Stay marketing and tech-savvy. Get the latest in martech - subscribe to MarTech Today.
The value of applying artificial intelligence in display advertising
From increased efficiency in campaign management to stronger personalization, columnist Adam Grow explains how AI is changing the face of digital advertising.
Technology is advancing at an accelerated rate, and the far-fetched dreams of yesterday are quickly becoming the innovations of today. For the longest time, artificial intelligence was an ambiguous notion of big ideas with futuristic applications — self-driving cars, drone deliveries, fridges that monitor food quality, and so much more.
This rapid pace of innovation has some theorists concerned with “singularity” — the belief that artificial super-intelligence will cause such rapid growth that human civilization will experience incomprehensible change. Some theorists predict that, at its current rate, singularity could come about in the next 30 years.
My point here isn’t to focus on if or when singularity will happen, but rather to convey that we are currently experiencing the impact of technological progress in our day-to-day lives and in the world of digital advertising. These advancements affect the way we interact with brands and the experiences that lead us to purchase.
And while many of the big ideas for artificial intelligence are still in their infant stages, the reality is that AI has many seemingly small applications that currently deliver digital marketing efficiencies for companies all over the world — from advanced consumer targeting and insights to highly personalized ad experiences.
AI’s place in advertising
Nearly 80 percent of US senior marketers believe consumers are ready for AI, and consumers already see the benefits of a world full of it. In a survey from Weber Shandwick and KRC Research, respondents worldwide believe that AI technology can help solve a lot problems: 72 percent stated a benefit of AI is the ability to complete tasks that are too hard or dangerous for people, 69 percent said AI provides easier access to relevant news and information, and 69 percent said it frees up time to pursue other activities.
While these points may align with a consumer’s everyday life, they are directly applicable to the value we see from using AI in advertising — specifically in programmatic display. AI allows marketers to more efficiently drive performance through relevant and personalized advertisements that leverage enhanced targeting algorithms and creative optimization.
AI provides efficiency in campaign management, but human touch remains imperative
AI is being adopted by many ad tech companies to improve the efficiency and relevancy of display campaigns, resulting in better performance, which is the ultimate goal of every marketer. Machine-learning algorithms used to drive efficiency across real-time bidding networks will generate $42 billion in annual ad spend by 2021, up from $3.5 billion in 2016, according to Juniper Research.
Though the value that AI brings to programmatic display is undeniable, we can’t forget that our human workforce is still a vital component of successful media buying.
Machine-learning algorithms do the heavy lifting of programmatic media buying, intelligently identifying consumers and serving relevant ads, but there are still pieces that can only be recognized and managed by a human. These include accounting for sales and promotions, the integrity of brand assets and anomalies that arise and can be accounted for by simply thinking.
I’ve seen the greatest results at our company when we combine the advanced targeting of machine-learning algorithms with the thoughtfulness and big-picture insights from our own media and campaign experts.
AI is a conduit for relevant and personalized experiences
Personalization has been a buzzword and a big focus in display advertising for some time, yet consumers are wary of their information being used for advertising purposes and the resulting influence it may have on their purchase decisions.
A survey in a recent eMarketer report found that only 28.7 percent of consumers felt that a personalized experience increased the likelihood that they would make a purchase, and more than half said it made no difference. Yet in a separate survey, 78 percent of US internet users said personally relevant content from brands increases their purchase intent.
So, what is our takeaway from these conflicting responses? Ultimately, it comes down to the quality and relevancy of the personalized message, whether consumers realize it or not.
Personalization technology has evolved to become a standard in advertising for many brands, but the technology must continue to evolve for brands to deliver powerful ad experiences among changing consumer behaviors, platforms and media.
AI provides a solution that is more than simply displaying products a consumer has viewed. Machine-learning technology considers other top-performing products to provide a variety of product recommendations that will ultimately drive a consumer to complete a purchase.
Going beyond products, this personalization includes the messaging or offer within the ad, and even the placements and timing of the ads. For new customer acquisition, algorithms can take top-performing products from retargeting campaigns and find similar products in a feed to serve to high-value lookalike audiences. Even consumers who have never been to your site can have a relevant ad experience.
Performance is delivered through machine learning and dynamic creative
Retargeting campaign performance used to just be a matter of having the right creative. If a campaign wasn’t performing well, it was probably because the creative needed to be refreshed.
To avoid creative fatigue and increase performance, marketers have historically turned to A/B testing. The standard protocol involves running two or three weighted ad variations through a server for a test period until near-equal impression volumes are reached, then the results are evaluated by an analyst before the winner is identified and then tested against new variations.
The process itself is extremely manual, not to mention lengthy and time-intensive, just to reach a solid conclusion across a few ad variations. Unfortunately, it is also difficult to make optimization adjustments or changes mid-test for fear of compromising the test and results.
Relying solely on the human element for this kind of testing process makes it nearly impossible to test or analyze the performance of multiple factors at once. Sad, but true — we simply aren’t intelligent enough to understand or predict human behavior across multivariate ad testing.
Dynamic Creative Optimization, also known as DCO, is an important application of AI in display advertising.
In the ever-cluttered world of online advertising, marketers must fight for the attention of their target audience. They need the right image, the right colors, the right products, the right call to action — all at the drop of a hat — for the hundreds of thousands of consumers that have visited their site.
Using behavior and purchase data, machine-learning algorithms create a wide variety of ads using simple templates and test them among consumers to find the most effective ads that lead to conversion, and DCO can do it all without any human bias mixed in.
It’s remarkable to think about the progress and technological evolution we’ve experienced over the last 100 years. Referring back to my earlier comments on singularity, for those not as familiar, I highly suggest looking up a chart for Moore’s Law. When visually represented, it’s staggering to see the level of technology advancements that have occurred in the last few decades.
As this growth continues, it will be interesting to see how marketing technology evolves to mirror and support the more consumer-centric advancements. It will be our job as marketers to push our industry forward in an effort to identify and utilize opportunities that further enhance our ability to make meaningful connections with consumers.
Opinions expressed in this article are those of the guest author and not necessarily MarTech Today. Staff authors are listed here.