ellipsis flag icon-blogicon-check icon-comments icon-email icon-error icon-facebook icon-follow-comment icon-googleicon-hamburger icon-imedia-blog icon-imediaicon-instagramicon-left-arrow icon-linked-in icon-linked icon-linkedin icon-multi-page-view icon-person icon-print icon-right-arrow icon-save icon-searchicon-share-arrow icon-single-page-view icon-tag icon-twitter icon-unfollow icon-upload icon-valid icon-video-play icon-views icon-website icon-youtubelogo-imedia-white logo-imedia logo-mediaWhite review-star thumbs_down thumbs_up

Look-alike targeting's new frontier

Look-alike targeting's new frontier Michael Horn

This past summer, Facebook announced that advertising had come to Instagram, which will enable advertisers to scale their Facebook campaigns considerably. Yet marketers are looking for not just more scale, but also more of the right audience. That's why one of the most popular features on Facebook's inventory is "look-alike" targeting.

Look-alike targeting enables marketers to expand the scope of an audience, but it takes more work than simply checking an "expand my audience" box. In fact, there's quite a bit of risk that if the look-alike audience isn't really similar to the base audience, it might just undermine the purpose of targeting in the first place.

Generally, there are two problems with look-alike targeting. First is the lack of standards for what "look alike" means. There is no commonly accepted definition. For example: take a quality target audience of frequent international business travelers from United Airline's DMP. Once all the clearly identifiable frequent international travelers are accounted for, the slippery slope takes over. Consumers who get added into a "look-alike" audience segment for this original group could be occasional international business travelers, or frequent first-class domestic travelers, or perhaps even people who read about international business travel online anywhere across the web.

The second problem with look-alike targeting is that it generally includes only third-party online or offline behavioral data, and ignores the increasingly vast amounts of data available that explains why the behavior occurred in the first place. Two consumers may have similar online behaviors in common, such as favorite news sites visited or brand of sneakers purchased, but might actually have totally different needs and motivations for why they did those things. If the marketer just assumes the underlying causes of the behavior are similar, the messaging might lose its influence.

In order to maintain relevance while increasing scale, marketers should embrace "think-alike" targeting. Think-alike targeting adds scale to the audience not by watering down the definition of "alike" behaviors, but by getting to the underlying sub-segments of the people based on why they behaved that way. In other words, think-alike targeting enhances the targeting criteria by slicing the larger audience into smaller groups based on new dimensions, not by diluting the behavioral dimension further. 
Here's how it works. Marketers first expand their target as far as they reasonably can based on the proven behavioral predictors. Then, that audience is enhanced with additional data elements and further analyzed to understand the underlying attitudinal and motivational sub-segments. So the original audience is now two, three, or more sub-audiences who are alike in their behaviors, but very different in their motivations. Each of these sub-segments is then expanded to find new consumers who have similar values, perceptions, or motivations, and the messaging and media reach strategies are adapted to locate and engage each sub-segment.

For example, if the original audience of interest was "people who searched for a credit card," look-alike targeting might seek to further increase the size of this group by searching for people who made other similar searches or visited financial websites. Instead, marketers should expect their agency and media partners to go deeper by layering on data that first exposes differences within the credit-card-seeking group like what their interests or motivations are. We can look then for sub-segments based on people who need liquidity versus people who aspire to free vacation travel versus people who want to shave a few dollars off monthly interest expenses. The appropriate credit card offer can then be made in a more relevant way to each sub-segment, and likely in environments where there are fewer competing card offers.

One of the most important additional dimensions to add is context. This is especially important for Millennials and younger online audiences who actually expect more relevant messaging from marketers, particularly on mobile. Marketers should not expect that people who search for a product on their phone while they are in a store should get the same message as someone who searches for it while at work on their laptop. The context of these two different people is at least as important as their product search information. One person may be just about to purchase, in which case an offer for a discount might be well worth it. The other person may be just browsing, in which case the discount offer might be premature, but then might need to be continued throughout the person's purchase journey or the marketer may risk losing the sale entirely.

It's still not practical for marketers to target individual messages to each consumer. Personas are still useful tools. Yet with all forms of communication, is important to try to build the messaging from the consumer up, rather than from the persona down. This is particularly effective when improving look-alike targeting with think-alike targeting. While both achieve scale, one simply dumbs down personas to even wider data consideration sets, while the other layers on more nuanced and specific data in order to enhance the personas and stimulate more consumer engagement.

Consumers have adapted to the reality that marketers know who they are, where they are, and what they are doing. Increasingly, they expect marketers to also understand why they are doing these things and reward the marketers who live up to their expectations of relevant engagement.

Michael Horn is SVP chief analytics officer at Resonate.

On Twitter? Follow iMedia at @iMediaTweet. 


to leave comments.