If you’ve ever bothered to ask your sales rep, or if you’ve attended any of my digital marketing training in the past 5 years, you are aware that similarly-named behavioral data definitions vary wildly (to say the least).
If you’ve never bothered to ask or even wondered about the definition of any behavioral data set – WAKE UP – please.
What do I mean by “definition”?
Behavioral definitions have three components: a qualifying component, a maintenance component and a removal component.
The qualifying component
What are the qualifications a user must demonstrate to be tagged with a particular behavior? Must they visit a page or section a certain number of times per period of time over a longer period of time of time? Must they be exposed to a particularly tagged content a particular number of times over a period of time? Must they have “liked” or “commented” a specific content in the past X number of days? These are all types of definitions what would qualify a user to be tagged with a specific behavior.
The maintenance component
How long is the user tagged with this behavior? A required part of a behavior’s definition is that the user maintains “said” behavior over time, or else it might no longer be relevant. The maintenance component takes into account that users who have particular passions or interest may at times be distracted from it and so lessen their behavior, but it would still occur at a particular pace over time to remain relevant.
The removal component
If the user stops behaving in a way that qualifyied or maintained their behavioral label, their tag should be removed as they are no longer relevant. This could be because they’re off on vacation or because they’ve deleted their cookies or started using another browser – it doesn’t matter. If the user doesn’t maintain their behavior, an advertiser doesn’t want them in the targeting mix.
No two behavioral data set is alike
Take for example the behavior “automobile”. If you don’t ask what the definition is for this behavior, one can assume it might mean interest in cars, or a passion about cars, maybe even intent to buy a car. However, when you ask what the definition behind the label of that behavior is, it can only mean the lowest or vaguest possible assumption.
If we keep the automobile behavior, one publisher my define that as visiting the cars section twice a week for a minimum of two weeks, and keep returning at least once every other week. Should two weeks go by with no visit, then the tag is removed from the user.
Ask any two or more publishers or ad networks how they define automobile and you will find two or more very different definitions for that same behavior, you might have assumed meant one thing before asking.
How else could a publisher identify an “automobile” behavior? One publisher might be more stringent and require three visits per week for three weeks – similar but not the same. Another might have a completely different approach to it by identifying people who’ve been exposed to 20 pieces of content with the tag “automobile” over a period of a month.
At first glance, any definition will make sense, but I encourage you, if you are the buyer and you value your investment, to play devil’s advocate with any definition you are presented with.
With the first definition (twice a week for 2 weeks in the car section) the user could be an auto enthusiast, always wanting to learn more about cars, but that doesn’t mean they even own a driver’s licence or that they have the age or available credit to even buy a car. It could also mean someone currently shopping, researching reviews and articles to figure out what to buy.
With the last definition (20 pieces of tagged content in a month), look to what type of content is available on that site. While much “auto” related content might be what you see value it, they may also report the news which could include “auto” tagged articles about a fender-bender involving “cars” or a manufacturer laying off or hiring. They could be about a dealer sponsoring an event or their latest stock market results…
This of course will vary from definition to definition, from site to site and from interpreter to interpreter. Know that not all sales reps know what their definitions are, nor are they all allowed to know or share what it is (for some confidentiality reason).
I think it is time to establish some standards the industry as a whole can agree with.
Because of this massive mix-match of valuable and valueless behavioral data definitions, I suggest that the ongoing 3MS initiative add standardizing behavioral data definitions to their list of upcoming projects to render the industry simpler and more valuable for all involved.