Behavioral Sequencing: Identifying Intent Before Intent

jamesmalins“Data-Driven Thinking” is written by members of the media community and contains fresh ideas on the digital revolution in media.

Today’s column is written by James Malins, vice president of cross-channel solutions at Adconion Direct.

How do auto dealers know you're about to be in the market for a new car? They don't.

The moment a third-party data provider tells an auto dealer a consumer is in-market for a hybrid vehicle, it's already too late. It's not that the data point isn't valuable, or that the consumer isn't actually in-market. But by the time the data point is identified, the consumer is likely already too far down the purchasing funnel.

This is a flaw in the way most digital intent and in-market data is identified and shared. Performing a search on an auto site or a travel site will denote you as in-market to the digital advertising world. However, it's more likely you were in-market before you started the search in the first place.

Most marketing revolves around intent – either creating it or identifying and targeting it. Advertisers want to drive awareness and consideration to create intent, and then convert consumers with reasons to buy. The success of this depends on how well you define your target audience. That, traditionally, starts with a lookalike model.

The in-market lookalike model for consumers looking for hybrid autos, for example, may stereotypically fit the general demographics and psychographics of someone in Portland, Ore., or anyone who lives in California.

But not every consumer in those places is in-market for a hybrid. Lookalike profiles are too broad. Looking like a purchaser and actually being a purchaser are two different audiences.

The behavioral triggers that occur before a consumer actually becomes "in-market" are located somewhere between that lookalike profile and the actual in-market profile for hybrids. These triggers denote that the person "might be considering" or is "almost in-market." The triggers can be anything, such as opening a credit card, a recent move, having a baby or a search for auto insurance. The key is not the behavioral attribute itself, but the amount of time that has passed since that trigger event happened.

Target, now infamously, did this to predict pregnancies. By studying buying patterns of a person preceding joining the baby registry, Target identified behavioral triggers, such as increased vitamin purchases or switching from scented to nonscented soaps, to predict due dates to within about a week of the actual due date.

Rather than using the lookalike profile of someone that is in-market or likely to buy, or simply targeting in-market behaviors, focus on the sequence of events, behaviors, video views, location visits and every other data point available across all channels. It's better to identify triggers that mark a user while competitors still rely on lookalikes.

A prospect is much more impressionable during this earlier stage and, in most cases, significantly less expensive to acquire.

Follow Adconion Direct (@Adconion_Direct) and AdExchanger (@adexchanger.com) on Twitter.

1 Comment

  1. Great article James. Very few data and tech companies are targeting intent at the right time because in many cases they are using the wrong data objects for their models as you suggest. At Dstillery, we like to use the term act-alike opposed to look-alike because we feel is captures the essence of how we identify intent using activities and sequences that occur across mobile, online and offline triggers. While some may argue that this is just semantics, we believe not all prospecting is created equally. The differences begin with the data used for modeling, and continue with how and when you reach your target audiences across all media devices to help marketers find those elusive new customers.

    Reply

Add a comment

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>