Most people get excited about the holiday season because it brings presents, a holiday party circuit (bonus excitement for open bars), and unabashed gluttony when faced with yummy food and beverages.
Know what excites me? THIS:
My holiday season just won’t reach its “merry and bright” potential if my fur baby isn’t donning gay apparel. Each year I’m on the hunt for a new holiday ensemble that suits my kitty’s svelte figure. And – while it may surprise you – I’m not alone.
What does this have to do with artificial intelligence? Well, that kitty is actually a robot.
Let’s dig in and see what this fuzzy friend has to do with today’s topic: bid optimization, step 3 of 3 in AI-powered advertising.
Quick refresher! At a very high-level, there are 3 main steps to AI-powered advertising. The first two are:
The “bid accordingly” piece is value-based bid optimization (step 3).
Unlike most things related to AI, this one is pretty straight forward. The definition is in the name.
Basically, value-based bid optimization is when the AI makes bid decisions based on the value of each impression for your campaign. It uses the impression-level value scores (from step 2) to:
OK, if you insist. Here’s a less boring way to think about value-based bid optimization:
Let’s say you’re advertising for FestiveFelines.com, a [totally made-up] kitty outfitting e-tailer that wants to increase sales of little numbers like this for the upcoming holiday season (mrrrrow…).
Let’s also pretend there are just 4 available impressions in the auction stream (instead of 4 gazillion):
First, Magnetic’s AI uses Connected Consumer Profiles and predictive modeling to assign value scores to each person.
Then the AI starts value-based bidding optimization to figure out who FestiveFelines.com should serve an ad to, and how much to pay for each impression.
The AI decides Alf is out; low-value, and not worth FestiveFeline’s dime. The remaining users are of moderate-to-high value to the brand. Taylor is least like to convert, Eleanor is more likely, and I’m most likely (and thus of greatest value). For simplicity’s sake, we’ll assume all else is equal (e.g. publisher, context, timing, etc., is the same, although factors like these will impact price).
The AI determines the right price for each impression to be:
Because 50% of the available impressions are high-value, the AI decides to put FestiveFeline’s money where it matters most and serves ads to me and Eleanor.
Bid optimization complete!
Now… obviously there’s a bit more math and science (e.g. value scores are calculated and assessed to multiple decimals places). But that’s the basic concept of AI-powered bid optimization: find the most valuable impressions, pick the best ones to buy, and pay the right price based on potential return.
Do you like spending more money than you need to? Do you want to miss out on the best prospects or customers? If you answered “yes,” then by all means, keep using antiquated bid practices and manual price-setting.
But if you want your advertising to be smarter, faster, and more efficient than ever before, you need AI-powered bid optimization. Here’s why:
Want to learn more about all things AI? Stay tuned for more from our “Explaining AI” series, or check out our previous posts to dig into topics like Data Science, Predictive Analytics, and Machine Learning.
Magnetic is an artificial intelligence company that uses machine learning to deliver smarter, faster, and more effective advertising. Our powerful AI platform continuously analyzes the attributes of over 350 million live user profiles alongside real-time inventory supply and bid opportunities to deliver highly performant audiences and profitable campaigns for our clients.