As we’ve talked about programmatic buying in the past, let’s start with a quick refresher:

“Simply put, programmatic buying is any ad buy that’s processed from one machine to another machine. It is a technology that automates the ad buying and placement, powering more targeted ad buys with the ability to respond in real time, through the use of software and algorithms.”

To distill it down further: An algorithm decides who will see your ad, what format they will see, and when and where they will see it.

But what about how—and why? Well, I’m glad you asked.

HOW

A pixel is placed on your site, allowing the algorithm to learn from your site visitors. From the day it begins observing user behavior, the algorithm begins sketching out a picture of our your “average” customer. Ex: They frequent CNN, not Fox News. They download coupons for diapers and milk from Target, and make big purchases from Amazon. They browse on their iPhone but make purchases from their MacBook. They’re more likely to click on a square ad than a rectangular ad. They browse in the morning and at lunchtime, but make conversions/purchases at night. In addition, and through these user behaviors, the algorithm is able to determine demographic info such as gender, age, parental status, income, and education.

This data is then used to create a seed audience, which in turn is used to develop a “lookalike audience.” A lookalike audience is an audience that exhibits similar characteristics and behaviors as your current site visitors. After the algorithm has learned enough about your current audience (learning time varies based on site traffic, but is typically 30 days or less), ads begin serving to the lookalike audience.

As ads are being served to the new lookalike audience, the algorithm is making hundreds of decisions every second:

  • Has this user seen our ad before? If so, what site was it on, what format (size) was it, and did they click on it?
  • Do they ignore ads with direct calls-to-action like “Buy Now,” and favor ads with CTAs like “Shop Today”?
  • Should we increase the bid to bring the ad further up on the page? Or should the ad remain at the page break in the article the user is reading?
  • Knowing the user’s behaviors, should the time of day be taken into consideration when serving our ads and/or increasing/decreasing our bids?

WHY

Rather than making broad assumptions about who you should target to grow your business, lookalike audiences use proven data taken from your current visitors to reach new audiences. These are users that aren’t actively engaging with your brand and who, based on current user data, are prime candidates for your brand. This data isn’t based on who you think your audience is, it’s based on who your audience actually is.

Using lookalike audiences in conjunction with a programmatic platform/algorithm allows this new audience to be reached via strategic, split-second decisions, ensuring that the right person gets the right ad at the right time. No matter the goal of your site or the business you’re in, the combination of lookalike audiences and programmatic buying has proven successful for clients of all sizes.

Got questions? Let us know. We’ve got answers!


Also published on Medium.