Not All Device Graphs Are Created Equal
I recently wrote a blog post about the importance of having a device graph that is addressable
and hinted that I had more to say about a more complicated, “nested” addressable device graph design that could have big implications for marketers.
Marketers want flexibility. They want to target based on digital data (mobile location, website visitation, retargeting, app usage) and offline datasets (mouthwash buyers, people who frequently eat at family restaurants, high-income households). They also want to target both on a household basis (all members of a household in-market for buying a new vehicle) as well as on an individual basis (women 25-54 who live in households that buy lots of laundry detergent). Thus, the ideal device graph is not only addressable but it is also nested.
Nesting the Device Graph
By nested, I mean that one starts with a household and, within a household, there are people and belonging to those people, there are devices. This allows for a device to have both household level attributes as well as important individual attributes.
In the example below, the household on the left has three people/personas, two adults 45-54 (a male and a female) and a young adult female, 18-24. The two 45-54 year olds both have a mobile device and the 18-24 year-old female has both a mobile device and a desktop personal device. In addition, this household has a desktop device which appears to be shared. The second household has one senior female who has both a tablet and a desktop device.
This type of nested addressable device graph that uses both the household, the individual and the associated devices enables marketers to target advertising based on offline household level attributes, as well as targeting based on person-level behaviors and also communicate in device-specific ways.
I believe this is the ideal graph for marketers, but how should one measure the value of the graph? What are the right metrics? When marketers target advertising based on household level characteristics such as pet ownership, or heavy purchaser of dish soap, the best way to measure the accuracy of the targeting, is by measuring the sales lift driven by exposure to the campaign versus a scientifically chosen control group. 4INFO has enabled over 200 such measurement studies. The best way to evaluate the accuracy of person-level attributes such as age and gender, at a device level, is by utilizing 3 rd party solutions that report audience characteristics.
Thus, it’s important to ask hard questions when choosing your identity and cross-screen device solutions.
First, ask yourself if you have interest in targeting based on offline characteristics. It’s often said that the best predictor of future purchase is prior purchase of either the advertised item or items in the category.
If you desire such targeting and you want to execute multi-screen campaigns, you need to work with an addressable device graph provider who can onboard these important offline segments. Also, do you want to measure sales lift? If so, this is also a case in which you’ll need an addressable device graph.
No matter the desired marketing goal, as users increasingly access media across a range of devices, zeroing in on how much credit to give one device versus another under one roof has grown in importance.
With marketers insisting on increased targeting flexibility, a nested addressable device graph, at a minimum, enables a sharper view of your targeted market and a more logical linkage between the household you intend to reach, the users residing within and the attributes of each user’s respective devices.