As I documented thoroughly in a previous post of mine, I hold the firm belief that great marketing can barely be noticed. It’s synonymous with content - serving itself exactly where and when people are ready to consume it.

With the sole exception of content marketing, this means figuring out exactly how to align your brand with content already being consumed by your target audience. As discussed in my previous essay, this is very difficult.

I frequently describe this “brand-content alignment” as a push/pull mechanism. The more well known “push” methodology means figuring out how to insert your brand into content while maintaining contextual relevance. This could simply mean it’s better to advertise your cookie dough on a cooking show rather than a gaming stream, but also often includes smooth transitions from organic content to ad, and creative product placement.

The “pull” mode is far less explored, partially because the end game is more neuroscience than it is marketing. A hyperfixation on catching the users' undivided, focused attention has pulled marketers even further away from a dream goal: multiple ads, one piece of content.

It sounds silly. Common reasoning says that if we have a user's attention with a piece of content, we should use it to overwhelm them with the upside of a single product, maximizing the chance of a conversion.

I like to use the example of an attractive female model posing on the beach. Marketers tend to assume that at this moment, consumers are most interested in purchasing a bikini. What about the beach towel, the sunglasses, the phone case, the sandals or the hat that random jogger was wearing in the background? Perhaps there is an understated interest for other commodities in the picture.

By algorithmically tagging all of these products, we can massively increase the amount of real estate available for advertisers. Like in today’s digital advertising ecosystem, pricing can follow proximity. Just as search ads listed higher in your search result are more expensive - tagged items in digital media that are closer to the foreground would also be priced higher.

With the advancement of GAN & other types of generative ML models, this becomes even more interesting. Part of the reason that Facebook & Google’s Ad networks are so successful is because of their ability to spin up Ads quickly, on a completely programmatic basis. The “push” methodology could be used to power a full fledged ad network precisely in this way. Advertisers could spin up campaigns, and by using generative AI, create exposure for their products in real time by “pushing” into relevant user-generated content in a completely discrete fashion.

The commonality in all of these techniques is seamless product placement. Never in the history of marketing have advertisers been able to truly blur the line between organic & sponsored content. The closest we’ve come is audience targeting with contextual queues - building a profile of user interests based on the content they consume. The problem is that common display formats on social networks and other publisher real estate stick out like a sore thumb next to organic content. When scrolling through social media or browsing the web, it’s easy to tune out ads that are visually separate from organic content, especially when those ads might pertain to your interests, but not necessarily the content you’re consuming at that moment. As expected, this context switching reduces the chance of a clickthrough on your ad to near-zero.

Rather than feeding interests extrapolated from content into a pool of interests for that user, and serving ads in a random sequence in the feed, we should maintain the link between interest and content. I call this link the “invisible pull.” If the pull mechanism means labeling visual content like hats and shoes, the invisible pull means developing an understanding of the thoughts produced by this content.

All conscious thought is built from sensory input. You might walk into a kitchen and smell the scent of baking cookies. The sensation is the smell of cookies, but the perception is infinitely more complex, and can produce thoughts like how Grandma used to bake at the family parties decades ago.

Modern digital media that’s optimized for sensory pleasure produces the same kind of thought. If tagging & pushing products into media added tons of advertising real estate - imagine what’s at stake when we can draw correlations between content and how it’s perceived in ways that aren’t obvious.

Many times, thoughts associated with content are specific to an individual (such as the family party). Through the use of big data, we can measure more universal “background” cognition (or daydreaming) generated in response to content consumption.

In the cookie example, it might be worthwhile to measure the interest in hand-knit blankets or vintage clothing, since they’re often correlated with grandparents.

When combined with audience targeting tools, this data becomes extremely powerful. It allows us to measure, on an individual level, precisely the kind of media attributes that stimulate specific memories and thoughts.

The magic of this is “infinite real estate.” In the last several years, content consumption has increased massively. More devices, more touch points, and more media consumed in less time. All of this has a limit. There’s only so much media we can consume in a finite period of time. When we abandon our obsession with dedicated Ads, we'll be closer than ever to unlocking the near infinite depth of video content & capturing the attention of our customers when it matters most.