“ZUCK DECLARES WAR ON THE ENTIRE AD INDUSTRY!”
Blog Image by: Cath Virginia / The Verge; Getty Images
The Zuckernator was interviewed on Stratechery, then The Verge wrote about his statements claiming that Meta’s AI will handle all phases of advertising, and all anyone who sells stuff has to do is “plug in their bank account,” then the Meta machine will start depositing money, effectively displacing and rendering obsolete the advertising industry. The $1.1 trillion ad industry.
The Cannes Lions people must have their skivvies in a wad of unprecedented twist.
Any of you who know me IRL have likely heard me talk about “text-to-revenue.” Not text-to-insights, text-to-creative, or even text-to-campaign. Text-to-revenue is the inevitable future, and is exactly where we are going with our own platform development.
Let’s have a wee chat about what this Zuckified rhetoric all means.
THE INSIGHTS INDUSTRY JUST BREATHED ITS LAST BREATH
If you work for Kantar, Ipsos, Nielsen, or similar company, polish your CV, and preferably learn to do something new. Now.
Let’s do a thought experiment:
There is an ecosystem of behemoth social media companies that have daily access to the lives of billions of people. This company trains machines on their daily machinations, conversations, photos, musings, browsing habits, consumption habits, and so much more. Their AI can leverage all of this to not only predict what they will consume, but when they will feel a need to consume it. In real time. They will offer this for free.
In contrast to the above, there is a consumer research company called Ipstar. They will have junior people with “research” in their titles, conjure up brainstormed opinion-based survey questions, executed on a “panel” of a few hundred (presumably) human people who will self-report, then the company will charge you hundreds of thousands of dollars for a lazily assembled powerpoint deck of the “results.”
Which option sounds more useful, more effective, more compelling to you? It’s not a trick question.
Zuckerberg said in the interview that marketers who think their opinions are superior, and thus, mess with META targeting algos—and I paraphrase—are mouth-breathing idiots.
A client showed me a report from one of the big research agencies, one they paid $300k for.
⁃ The “segments” were not relevant to the product;
⁃ There were artifact images in the margins of the deck, meaning the research interns there could not be bothered to delete the ones they decided not to use;
⁃ Descriptions of traits at the extremes were basically the same, using the same wording. In other words, at the “Vanilla” extreme, it described them as “loves vanilla!” At the chocolate extreme: “Loves vanilla!” FFS.
⁃ The one “big spender” segment literally said: “price conscious and seeks finesse.”
If only I had a dollar for every time I found myself seeking finesse. WTF does that even mean? In any other industry, they would be justifying themselves in front of a congressional body. But not advertising.
My team ripped the report apart, pointing out all of the faulty logic, mistakes, carelessness, and then offered to show this travesty to the client. Twice.
They refused.
What’s the word for this again? Ah, yes, I remember now. Complicit.
THE AI/HI CREATIVE PRODUCTION INDUSTRY JUST BREATHED ITS LAST BREATH
A couple months back, Appier paid $38 million for AdCreative.ai, a company that produces shiny, Gen2 images. Obviously, a plurality of individuals at Appier thought this was a smart idea. Here is the issue: a Gen2 idea is a Gen2 idea. It’s based on liking, and those of us who practice Gen3 marketing know that liking is not what drives purchase.
Old school scenario: a bunch of human creatives come up with ideas, have the artists mock them up, then they sit around and decide which of the ideas they like best. Then they ask the focus group, then the client, then their mums. When peak consensus on liking is hit, there is adequate comfort, and the idea runs.
AdCreative.ai scenario: the AI conjures up a bunch of shiny “creative” images based on some prompts. A bunch of humans sit around and decide which they like best…and then ask their mums.
GIGO.
Both scenarios are based on liking. The resonance response, or the release of two brain chemicals, is the antecedent to action. This is science and the basis of Gen3 marketing. It has created every customer, ever, and has no relationship to liking.
Compounding the issue of Gen2 liking-based AI “creative” asset generators, was the looming probability that any of the big tech companies can—and will—and have—implement the same capabilities. This move to automation won’t stop with META. GOOG, MSFT, AMZN, and others no doubt have their own versions just waiting to be lightning bolted to life. Free of charge to users.
Icon.me, Adcreative.ai, Smartly.io knew this was coming. AdCreative managed to secure an exit. They are about to be vaporware. I reached out to a number of people at Appier to discuss the situation and was greeted by the sounds of silence.
As for regular carbon-based humans doing creative, the point I have been trying to make all along is that it doesn’t matter if the manscaped chin stubble contingent think the ads are shit, as long as they prompt action via the resonance response. If ads trigger the resonance response, that equates to branding, decision, sales, all in one motion. That’s what advertising is, fundamentally.
Yet it’s astonishing how many think advertising is entertainment-art that merits a rosé-soaked festival in the south of France. So many say “creative ads sell more!” The problem with this claim is that creative in this context is synonymous with liking, and we know how that fares in side-by-side taste tests. Liking is the wrong yardstick for “creative quality,” like measuring gallons in millimeters. Plus, what qualifies as “creative,” one summer-scarved marketer to another, really depends on what kind of expensive eyewear they are sporting.
“Creative” in advertising is what triggers a dopa/oxy release. That’s it, no matter who does or does not like it. Machines are already quantifying the hell out of this on an individual user basis.
What about content factories like Tag? Adding to Dentsu’s current woes, they paid $516M for Tag in the UK. Now all that capability will be built into Big Tech’s AI, from ads to posts, to content of every kind, tailored per user. Pretty soon, you won’t even need to scroll. Open your phone and mainline dopamine, watch that dragon burn.
THE CONSIDERATIONS SMORGASBORD
Now that we understand:
A) Big Tech has access to literal galaxies of personal data updated by the millisecond, and existing consumer insights companies have no way to compete with this;
B) AI, while not yet 100% ready for prime time, will be soon, and is going to munch the lunch of creative production;
We can now move to the general discussion of this transition.
BUT, BUT…BRAND!
Fueling the collective “pshaw” response from agency heads is exactly what you would expect, fruit hanging so low it would not escape a single business-oriented mind: brand and brand safety.
How could Zuck’s machines possibly be any good at brand?
In Part 2:
- “I’m in marketing, what the hell do I do now?!”
- How to think about man vs. machine vis-a-vis brand.
- The weaknesses of the proposed model.