Advertising,

AI in Advertising: What it Means for Marketers

Written by Tim Edmundson

Oct 29, 2018 10 Min Read

There’s a lot of buzz around AI and advertising these days. You would be hard pressed to visit any industry news or ad tech website and NOT see something related to AI. They all know it’s the next big thing, and they want you to know it too. We’re also excited about its possibilities; we’ve got an AI-powered, machine-learning optimization engine we’ve affectionately dubbed Dynamic Spend Optimization, or DSO for short, that’s driving awesome performance for brands.

There isn’t an industry around that won’t be changed by AI systems. Major companies like Google and Microsoft have invested heavily into the technology’s potential because they see a future where AI can influence the trajectory of their businesses. There are countless industries that will feel AI’s impact, whether it be self-driving cars, financial services, or even fast food.

But with so many applications in so many industries, the meaning can get muddled. It’s safe to say AI has become a buzzword. We think AI has a specific application in advertising though, and while that’s not to say its role won’t continue to evolve as more tech emerges, we’re confident AI has a very strong role in advertising now.

 

AI Helps Marketers be Efficient

There’s no denying that today’s marketers can be data junkies. Before the days of the internet, campaign data was sourced through surveys and best guesses. With the advent of digital advertising though, thousands of metrics have been developed and available for analysis. Which is great — more data the better — but advertising went from not having enough information on performance, to way too much.

It’s not humanly possible to sift through the countless number of data points that make up digital campaign performance, especially in a quick enough time to take advantage of the opportunities they present. This is where AI can have its biggest impact for brands. Within that data are stories that lead to better performance. If there’s a trend where users who eventually convert are served an in-view ad on a certain publisher, on a certain day of the week, it goes without saying that serving ads during that time and on that publisher would be extremely valuable to a brand. This isn’t something a human will catch quickly, or they may even miss it, but AI can shoulder that burden and identify the behavior quickly. That’s essentially what Dynamic Spend Optimization does; it sorts through billions of data points to identify the trends that lead to conversions. It does what humans can’t possibly do, and it does it well, which is the perfect example of what AI brings to the table for marketers. This creates efficiency in a system that has traditionally had trouble reacting quickly to new information.

This line of thinking has been proven out recently in a study conducted by MAGNA. We did a writeup on the results, but a quick summary is as follows.

  • They pit human-managed campaigns versus machine-learning powered ones, and the machines came out on top.
  • The machine-managed campaigns connected better with people who were more likely to convert, and drove better brand interest, preference, and most importantly, purchase intent and consideration.
  • We’ve seen similar results with campaigns managed by DSO, and as time moves on, AI’s worth in driving campaign performance will be clear.

 

AI Can Predict User Behavior

AI has a knack for predicting what users will want, or do, next. This can range from a product recommendation feed on an ecommerce site, to recommending additional content users may find interesting on a streaming service. Marketers can leverage this to up-sell their customers; if data shows users generally buy a certain combination of items together, or buy items in a particular sequence over a set amount of time, they would do well to serve ads with those items to users who have purchased an item in that set. This example isn’t as monumental as the one above — humans are capable of identifying these types of trends. But this is a great example of AI handling something that would normally take a human time to complete, and completing the task nearly instantly. AI can release people from time-consuming tasks and free them to work on things that can make a bigger impact, whether it be strategy or creative work.

Humans will always be better than machines at certain things, and vice versa. Sorting data and identifying trends is one thing, but imagining creative campaigns and strategy is quite another. It’s up to marketers to harness the advantages of AI, use it to create efficiency, and funnel it toward better marketing.

 

The Human Element Matters

Netflix is an industry leader when it comes to recognizing viewing trends and capitalizing on them. Their algorithm’s ability to parse through tons of data and make it actionable has resulted in a content recommendation system that has a lot more hits than misses. They even use viewing trends and habits to create new content itself, with House of Cards famously getting the green light due to their data showing it had a strong chance of being a hit.

But there can also be reason for caution. AI running on its own, unchecked, can sometimes lead to issues. Take for example the recent controversy surrounding Netflix’ methods of serving users recommendations based on their viewing history, leading to charges of “racial targeting”. This stems from some users reporting they would be served content recommendations using black or brown actors, just because they themselves were people of color.

Netflix has claimed there is no controversy, saying they do not target based on race, gender, or ethnicity. However as The Guardian’s Lanre Bakare writes, “…the fact remains that a user who enjoys watching Spike Lee or Ava Duvernay films are more likely to be shown ads using black and brown actors to promote a film or TV show, regardless of their prominence in the production.” Netflix’ algorithm sifts through data, identifies trends, and then applies what it thinks is best. Problems arise however when it ignores real-world circumstance and sensitivities, which can lead to brands ending up in hot water. AI, after all, is just a machine. This is a lesson to every marketer out there — AI can make things efficient but it doesn’t carry a human’s understanding of context.

This is a prime example of why the human element will always be a big part of AI’s effectiveness. Marketers need to ensure they act as a sanity check. AI may be great at sorting data and finding trends, but it can’t catch elements that a sound-minded person would know to be problematic. AI needs humans to shepherd it, responsibly, and channel it to ensure it is being used in the right way.

 

AI Makes Marketers Better

Marketers and AI are destined to have a very close relationship. AI allows them to be better marketers; it removes inefficiency from tasks and duties marketers are already handling. This allows marketers to go beyond menial tasks and focus more on strategy and other high-level tasks that can lead to bigger impacts. AI is nothing to fear, and marketers would do well to embrace it going forward.