On Wednesday, Twitter had a blog post about "Experimenting with new ways to tailor ads." They provided the following example:
Let’s say a local florist wants to advertise a Valentine’s Day special on Twitter. They’d prefer to show their ad to flower enthusiasts who frequent their website or subscribe to their newsletter. To get the special offer to those people who are also on Twitter, the shop may share with us a scrambled, unreadable email address (a hash) or browser-related information (a browser cookie ID). We can then match that information to accounts in order to show them a Promoted Tweet with the Valentine’s Day deal.
It is an interesting move by Twitter to engage in behavioral targeting, which is a proven tactic in online advertising. What we're seeing more of in the industry is what data is brought to bear to inform and improve re-targeting.
For example, site re-targeting was huge when it first came out. The idea of being able to serve a "hey, you, come on back and finish that order we know you were looking at on our site the other day" online ad to a cookied user who had come to your site was fantastic! It still is.
As re-targeting grew, people got more creative with data, especially ways to tie various online data sets via cookies or other means. In Twitter's case, an email address or cookie would be tied to an advertiser's first-party data. I've also worked with brands that hand over their first-party CRM data (email addresses -- which are needed as the primary key to match to another data set -- and key segmentation variables) to a DMP/DSP that attempts to match user's cookies to re-target them across ad networks. In some cases, an intermediary like LiveRamp helps marry the data sets to facilitate this process.
The other targeting trend I am seeing that I think could show promise is 3rd-party data append of offline data to online behavioral data. For a long time, classic direct marketers have appended data from the likes of Experian or Acxiom with their CRM data to learn more about their customers' interests, financial profile, or shopping behavior. The challenge for a long time is with anonymous cookies, how do you tie that to a real person offline. Email was and still is a great primary key as consumers often use a specific legitimate email address for online shopping, signing up for newsletters or contests, etc. and usually keep that email address updated. That led to companies like Rapleaf whose main business is matching users via email addresses.
Then Facebook took it to a new level earlier this year when they announced a partnership with major 3rd-party marketing data companies, such as Epsilon, Acxiom, and Datalogix. Those names sound familiar? With their new self-serve Partner Categories ads, Facebook claimed it was a new way to target ads to more categories of people. For example, a local car dealership can now show ads to people who are likely in the market for a new car who live near their dealership. To date, advertisers have been able to show ads to people based on their expressed interests on Facebook. Now with Partner Categories, they can also show ads to people on Facebook based on the products and brands they buy online or offline. This has the potential to marry tried and true techniques of direct marketers with the power of the social graph!
And then there's FBX -- Facebook Exchange, which allows real-time bidding and behavioral targeting with data from third-party Web sites.
It's too early to tell if all this is working at Facebook. But in one early report of Partner Categories tests run by Facebook partner Social Code using a wonky KPI called "engagement per Like (EPL)" showed promise.
For a data-driven strategist like me, it's fun to get my hands dirty with first-party CRM data, 3rd-party data appends, and these social networks. What an exciting time in the industry!
##
No comments:
Post a Comment