Mar 15
AI: Advertising’s Refrigeration Moment
While many see AI as an existential threat to the advertising industry, I believe it’s actually the forcing function the industry has needed for years. I believe this because I see Generative AI as advertising’s refrigeration moment.
In its “refrigeration moment,” three groups of groups of companies will emerge from the advertising industry as we now know it:
- Those who ignore or underestimate the impact of AI on the advertising industry. Much like ice delivery companies after the advent of refrigeration, these companies will suffer mightily.
- Those who seek to directly monetize AI. Much like those companies who sold refrigeration devices, the companies who equip agencies and brands with AI tools will make good money (but that’s not where the real money is).
- Those who create new types of commercial creativity that are only possible because of AI. Much like Coca-Cola was able to build an empire because of what refrigeration enabled, the real money that will be made in advertising during the AI age will be those who use the technology as a springboard. The ones who win big will be those who create new forms of commercialized creativity that simply were not possible before AI.
But how will this happen and how will the real money get made? Let’s get into it.
Before we get started: yes, this time it’s different
Having been in and around advertising for a long time, I’ve seen first-hand that waves of panic are nothing new for the industry. Every few years or so, some change comes along that has the industry running for its battle stations (a lineage of panic that dates back through crowdsourced creativity, the growth of in-house creative departments, the emergence of social media platforms as creators of advertising and not just sellers of it, and on and on).
So you might think that AI is going to be like these prior moments of ad industry panic: executive teams will have a bunch of meetings about it and there might even be a task force formed, but business will more or less go back to normal in a year or so.
The reason why AI is different is that it’s not a change within the industry (like the growth of in-house creative departments). Instead, it’s an underlying technological change that transforms how a brand can build an audience. Except this change is coming much, much faster than anything we’ve seen before.
Ruh-roh. Better form that task force fast. But before you do, let’s look a bit more closely at the real opportunity.
But AI opens up Coca-Cola money, not just refrigeration money
As Chamath Palihapitiya points out in this video below, and as I alluded to earlier, the true fortune made at the advent of refrigeration was made by Coca-Cola (not the refrigerator makers and definitely not by the ice delivery folks.)
The innovators at the heart of The Coca-Cola Company realized that refrigeration enabled them to sell the world a new kind of refreshment: all they needed to add was a secret formula and some marketing magic (and — well yeah — maybe a little cocaine).
But Chamath goes on to explain:
“If you take 1,000 of the same inputs and give it to Facebook and Microsoft and Google and Amazon, they’ll come up with the same machine learning model. But if you have one extra thing, one little ingredient, that all of those other companies don’t have, your output can be markedly different.”
I believe that one of the greatest challenges facing the advertising industry today is that agencies are increasingly using the same inputs and as such are generating increasingly similar outputs, much like the identically-trained machine learning models referenced in Chamath’s quote above.
However, properly harnessed, I believe that AI has the potential to become that “one extra thing” that enables creative output that is markedly different. Indeed, I believe that GenAI can become the springboard to a new era of commercialized creativity, generating Coca-Cola money and not just refrigeration money.
To achieve this, I believe that three significant shifts are necessary.
Shift #1: Transformation, Not Advertising
For far too long, agencies (and consultancies, for that matter) have been hired to be Winston Wolf, but incentivized to try and be Bonnie’s breakfast companion. For those of you who haven’t watched Pulp Fiction as many times as I have, what this reference means is that advertising agencies are always at their most valuable when they are in the transformation business, but still too often monetize their business by charging for the time and people it takes to make advertising artifacts.
When an hour spent solving a core business challenge bills out the same as an hour comping a social post announcing the annual Arbor Day sale, the financial structure rewards the scale of the team rather than the scale of impact; and the longevity of a client engagement, rather than the incisiveness of it.
GenAI will finally break this model. Clients will increasingly use GenAI to do things that were previously done by agencies (it’s already happening). But the key here is that agencies should not react with umbrage and argue that there’s no way that GenAI can do as good of a job with the aforementioned Arbor Day ad; rather, agencies need to step back from the scope that GenAI can create in a way that is “good enough,” and focus on clarifying and monetizing the part of the scope that needs to be far more than good enough. I believe that agencies need to rebuild to be fully in the transformation business instead of the advertising business.
Your initial reaction might be that any reduction of scope is bad business by definition. But the reality is that the average agency new business pitch is already 35% smaller than it was the year prior, and that’s before agencies had to compete against GenAI. So figuring out a way to build a business around Winston Wolf-like moments isn’t surrendering revenue; it’s proactively ensuring your survival.
