The current construct for big data’s assignment of value is assuming that what is always countable is always meaningful. This simply isn’t the case. There are ample non-affirmative causes for behavior; in fact, the majority of behavior causes are non-affirmative. I’ve said hundreds of times that the sum of human behavior is not renderable into machine-readable form.
Not to get all “Wittgenstein,” but if we think of data as constituents of language, there are some things language can’t “say;” it can instead only “show.” And it can only do either in some cases, not all cases.
Most of big data in digital advertising still relies on clicks. They are, in fact, the building blocks for most act-alike profiling. This is not to say the “click” is not worthy of attention, but it may no longer be worthy of our study.
Most value ascribed to advertising is an assigning of value after there is a determination of what’s meaningful. But the underlying presumption is that countability is a necessary condition for meaningful, and therefore valuable.
This is sensible when it comes to reading results; it does little when it comes to reading causes. It’s a big rain puddle of “what” without an umbrella of “why.”
This is why we have the kinds of proxies we do for the softer brand metrics, because we hope it will simulate enough of the “whys” to be predictive. But even these rely more often than not on affirmative rather than non-affirmative response. This ends up leading less to new opportunity of new growth through brand enhancement and product sales, and more to simply better articulating existing business opportunity. As an example of how affirmative-only data isn’t even always very good at this, either, try this experiment: what is the total number of “auto purchase intenders over the next 6 months” identified by any of the big data houses? Six months from now see how many autos were purchased.
The challenge is figuring out how to better see the shape of behavioral “dark matter” through other proxies — the non-affirmative behavior.
What advertising and marketing try to do are two things: solve puzzles and mysteries. The data helps with the puzzles. The puzzles consist of the countable, the visible, and a picture consistent with logical form, if you will. The interpretative intelligence of the humans helps to solve the mysteries. Those consist of the empty spaces, the trackless gesture, the Piero Sraffa stroking his chin with his fingertips; the meaningful but not countable.
The challenge is that big data’s robot army can only count what is countable. And what is countable is not always meaningful.
There are only 6 basic human emotions (some recent research suggests there are 4). In that recent research, only one is “positive.” This means that, if what we are counting in advertising is positive response to stimuli and optimizing on it, we only have 25% of the story. What we need to focus attention on for long-term sustainability of business objectives is optimizing against the not-click or the not-sale. As Frank Herbert once wrote in God Emperor Dune, “The empty spaces are always worthy of our study.”
The empty spaces are always worthy of our study, but the robot army can’t study, because it can only respond to the countable.
The computers will only get good at recognizing positive habits with response. The part of human behavior that cannot be rendered into machine-readable form will have to be passed over in silence.
Jim Meskauskas is a co-founder and Chief Strategic Officer of Media Darwin, a consultancy specializing in strategic planning of commercial communicative action. He’s a medialogist who has spent the last 20 years living, breathing and thinking about how to use media to move people to action. Outside of that, his likes are horror movies, Southeast Asian cuisine, his wife and his cat — not necessarily in that order. His dislikes are mean people, people who text while walking in or out of the subway entrances, pestilence, war, famine and death.