February 4, 2014
Ever since it was introduced, retargeting has been a boon to direct response marketers. Retargeting, or sometimes called remarketing, has been so effective that all the networks have created their own ability to retarget. Recently, Facebook has made its own retargeting pixel, the Web Site Custom Audience, available to PMD agencies.
There is, however, one limit to Retargeting: volume. Since retargeting requires that the audience must have interacted with the web site before (e.g. visited the web site, submitted an email address, clicked on a display ad), the volume may not always be large. Wouldn’t it be nice if you can use the “seed audience” of your converting visitors, and “look for” similar audience in a particular network ?
Look-alike Targeting addresses this problem. Look-Alike Models are used to build larger audiences from smaller audience segments to create reach for advertisers. In theory, they reflect similar characteristics to a benchmark set of characteristics the original audience segment represents, such as in-market kitchen-appliance shoppers. Generally, the higher the marketer’s tolerance for loosening the model, the larger the segment can become.
Lookalike audiences can be used to support any business objective: Targeting people who are similar to sets of customers for fan acquisition, site registration, off-Facebook purchases, and coupon claims, or simply to drive awareness of a brand. It can also be used to find audience who put an item in the basket of your web site, but didn’t pay for it.
Look-alike audiences has been made very popular recently by Facebook but brand managers and agency managers should know these four little known but important things:
1. Facebook is not the only network that gives advertisers like you the ability to use Look-alike audience.
Google Display Network and Yandex also offer lookalike audience capabilities. Facebook’s lookalike audiences uses social graph data. Google Display Network uses behavioral targeting to find similar audience to the “seed audience”. Yandex uses its Crypta-powered targeting technology and its analytics tool called Yandex Metrica to predict lookalike audiences.
2. Make sure your seed audience is clean and well-selected or else the Look-alike targeting algorithms will not work.
Whether the network uses social data, search data, or behavioral data to deliver lookalike audiences, the network’s predictive models for lookalike targeting must lead to higher conversions. Furthermore, lookalike targeting predictive models are only as good as what the inputs are. Garbage in, garbage out. Make sure your seed audience converts really effectively before you expand the seed audience with look-alike targeting. Go for quality before quantity for your seed audience.
3. With Facebook, you can set up your own Lookalike audience with the power editor.
It is little known that Facebook’s Power Editor can be used to set up Lookalike Audience. Next week, we will walk through the Power Editor to show you How to Set Up Lookalike Audience in 5 minutes.
4. Networks must ensure that consumer privacy is not being violated as they apply their algorithms to search for Lookalike models.
Theresa LaMontagne, managing partner, senior practice lead, MEC Analytics & Insight, stated that “Unfortunately, buyers are limited in their ability to reliably gauge the quality of the underlying data such as quality and accuracy of the underlying targeting source data and the definition of the data sources used including the size and age of the underlying data set. Marketers also need audited assurances and standards to ensure consumer privacy is not being violated.” Sensitive areas that networks must not step on are private message communications and emails.