It's happened to all of us: You do a few web searches for "Let's Rock Elmo" toys, and for the next few weeks you're barraged with ads for the furry red munchkin. Welcome to the world of retargeting.
By now, many large online advertisers have incorporated retargeting into their digital mix. For the most part, retargeting -- or the process of serving relevant ads to visitors after they leave your website -- delivers decent results. It boosts conversion among a select group of consumers who showed initial interest in your products or services.
But the truth is, when it comes to standard retargeting, there isn't much true "targeting" going on at all. Mostly, retargeting is a nuts-and-bolts type of system: Person X visited your website, looked at Product Y, and then left without buying. When Person X then travels around the web, they'll see ads for Product Y.
The basic way to execute a retargeting campaign is to place one bid, say $2, across ad exchanges for each consumer who's been on your website in the past 30 days. In other words, most retargeting providers simply bid $2 on anyone in the retargeting pool, on any page, at any time of day, in any slot on the page. Hammer, meet nail.
The result is that every single person who visits your site has the same ad served to them as they surf around the web -- repeatedly. This basic retargeting method is simply a numbers game that serves neither consumer nor advertiser well. Yes, an advertiser can reach a lot of people. But the vast majority of them won't be interested in your ad. And because they are so overexposed to the same ad, some will become hostile toward your brand -- obviously, the opposite of what you intended. Still, because this one-size-fits-all approach often looks good, measures well, and delivers good metrics relative to other tactics in an overall marketing plan, many advertisers are satisfied with their retargeting results.
But a new breed of retargeting -- call it "retargeting 2.0" -- powers the retargeting mechanism with artificial intelligence (AI). A retargeting system that continuously scores prospects, like a FICO score that changes every second, goes far beyond just showing people ads for a pair of pants they looked at. Instead, it shows them the best ads for products they are most likely to be interested in at that moment -- whether that's pants, shoes, a belt, Elmo, or a rental car.
AI is the secret sauce behind a new breed of retargeting that actually predicts which exact ad an individual will want to see -- when, where, and how often -- and then serves that ad to the individual in real time. These technologies analyze prospect's behavior: how many times they've seen an ad, whether they've responded, and how they've responded. They then analyze the performance of ads on specific pages, group prospects into dynamic buyer categories based on their real-time behavior, and then retarget relevant ads to each segment with the right frequency over the right period of time.
An advanced retargeting technology platform might learn to score a single impression based on multiple data attributes. For instance:
Using this scorecard, and accessing the data on thousands of computers in real time, a sophisticated retargeting technology platform can then evaluate billions of real-time biddable ad opportunities a day. Based on the data, it serves only those ads most likely to drive results for the advertiser.
Retargeting continues to evolve. As AI processes grow more sophisticated, it's already moving from 2.0 to 3.0. Advertisers need, and are now in a position to demand, a sophisticated, technologically savvy retargeting provider that delivers AI solutions. Artificial intelligence has the power to turn retargeting into a precision science that delivers the right ads to the right consumers at the right time. Retargeting that learns is simply better retargeting -- for brands and consumers.
Eric Porres is chief marketing officer at Rocket Fuel Inc.
On Twitter? Follow iMedia Connection at @iMediaTweet.
"Dart hitting a target on black background" image via Shutterstock.
Not a People Connection member?
This works well for an e-commerce site with many offerings (read AMAZON) that can fire off a more meaningful message. It may not work well enough for single-minded sites (read Eyeglasses).
Love the scoring methodology example. Super simple. Easy to wrap your head around. Easy to explain to non industry insiders. Also highly extensible when needing to generate a customized set of factors. However, your example only addresses the question of "when is a good time/place to bid on an ad impression?" The example stops short of answering the follow-on question "RTB impression has been bought, now what ad content do we serve?" Would love to see a Part 2 that dives into that half of the question.
Full Summit Calendar | Request Invite
1 The best social media campaigns of 2013
2 The most meaningless (and hilarious) job titles on LinkedIn
3 6 signs your agency is dying
4 5 requirements for a sustainable career in marketing
5 6 social media network updates that you missed