Shift #2: Narrative Violations, Not Brand Narratives
There is nothing that advertising agencies revere more than the brand narrative (except, maybe, the memory of commission-based fee structures). Brand narratives have been the very essence of Madison Avenue myth-making: spinning stories from thin air that turn pedestrian products into icons. Indeed, Coca-Cola itself created one of the most powerful brand narratives ever.
While storytelling will always be one of the cornerstones of marketing, I believe that the age of AI will require a very specific approach to storytelling: one that seeks to create narrative violations instead of brand narratives. To delve into why this is the case, let’s spend a minute looking at how the current models of AI actually work.
At its core, what LLMs (Large Language Models that power AI) do is most accurately described in terms of pattern completion. Your inputs give an LLM the beginning of a pattern, and it uses its model of the statistics of human language to tell you what words are likely to come next to complete the pattern. You give an LLM the outline of a story, and it generates pattern-conforming content that fits our collective mental model of what feels “true.”
In other words, AI definitionally generates the familiar. And so, as AI is used to generate more and more of the executional manifestations of marketing, it will lead to increasingly similar brand narratives. This is why I believe that advertising will need to take dramatic steps to re-focus on creating narrative violations instead of brand narratives.
When Bedrock Capital introduced the term narrative violation a handful of years ago, it briefly became so popular that it became the kind of popular narrative that the idea was designed to escape. But I believe, much like disruption and zagging have in the past, narrative violations can become a very useful strategic framework for brand innovation in the age of AI.
Increasingly, it will not be sufficient for a brand to have a potent brand purpose or a compelling brand narrative. To thrive in a world wherein patterns are increasingly completed by AI, brands will need to establish and deliver on their narrative violation that breaks free from the patterns that surround them.
So, in the Age of AI, transformation replaces advertising as the business that we are in. And narrative violations replace brand narratives as the strategic framework. But what about the communications itself?
Shift #3: Springboards, Not Moats
While the two shifts explained above will enable those in the advertising industry to escape the ice delivery business in the age of refrigeration, they won’t quite lead to Coca-Cola money.
To reach the full economic potential of AI, an additional third shift is necessary: identifying one of the many ways in which GenAI can become a springboard for what is only possible in advertising because the technology exists, rather than furiously focusing on trying to dig a moat to protect what AI threatens to replace.
To begin to explore what one of these spaces might be, let’s start with Ryan Reynolds. In one corner of his ever-expanding Midas touch empire (see also: Aviation Gin, Wrexham FC, etc), he recently made the following ChatGPT-created ad:
Rather than talking about whether I think the script is any good, I’m going to focus instead on the fact that the ChatGPT-created ad for Mint Mobile could be placed anywhere in any media placement. And this ad’s infinite flexibility is in fact its Achilles heel — and one of the areas of enormous opportunity for creativity in the age of AI.
For the vast majority of advertising, the ad’s meaning and impact are completely independent of the context in which it is placed. To be clear, I’m not suggesting that advertising is placed randomly: clearly, FAANG companies and many others have made billions based on a tech-enabled understanding of who is seeing the particular ad. I’m talking about cultural context: syncing the specific meaning of an ad with the very specific cultural milieu into which it is placed.
One of the most famous context-enabled campaigns in advertising history was for Absolut vodka. Take this example, custom-created for the one moment in time when Steven Spielberg was awarded the AFI Life Achievement Award:
Custom-to-their-context ads like these tend to clean up at award shows but are rarely done because they are unimaginably inefficient: teams of people at agencies are needed to research the nuanced understanding of individual media placement and in turn, create a custom-crafted unique ad for every placement — and each ad only runs once. Doing this at any scale has been de facto impossible.
But, with generative AI, I believe these infinitely-custom campaigns can become imminently possible.
You would start with an idea that creates the narrative violation (an idea that can only be generated by human imagination, because of the pattern completion nature of AI we discussed above in shift #2).
But then, instead of traditional creative campaigns, agencies would apply systems thinking to prompt generative AI to generate an infinite number of unique ads, custom-matched to the specific context in which they appear. Quite literally, millions of custom “Absolut” ads: each tailored to its unique digital environment and moving at the speed of culture.
Summing up how to win in advertising’s refrigeration moment
The three shifts explained above outline how the advertising industry can re-think the three pillars at the core of its future.
- It can re-think its business model by shifting from being in the advertising business to being in the transformation business.
- It can rethink its strategic approach, by creating narrative violations instead of brand narratives.
- And it can re-think its executional approach, by focusing on how GenAI can springboard to new forms of commercialized creativity — instead of digging moats to try and protect yourself from what GenAI might replace.
And if you don’t make these shifts? Well, deliver as much ice as you can while you still can.
