How AI is curing advertisers’ privacy problem

Gone are the days of running a few big, bold spreads in a popular magazine or launching a digital homepage takeover on a few high-traffic publisher sites. By next year, the global cross-platform and mobile advertising market is expected to hit nearly $300 billion. So, rolling out a campaign that spans social, display, streaming and more isn’t the exception. It’s the (new) rule.

While this approach casts a broader net and enables brands to meet consumers where they are, two challenges continue to greatly worry advertisers: consistently and effectively targeting ads to relevant audiences and, from there, attributing performance to business results. These concerns exist across platforms — and the more campaign extensions, often, the murkier the targeting and attribution efforts seem to become.

Now, though, there’s a new layer. Enhanced privacy regulations mean advertisers are losing access to granular details that prior have helped hone campaign targeting and overall performance. 

The problem with modern advertising 

Enhanced privacy policies are making digital advertising even more challenging. While identifying and effectively targeting discrete audiences at scale has always been a tall order, data that once guided these decisions is now off the table. 

But it goes deeper. Much of the data advertisers can collect is purely estimation — many users opt out of onsite and in-app tracking. Up to 88% of Facebook users globally — and 96% in the U.S. — have opted out of app tracking, for example. Pair those staggering numbers with platform-specific privacy policies and advertisers are staring down incomplete metrics more often than not. 

Leveraging AI in campaign development and targeting  

Increasingly, brands have been tapping artificial intelligence (AI) to improve targeting and attribution. AI can help advertisers reach desired audiences based on creative inputs. Compare this approach to the old guessing game where pre-planned audience segmentation drove strategy.

By leveraging AI, both platforms and advertisers can better optimize broader-based targeting, relying on the media platform’s algorithms to collect countless data points while delivering the right messages to the right users at the right time.

AI is also improving measurement and attribution, bridging the insight gaps modern privacy policies are creating. Using predictive modeling, advertisers can now efficiently and effectively fill these holes.

Finding your audience needles in a massive haystack 

AI-driven algorithms can analyze vast amounts of data, identifying patterns and insights that help find, engage, and activate your ideal customers. No human media planner would be able to execute at this level and at this speed.

Equally important, with AI, advertisers can go beyond demographic data, tapping into behavioral and contextual signals to deliver more relevant ads to audiences. Lookalike modeling can still apply here, too — working off of a specific data set; advertisers can create similar audiences based on how closely individuals and segments align with the original audience. Many connective insights that modern media algorithms can uncover were previously hidden from advertisers and media experts.

Relevancy at scale 

AI can help advertisers deliver far more relevant messages to individual consumers based on their interests, browsing history and other key factors. This approach can help increase engagement and conversions by delivering more noticeable content when, where and how consumers are most likely to take action.

As the algorithm feeds many different creative elements to a broader audience, it quickly learns who responds to what ad types in what environments. This enables AI-powered systems to optimize delivery based on behavioral signals. It recognizes that this type of consumer who sees this ad on this platform at this time, for example, is likely to click, browse and take a critical action — be it a purchase, email opt-in, or other KPI.

As a result, there is a movement towards broader targeting and more creative variations within the social media industry because the algorithms can target sub-segments within the broader audience by using many more creative assets and machine learning (ML). And they can do so much more effectively than human media planners with their manually pre-planned segmentation and message maps.

That leads to another AI use case: Scaling content and creative production to keep pace with campaign demands. Using generative AI tools, brands can create wide variations in copy, text and even full images and video more efficiently. The more creative variations you have, the more the algorithms can learn and deliver. Generative AI breaks the compromise that advertisers have historically faced between rising production costs and better media delivery through algorithms. Quality creativity in a multitude of versions can be fed to media algorithms at increasingly lower costs.

Developing and delivering dynamic ads 

Ad creators now use a deconstructed approach, uploading multiple images, videos or copy variations and then letting combinations of machine-constructed ad versions be served up via AI. The goal: Feed the algorithms enough content elements to support a statistically sufficient number of user data interactions between creative and audience types. AI will measure and optimize toward the best combination for targeting.

Targeting ads in this manner has become much more difficult through cookies and MAIDS (mobile ad IDs). Done manually, identifying and acting on such performance insights could take weeks for data acquisition, analysis and production. AI can change ads based on a short run of a few days displaying the ads in different combinations. Having this level of adaptability gives advertisers a chance to reduce waste and maximize impact, making even more out of every campaign dollar. The consumer experience is also enhanced through more relevant content and ads.

The future: modeled conversions

The most interesting emerging area of application may be leveraging AI for attribution. Even now, many platforms estimate ad conversions because privacy policies and data limitations heavily limit direct tracking. AI-powered marketing mix modeling can help forecast campaign performance at a speed and cost that is far below historic norms.

AI can also help brands navigate around the need to handle cookie files and interpret consumer response data while still adhering to privacy policies. Many CRM platforms increasingly have these features built in. Given the insights emerging from AI-powered modeling, new opportunities are emerging for data connectivity via SaaS-based advertising. 

Today’s digital advertising landscape is challenging but also very promising. To optimize the success of your campaigns, stop trying to outsmart the machines. You will not win.

The predetermined segmentation schemes of yesterday will not help algorithms get the results you seek. Instead, lean into AI-aided algorithms by feeding them more messages, more images, more video and more templates. Give AI the content and creative elements it needs to optimize.

That being said, don’t forget that your audience is human. Brand and creativity still matter, and this is where the human element remains supreme. A campaign hook built upon a big idea will only help AI deliver even better results. So focus on creativity and brand values and let AI take care of the in-the-weeds execution. 

Max Cammarota is director of paid social and performance media at Beeby Clark+Meyler (BCM).