AI and machine learning have impacted several industries far and wide, but perhaps one of the greatest disruptions is happening in advertising. Below are just a couple of examples.
The Trade Desk, an online advertising marketplace, receives an estimated 9 million ad bids per second. The company uses an in-house artificial intelligence engine called Koa to analyze bids in real-time and make strong advertising recommendations, ensuring that the right ads are seen by the right consumers at the right time.
In November, Lexus released the first-ever AI-scripted ad. IBM’s Watson built the 60-second commercial’s basic flow and outline, and the creative agency fleshed out the full story. The commercial was well-received and highlighted a key way for agencies to incorporate AI into their daily work.
Some ad executives speculate that a full embrace of AI could mean their demise — that the more sophisticated the technology becomes, the less purpose a traditional agency would serve. But in my opinion, this fear is unwarranted. Agencies don’t need to run from AI; they need to learn more about it and use it to bolster their processes.
How AI Is Shifting The Ad Industry
Both AI and machine learning are innovating advertising in a variety of ways:
• AI corrects the spelling of keywords during online searches, which leads to the display of relevant ads and takes consumers to the right webpages.
• Machine learning helps with targeted advertising by using data and advertiser preferences to reach the right consumers.
• Speech recognition is what drives virtual assistants like Siri, Alexa and Google Assistant. On the surface, these assistants help consumers order groceries, dim the lights or pull up webpages. But they also play a major role in helping consumers search for and discover products, services and information. And their searches produce data that improves the accuracy of the speech recognition and ads that users see.
• Like the Lexus ad mentioned above, machine learning is being used to study advertising trends and to generate new content.
• AI can also “synthesize data and identify key audience and performance insights,” which leaves more time for agency staff to do more strategic and creative work.
Agencies are warming up to the idea of AI: CMOs are investing in AI technology, even as budgets shrink. And those who jumped on board right away are seeing major ROI: 82% of early adopters reported positive returns. Those who are resistant could find themselves in competition with partially automated agencies that have more eyes, ears and data to outshine them. So that begs the question: How can agencies prepare for an AI-driven ad industry?
How To Incorporate AI Into Your Agency’s Work
There are a couple of key ways to weave AI into your everyday work without tearing apart your current workflow:
Start small with A/B testing.
Our agency uses A/B testing to build audience sets, which we’ve found to be instrumental in our process. It saves us countless hours of reviewing ad sets and converting successful conversions into new sets.
But A/B testing doesn’t need to be so advanced. It can be as simple as releasing two similar ads with different slogans or sending out two newsletters with different subjects. AI measures the performance over a specified period of time and reports which version performed best. Then, you can decide how to move forward.
Although the ability to perform these tests is now mainstream and has been incorporated into most native platforms, getting it right can be tricky. If you are unfamiliar with A/B testing or have had frustrating results, I’d recommend trying the following:
• A/B testing should be run for a minimum of 1-2 weeks but no longer than 4-6 weeks.
• Test apples to apples. If you’re testing the effectiveness of copy, it should be tested solely against other copy, with no other changes to the ad.
• Understand that it takes time. Let your testing run to completion before deciding which direction you would like to take based on the information gathered.
Shift to audience buying and planning.
The traditional ad model focuses on media planning and buying, but AI has created a more consumer-centric world. It has given us the ability to focus on “audience buying” models as opposed to the typical “media buying” models we are all accustomed to. The data that’s aggregated from ad platforms provides insight into where our audience will show up and how. This information can then be used to purchase segmented space within particular areas of a given space. This information is typically dealt with at a much larger scale by publishers selling ad space than it is by an entrepreneur running Facebook ads, but the concept is the same.
For example, you are now able to look well beyond website traffic as an indicator of where an audience may or may not be. Start with market research. Develop an audience set based on that information, and A/B test it. You can always refine and test it again. Don’t just let the data pile up; take the time to understand it and use it effectively to define and target the right consumers.
While I’d recommend using AI for what can be automated, it’s important to place a higher value on the aspects of your business that can’t be. Ensure your clients understand the passion and expertise you bring to insights and creative. Also, don’t try to clear the tremendous barrier to entry of making your own AI systems. Instead, form partnerships with companies that already have great technology. And while partnerships can seem exciting, remember to always ask yourself: What values are they bringing to the table? How does this move my business forward? And what type of training do they offer to use their technology/platform?
It’s true: AI and machine learning are changing the advertising industry, but with the right mindset and strategy, you can make a seamless transition and still create value for clients and consumers. You might even stumble upon a creative breakthrough. This isn’t the end of the agency; it’s a rebirth.