#26: Last Year / This Year with AI + Custom Algorithms
Quo Vadis Webinar
Reading Time: 7 minutes
Are you or is someone you know itching to learn a lot more about AI and Custom Algos? Well, that’s great news! Join our incredible Quo Vadis guests on April 20 at 11am ET to learn everything you need to know and more.
Our Round Table Guests
It's all about curiosity and education for our Quo Vadis community. As your faithful host, I'll bring a few provocative frameworks to drive the conversation with some really fine folks.
Between Eric and Sara, they pretty much know everything there is to know about custom algos. They are in the 100th percentile of knowledge. How do we know? Because if Bezos and Zuck are right about needing to log 10K+ hours to really know your stuff, then Eric and Sara are standing at the mountaintop with a crystal clear view.
Then there is Seb, by far one of the best journalists in the digital media space. One of the big reasons he gets it so well is because he has a degree in computer science. With way more than 10K+ hours under his journalist belt, Seb too is 100th percentile.
When we’re done with our webinar, you’ll be on your way to that exclusive space on the knowledge bell curve!
What We’ll Cover
What is AI really all about in display advertising today?
What's changed since last year?
Where will advertiser adoption be by year’s end?
The custom algo market size pie is certainly growing, how much will the pie grow?
Will advertising in the metaverse be driven by programmatic and custom algos?
You'll no doubt get great answers and a fun discussion from our top-notch round table chat.
Get your questions ready!
Something to think about
As they like to say in insurance advertising, “15 minutes could save you 15%!”
In the custom algorithm business, a little extra fee not only saves hands-on-keyboard people lots of expensive minutes, but the machine squeezes way more out of campaigns than a human could ever do. It’s a classic benefit/cost no-brainer.
As The Trade Desk points out, “At any given second, its platform processes around 9 million ad opportunities, data from 130 data providers, and inventory from 110 inventory partners.”
That amount and velocity of data present our human brains with an insane number of decisions to make. And as behavioral economists recognized many decades ago, we humans are not so great at making sound and rational judgments when overwhelmed with that amount of information and that many options.
No matter how talented, educated, or sharp, there is not a hands-on-keyboard person in the entire world who can figure out optimal campaign optimization choices, type fast enough to get it implemented in time, and then repeat this same exercise the very next second. Not possible. Full stop.
How do behavioral economists know this? Take the test.
You’re at a dinner party this weekend.
A friend introduces you to a woman named Genevieve.
He tells you that Genevieve recently graduated from Bryn Mawr College with a B.A. in Philosophy, where she was active in the Occupy movement and edited a literary magazine.
You’re interested in talking to Genevieve about the philosopher Hegel, the subject of her senior thesis, but your friend jumps in and asks you to rank the following statements about Genevieve in order of their probability.
Given what you know about Genevieve, rank the following statements from most likely to least likely.
Genevieve is a feminist.
Genevieve is looking for a job as a sanitation worker.
Genevieve is a feminist who is looking for a job as a sanitation worker.
Programmatic Case and Point
Allison is a programmatic specialist at an agency trading desk in NYC. Her colleague Jeremy is traveling today, so he asks Allison to look into a client’s campaign and make the right optimization decisions to boost performance.
But when she logs into her DSP, she gets stuck. The campaign is targeting tons of elements.
Allison is overwhelmed. She’s already thinking about how that after-work Don Julio chupito would calm her brain activity right about now. She has to think about soooooooo many options… her head is throbbing...
Desktop, mobile, CTV (3)
One 6-second video ad
One 300 x 250 display unit
Across 10 geographies
All the main browsers (5)
10 finite audience segments
40 SSPs are enabled, sending gazillions of bid requests every second
7 page positions
100 carriers
7 video placements
3 video sizes
“Holy f*ck!” Allison says under her breath. “How in the hell can I make the right choice when I have 882,000,000 to choose from? That’s not a fair game!”
She frantically texts her friend Melissa, “I need that shot now!”
Then she ponders the reality of invisible options, “what if there are only 100 best-optimization combos that drive the best results + or - 5% either way?”
She only has a 0.00001% chance of solving the puzzle.
“Wow… those are really shitty odds!” as she stares at the screen.
This moment of total desperation reminds her of something Ben Mezrich pointed out in his book Ugly Americans:
“You walk into a room with a grenade, and your best-case scenario is walking back out still holding that grenade. Your worst-case scenario is that the grenade explodes, blowing you into little bloody pieces. The moral of the story: don’t make bets with no upside.”
So what does she do?
Unfortunately, she uses her best judgment and hopes for the best. The odds that her best judgment is even close enough are slim to none.
If we humans suck at statistically driven optimization, then what are we good at?
Just because we can’t make optimization decisions better than a machine does not mean we aren’t totally superior at other things. When it comes to programmatic advertising, we humans are better at two things for sure:
Setting goals
Setting constraints
Rule of Thumb: Allision’s problem can be explained by how Ben Horowitz judges CEOs. Does the CEO know what to do and can the CEO get others to do what he/she knows.
The same two questions can be applied to Allison's dilemma. Does Allison know what to do (set goals and constraints = yes) and can she get the machine to do what it knows?
Triple Whammy Effect
Marketers are always looking for ways to self-fund new advertising frontiers. And they are also well aware that their programmatic working media has probably been terrible for years ($0.10 on the ad dollar at best).
As we like to say at Quo Vadis, every large advertiser has programmatic mice lurking about their house. Depending on the advertiser, some know about the mice and some don’t. Some care, some don’t. Some have tried to do something about it but most have failed or not tried at all.
And that’s why programmatic working media is either unknown, inaccurate or totally made up.
Buying an inexpensive mousetrap is as easy as a trip to your local Home Depot. But as Peter Drucker wisely tells us, “What you have to do and the way you have to do it is incredibly simple. Whether you are willing to do it is another matter.”
If Quo Vadis was running the programmatic show, we’d tighten our transaction fees and be maniacal about maximizing quality impressions. We’d focus on doing advertising first and spending money second.
You know, the kind of ads that are actually viewable way beyond MCR’s low hurdle rate (50% pixels, are you serious?) and to real human people that buy stuff.
We’d end up with so much found working media that we wouldn’t know what to do with it. And that’s exactly how marketers can self-fund innovations like custom algorithms. With clean working media data, we’d also be giving our AI algo true signals to make the best advertising choices. That’s a triple whammy.
Adoption Curve = Advantage
Whether you’re thinking about custom algos, artificial intelligence trials, attention metrics data collection, working media innovation, or creating first-party data to book your data as a real depreciable asset, you definitely want to be on the early side of the curve. Things are moving too fast not to be.
Now that you have a taste for what we’ll be chatting about, what are you waiting for?
Where can I find the webinar's recording?