data analysis

Are We All Data Analysts Now?

Jim1There isn’t a day that goes by when one of the dozens of newsletters you get from MediaPost every day hits your inbox without a headline touting the latest data-driven mash-up, team-up, partnership, or can’t-live-without-solution that will make our marketing and advertising lives better.

They are all based in one form or another on taking myriad data from a variety of sources and normalizing them into a hoped-for cohesive whole that can ultimately be used to better deliver advertising and, thus, get digital media users to become first-time or repeat customers.

Some of these sources – in fact, more and more of them – are social in origin. By that I mean social media networks are the source and repository for these data. I’ve written in this space before how important that source can be. But I’m here now to caution against taking a data-centric only approach to determining everything there is to know about how to activate people to action, be it for something as banal as buying our stuff to as potentially culturally altering as creating stories or music. A few months ago Havas and Universal Music group announced an alliance to comb through billions of bits of behavioral data as it connects to music in order to uncover “unsung” insights into the human being. Artists can then ostensibly use the segments drawn together by this exercise to better advertise their wares; and advertisers can ostensibly use this to better associate their wares with a variety of musically driven behaviors.

Now, I think it’s important to remember… Users (a horrible pronoun) aren’t good at anything other than doing what they do. The best we can do in response is watching them. Of course, of the 6 basic emotions that motivate action, only 1-2 is affirmative (there is actual action; that which most data driven enterprises depend). This means that, basically, 75% of human response is unregistered. This means that while the amount of data we have might be enormous, it is a whole lot about only a minority of what drives human behavior.

I will, as I often do, turn people to John Searle’s Chinese room thought experiment.

One can argue that appearance of understanding is as good as understanding itself. But the questions we are left asking is:

1. Do people share real data that is analogize-able to broad swaths of behavior?
2. Is online behavior the same as “action,” I.e. Does click stream simulate walking down the street?
3. Are future preferences predictable from present consumption patterns? (I used to read Cerebus the Aardvark and listen to Tribe Called Quest; does that mean that future me will only get news from hip-hop alternative comics?
4. Do you really think that all of human behavioral motive is renderable into machine be readable form?

I want to be clear: I’m a big fan of data and make a living off of its application in a wide variety of ways.

Couple things:

1. How predictive can these systems every really be?
I’m reminded of psychohistory and the laws of mass action from Asimov’s Foundation series: it can predict the future, but only on a large scale, where a large number of people behave at the same time in similar ways but individually and without coordination. There are some markets that are both mass enough and slow moving enough for this to work some of the time, and I’d argue that this is where success is found with our big data methods. But there are plenty of black swans out there, and ample individualistic human motive (at least in the West) to produce exceptions that outnumber the rules. Hence the ultimate unpredictability of financial markets, weather, politics and sporting contests. As a contrarian, my objective is to produce consideration of otherness, not rejection of is-ness. So, I find myself frequently saying, “yes, but…”

2. What are the desired outcomes?
For marketing, it is quick, easy, low-friction growth-hacking. Though I sometimes wonder how much real growth. Sometimes it feels as though there’s just a shifting around, like the old axiom “energy is neither created nor destroyed; it is merely transferred.” Does the kind of systemization we usually are talking about when it comes to application of data produce growth that is born of something other than extracted efficiencies from process or managed risk in the form of lower price? That’s my one outstanding bete noirs: not that cost control and process efficiencies aren’t absolutely important, just that they eventually come to a dead end. Next would be identifying the existing consumer to more efficiently deliver advertising that is meant to remind that existing consumer the brand is there and instigate action. Following that is taking any learnings from that to find other act-alikes. The suspicion here is whether or not those act-alikes are really new to my brand. Could be that I am really only identifying existing customers that have not been identified before via the marketing panopticon. What impact on growth is there in absolute terms over regular sales had I maintained the status quo.?

Like I said before, don’t get me wrong. I don’t have a single client for whom I do not use programmatic everything as a major digital comms channel. But while we seem to have successfully wrestled quicker, easier, lower-friction processes, I think growth in absolute terms remains almost as uncertain as it’s always been.

For other activities the desired outcome is influence – influence on politics, on art, on media, on social activism. We’ve seen ostensible influence on politics; the Obama elections are often put forth as case studies for this. That’s the argument behind Sasha Issenberg’s The Victory Lab. On art, we could say that the outcome is something like Jonas Lund’s “The Fear Of Missing Out” installation. On media, it can be seen in Upworthy/HuffPo/BuzzFeed/ABC News/Vice…. you name it.

I’m curious to see if all this collection of data and the application of pattern recognition will produce progress as well as change, particularly in the medical/biotech field, where the outcomes are more meaningful. And I’d love to see it produce social and cultural progress in more meaningful ways. But I’m admittedly becoming cynical about progress when it comes to marketing. Can’t shake the nagging feeling that we are really just optimizing towards homogeneous mediocrity there.

Luckily, I remain persuadable!

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.