactions across devices.
Both online and offline signals can be connected in near real time to understand how the market is interacting with a brand’s messages. While signals mostly ignore the effects of one media channel on another, regression modeling delivers a powerful, cross-channel view that course corrects shorter term decision making driven by signal information.
The next step–simulating human behavior based on strong situation awareness–is far less daunting than it seems. You can attack it in incremental steps. From situation awareness you proceed with the AI concept of “multi-modal data fusion” to enable decision making.
Multi-modal data fusion for human-machine interfaces can include processing such inputs as gestures, voice, and keystrokes. In media management, multi-modal data fusion requires processing signals and modeling across all media channels while taking into account marketplace, seasonal, economic, and competitive dynamics.
To gain insight into how incremental steps take place, think about all the safety tech in today’s cars that take situation awareness and simulate behavior. Sensors pick up close objects and start to apply the brakes. Adaptive cruise control slows the car down when approaching a slower vehicle ahead. You can benefit from lower-level forms of AI without having a completely self-driving car.
Likewise, you can take small steps in behavior simulation to have significant impact on media efficiency and efficacy. Over time, the simulation of human behavior can increase as data is better fused and decision making becomes more accurate. Unlike self-driving cars, however, the need to completely remove the “driver” is not essential or even desirable to achieve media management benefits.
If you want to see what the future of advertising media management looks like, consider passenger ferries in Finland. On December 3rd, 2018, Rolls Royce Industries (not the car company) and Fin Ferries (the national public passenger ferry system in Finland) launched the world’s first fully autonomous ferry, traveling from Parainen to Nauvo, Finland.
The ferry system employs sensors for situation awareness, simulates human behavior by fusing the sensor data to navigate the ferry, and fully connects the AI platform to the people who manage it. Situation awareness comes from electronic charts, radar, sonar, weather forecasts, and other inputs to identify anything that can affect the course of the ferry.
Most importantly, the ferry is fully connected to a control room hundreds of miles away where operators can intervene at any moment and take over complete control of the ferry. Fail safe systems provide for connection failure.
In this example, Rolls Royce Industries is the media agency that provides the AI platform, Fin Ferries is the brand marketer, and the ferries are the media. Rolls Royce developed a platform that takes the situation awareness data provided by commercial devices, fuses the data from different modes, and applies AI to manage the boat. The control center – the human interface to the platform – can be located and managed by Rolls Royce Industries or Fin Ferries.
What does this all mean for media agencies, ad tech, and brand marketers? For a media agency, it requires no less than transforming into a technology company to provide proprietary AI-driven media management platforms. If they don’t, then media management AI platforms of the near future will relegate them to smaller service organizations with very little differentiation and low margins. They will provide ancillary services to a core technology platform in a similar fashion to firms that help manage Salesforce CRM systems or Oracle supply chain solutions.
For brand marketers, AI applied to media management will greatly improve the efficacy of their media investment as the AI platforms get stronger and stronger. Brands will either rely on the newly transformed media agencies to manage their AI platform of choice or will staff to manage it themselves.
Decisions around which AI platform to choose will be made in a similar fashion to any other cloud-based enterprise software solution. The conversation will shift from “who is my team?” to “what platform do you have and who is managing it?” Media agencies with AI platforms may even give them names, such as Alexa or Siri, and will focus on how their AI implementation is unique, better, and proven.
For media agencies, this moment in time represents unprecedented business opportunity. Over the past twenty years, competition and pressure on compensation have only increased, no matter the scale. The linear relationship between revenue and the need for people has been inescapable. By transforming into an AI-driven technology platform, media agencies effectively unlock a far more scalable and profitable business model.
Media agencies are also in far better position to develop these platforms than standalone ad tech companies. Fusing situation awareness data into simulated human behavior requires a deep understanding of how media decisions are made across a wide breadth of brand business situations. Building a generic model and simply letting it learn from its mistakes through machine learning is never going to happen.
Media agencies have instant experience across a wide breadth of client engagements with different business objectives, strategies, and outcomes. This experience reflects the complex differences in building awareness, driving retail traffic, generating web site traffic, or even turning a business around. There’s no better place to start with AI-driven media management than inside a media services organization. The challenge is transforming from a services business with technology subscriptions to a technology company with additional services. This is certainly no small task, but as the saying goes, “if it were easy, someone would have already done it.”
Today’s ad tech companies–while having the staff, skills, and culture to develop media technology platforms–currently operate in distinct silos. They attack smaller problems in the media landscape, limiting their ability to solve for AI in media management and analytics. These firms will evolve into sensor organizations that will provide inputs into a powerful AI platform. If ad tech companies attempt to tackle true AI-driven media management, their progress will also be slow because they cannot scale their experience quickly.
AI will certainly change how people work and what roles are required, but it will always put a premium on media knowledge that goes beyond what a platform can know in the moment. The highest order value that media managers provide today remains perfectly relevant in an AI-powered future. During the near-term period of solving for AI in media management, an even greater premium will be placed on media strategy, marketplace understanding, and analytics.
Late in 2018, Pew Research conducted a survey of more than a thousand professionals involved in the development of AI. A full 63% indicated that AI will result in a world that is better for people rather than creating some dystopian vision of robots taking over the world. If media agencies understand the future now, they can prepare for it in a manner that doesn’t just sustain business, but actually accelerate it.
If you don’t believe this is going to happen, then look across industries today where AI development is focused – transportation, healthcare, and manufacturing. AI investment concentrates where there is the most at stake. With some $220 billion spent each year on advertising just in the U.S., there is a lot at stake.