Data remains our industry’s greatest challenge and opportunity. Even as marketers and agencies get comfortable with the necessity of data-driven marketing, there are lingering issues tied to our
“bigness” obsession. On the ad tech supply side, we must continually decide how to solve large data engineering problems. At the same time, marketers must embrace the possibility of rich
data without fixating on its complexity, obsessing over its bigness, or wandering into common mistakes that can steer one astray. For all of us, it comes down to simplifying the way we regard data in
the first place and how we gauge consumer intent.
Understanding the Difference between Passive and Active Signals
In an industry that still touts
“behavioral targeting” as the ultimate data-driven marketing methodology, consider this: is the behavior you are being sold any more than a page being loaded (read) vs. unloaded? So-called
telling “behavior” may be nothing more than a loaded URL.
Read is a classic behavioral signal, but it’s problematic, because it’s passive. For
example, when a browser loads a Motor Trend page, we may presume the user is an auto intender because they must have read that page. The issue with “read,” is that we don’t know why
the user loaded that page, or even if it was intentional. Everyone will at one point execute a passive signal. That doesn’t mean they are in market. Cats jump on keyboards all the time and
therefore may represent a substantial portion of false data.
As an increasingly available alternative, active or “said” signals are more interesting and indicative.
They reflect a demonstrative intent by the consumer toward a particular avenue of interest. Consider a consumer adding a product to a wish list. Here, we aren’t guessing the consumer wants the
product; they’ve clearly said so.
Marketers seeking data partners often get tricked by the volume promise. But I’ll suggest that generating a large volume of data is
no trick. Ask any provider in our realm, and they will tell you: There is a ton of data out there and endless segments to be modeled.
That “tonnage” proposition alone
unfortunately tends to sell itself. It woos us. Any provider can sell you the concept of “reaching 300 million uniques,” while dodging the question of how that can even be possible, given
the population at hand. Think about auto intenders, for example. How many cars are sold per year in the United States? Ten million to 20 million? That’s the number of auto intenders. Why?
Because their intentions were indicated by crystal-clear commercial signals -- not just browsing or reading behavior.
We also tend to forget that your true audience reflects those
likely to be persuaded by your creative. A common mistake in our industry is to become enamored by very large audience segments as a sign of scale. The truth is, the larger the audience, the less
selective it is. The less selective it is, the less persuasive your media will be to the audience. The less persuasive, the less effective and more expensive. In the worst, very common case, you are
winning the impressions on the worst users (as you bid on all of them) and thus having the least impact.
Embrace the Power of Commercial Signals
Surprisingly, there is still a lot of demographics-based advertising. When you lack commercial signals, demographics are better than nothing, but this data is coincidental, not
causal in nature.
Is it better to market a Toyota to people saying they want to buy a new car this year, or those who are 18-35 and male? You tell me. To the point on
selectivity, one of those groups is much more selective than the other. In reality, there are plenty of car sales to women, as well as many to people older than 35.
It will always
be better to market with the more selective and intent-based information -- to which we now have access -- than with coincidental data. That approach is your smart alternative to super-sized
segments. Active commercial signals will be the best yet, because they require no guessing at all.