Commentary

Digging Into Our History of Intent

In the early days of behavioral targeting, "recency" emerged as one of the first very rough measures of data quality and intent. With the goal of targeting those users who were closest to the bottom of the purchase funnel, knowing that a person has read content within the last 30 days related to cars, bassinets, or golf clubs was supposed to be a signal of their intent to buy. All according to the category, intent signals that were 30 or 90 days old probably indicated that a purchase had already been made and the consumer was out of the market.

But in the social space, users leave a longer and more detailed trail about their interests that marketers can use for longer-term profiling and segmentation that can be hit again and again. After all, some aspects of our tastes rarely change. Once a sci-fi loving, first-person shooter playing, "Twilight Zone" watching, Magic card aficionado, always a sci-fi loving, first-person shooter playing, "Twilight Zone" watching, Magic card aficionado…I like to say.

One of the hot local mobile platforms, LocalResponse, has been using social signals for real-time targeting now for a couple of years. The company made its name first by harvesting billions of public social media declarations (on foursquare, Twitter, Instagram, etc.) and sending them Twitter messages with relevant promotions. While the platform continues to use public social signals to indicate user intent, it has expanded its message delivery to display advertising onto desktop, apps and mobile Web. You might tweet about or check in at your local gym and find a fitness-related ad in your desktop or app travels later that day.  

In a new twist on the targeting model, however, LocalResponse is finding that there is data to be mined from a longer view of social posting or what President Kathy Leake calls “historical intent.” They can look at patterns in social posts to create audience segments that are especially appealing to entertainment clients who are looking for taste groups. “These types of clients need to market throughout the year by genre,” she says. “Movie companies need to remarket to people who watched a romantic comedy in Q1 for a new release in Q4 in the same genre.”

Gaming companies especially have a need to remarket to existing customers of specific titles, since expansion packs and sequels are a central part of the business. Of course, the technique can also identify games frequenting competitive products. In the case of EA, for instance, they used historical intent to target a mobile in-app banner campaign for "Madden 13." They identified through their social declaration people who were expressing anticipation of the annual release. But they also grabbed people posting on competitive titles as well as people signaling a taste for the NFL, the major gaming consoles and even specific NFL players.

The audience was acquired in the mobile ad exchanges and targeted mainly mobile apps. The campaign produced CTRs in the platform’s typical range of .7% to 2%. EA is a long-time client, and of course represents a category whose audience tends to be tech aware and very vocal about their entertainment tastes in the social sphere. They leave a lot of traces of their taste. “Three years ago a client like EA was mostly spending on TV and on desktop,” Leake tells me. “We see a shift from clients like these to increasingly spending in mobile.”

Longer-term, however, it seems to me this is a method of targeting and tracking that lends itself to deeper metrics than the click-through, and Leake says they are working on post-impression tracking and the like for future campaigns. It seems as if these are pools of users whose tastes and pronouncements could be tracked and more deeply understood. After all, using social signals historically lets you establish some well cookied and profiled user bases whose subsequent social behaviors and posts then can be tracked after the campaign.

   

2 comments about "Digging Into Our History of Intent".
Check to receive email when comments are posted.
  1. Rob Schmults from Intent Media, July 8, 2013 at 2:14 p.m.

    Is this really intent that's being uncovered or does it more logically belong in the segmentation bucket? Not that there is anything wrong with that or that it diminishes the value of what Local Response is doing. The competitive titles example seems to uncover people interested in football console games -- exactly the right segment for EA to target, but it doesn't qualify as "intent." In the case of people posting "can't wait to buy Madden 13," these people are clearly expressing their intent (as well as fact you probably don't need to market to them). And it's a good example of how intent typically emerges: it's less about historical mining of data and drawing inferences and more about the user expressly exposing it, be it in a post, a search, or some other consumer-powered vehicle. The key thing for the marketer is to be there when it happens, as intent typically has a short shelf life as Steve Smith rightly points out.

  2. Jeremy Geiger from Retailigence, July 11, 2013 at 7:31 p.m.

    Agree with you Rob that intent is stronger when the consumer is expressing it explicitly. And within that, search in a mobile shopping app, is probably of greater value that a search for "bike" on a general search engine. That's why we think it's important that retail and brand marketers to be visible in such places, in response to such searches. To that extent, there are many ways intent data can be used intelligently in aggregate, to signal demand and drive effective marketing initiatives. For instance, if a retailer or brand could know which specific locations and geographies have the most shopper searches for their products, it would be a clear signal of demand and would be very effective for that retailer/brand to concentrate marketing in those areas.

Next story loading loading..