You already know past all argument and evidence that Big Data and Analytics is one of the great inventions in human civilization. You need only read the pop press and not a little peer review stuff to know that. But, just for the hell of it, let’s look at an example to test the proposition. Start with Facebook and the dustup over its alleged suppression of conservative news on your FB News Feed.
There was one big problem: Facebook’s trending algorithms, which identify the most-talked-about terms, were not very good at discerning what was and was not news. Left to their own devices, roughly 40 percent of what Facebook’s algorithms dug up would be junk or “noise,” a result of many people using the same word at the same time across the network. The algorithm might pick up a sharp rise in the word “Skittles” and deem it a trending topic — not exactly the events Facebook had in mind.
That is where humans came in. Facebook enlisted a set of 20-somethings as curators, copy editors and team leads, charged with sifting through the material the algorithms unearthed. They were crucial, they were told, to improving Facebook’s ability to discern, over time, what constitutes news.
Instead, these people said, Trending Topics was a fledgling, ill-managed group — made up largely of recent college graduates with little work experience — where individual judgment of news was encouraged. That led to inconsistencies in how the most popular stories were presented, along with departures from the team, eventually landing the group in the controversy that spotlights Facebook’s huge role in the type of information people see every day.
Of course, this is untrue, idiotic, and provably so. Consider an editor of a technical magazine writing about Facebook’s problem.
I was actually taken aback to learn that there even was a “team” of humans editors behind Facebook’s Trending section. It seems strange that in 2016, an age when computers are increasing their ability to parse the nuances of language at a breathtaking rate, that some Homo sapien (let alone a team of them) would still be involved with this highly regimented process.
The truth of the matter is that any system that includes humans in the mix can never be perfect. As long as humans—we petty, fallible, emotional, illogical things—are involved, there will always be room for subjectivity, error, or outright malpractice (i.e. the human element).
The good news is that humans are increasingly replaceable! In fact, the technology to replace Facebook’s human editors with cold, unshakable algorithms probably already exists. According to a post by Facebook’s VP of Search, Tom Stocky, the Trending section is currently a collaboration between humans and algorithms.
So, once again you can slumber in bliss knowing that Big Data and Analytics works so well, you don’t need humans, even Facebook humans, to locate and deliver objective truth to you at the press of a button. Or not.
You cannot square the PC Magazine editor’s claims about “cold, unshakable algorithms” against the Times quotes about Facebook trying to write those algorithms but delivering Noise instead of Truth about 40% of the time. Their algorithms still can’t understand language as well as a child. You’ve got to accept one assertion or the other.
As we’ve seen in the Big Data Case Study, while the play involves numbers and counting, it is still more persuasion than science. Humans write the code that collects, analyzes, and acts with and upon data. You can solve this problem with news as an example the way Google does. It first defines a limited set of websites as legitimate news sources, then aggregates and delivers them to you with its secret formula. At first take that sounds like the PC Magazine’s cold, unshakable algorithm, but think again. Google’s code selects only from a pre-defined list of information and that information has already been created, edited, and published by humans, not algorithms.
Defining Big Data is the heart of the persuasion play. For example, the US Census has been collecting and analyzing data on every American since 1790 every 10 years. Why was that not Big Data and Analytics?
Because it didn’t involve computer networks with apps ‘n iGizmos. Big Data became Big Data only as a persuasion play to convince Other Guys whether as consumers or customers that sophistical statisticians had invented something new that could discern important TACTs in gouts of digital bits and bytes. Anyone can make that claim, but providing good evidence in support is difficult to find. Just think about this silly Facebook anti-conservative news bias story. How is it possible for the greatest communication tool in human history, built by the smartest Big Data coders ever, to fail at something as simple as delivering the news to your digital door?
The reason is persuasion obvious: the Facebook guys are lousy statisticians and worse AI theorists. They may know how to write code for an interface, but they know nothing about statistical analysis or the vast field of study loosely called Artificial Intelligence. But, they do know how to sell themselves as Big Data guys who know a lot about Big Data. They also are looking more and more like digital Other Guys who think they know a lot about persuasion.
I’ll now pivot on that point to observe what appears to be a tangential argument: Ad sales. The ad industry annually holds what is known as the Upfronts which is a meeting between media content companies, both traditional and digital, and ad agencies. The media companies pitch their upcoming season and the ad companies make early buys for time and space. Get this from Les Moonves, the president of CBS, your grandfather’s Oldsmobile ad platform.
“After years of predictions that digital would render traditional media obsolete, that is obviously not happening,” said CBS Corp. CEO Leslie Moonves during his network’s upfront presentation, pointing out that digital advertising often lacks accuracy and credibility. “There is a clear shift in advertising back to network television.” He was a bit more pointed earlier in the day, saying “the bloom is off the rose.”
Other TV lamers noted this about digital platforms for advertising.
At Fox’s upfront event, sales chief Toby Byrne called digital content “subprime” video. “The digital metrics game is rigged,” he said.
Byrne’s shot across the digital bow was meant to shine a light on a bigger issue: As digital companies vie for a greater slice of the $17 billion video advertising market, a lack of true apples-to-apples measurements forces them to game the system.
“What they’re saying at the NewFronts is true,” said TNT and TBS president Kevin Reilly. “But having looked under the hood of a lot of these businesses, they get to a certain point and they realize, ‘We’re stuck.’” While ad dollars will continue to drift toward digital—YouTube’s big boast during its Brandcast was the $250 million that Magna Global shifted from TV to digital (though still a fraction of its overall spend)—there has been some buyer’s remorse.
“A lot of clients have found they moved too quick, and they didn’t get the results,” Reilly said.
Even a former YouTuber who’s now a TV star, Adam Conover, host of truTV’s Adam Ruins Everything, took shots at digital. A self-described “former foot soldier in the digital revolution,” Conover didn’t mince words when talking about the move from digital to linear.
“Maybe the power of TV is why a lot of digital-only powerhouses are starting to pivot towards higher quality linear video, more like … television,” he said at Turner’s upfront event.
“As a former millennial YouTube star, let me say publicly for the record, ‘Thank God I am on truTV now,’” he told buyers at the packed Theater at Madison Square Garden. “It’s so much better.”
In all those attacks, you may have missed the Big Data indictment. I’ll requote the key line buried in the middle of that long series of quotes.
As digital companies vie for a greater slice of the $17 billion video advertising market, a lack of true apples-to-apples measurements forces them to game the system.
We’ve observed this in the Facebook Case Study and I think it generalizes to all digital platforms. When you try to count the persuasion change the old fashioned way – did the Other Guy change? – the digital Big Data world cannot give a straight answer that opens the Windowpane. Worse, when you do count the change the old fashioned way, the Windowpanes don’t open; you get effect sizes that even Tooth Fairies find amusing.
Facebook is worth a $300 billion market capitalization because it is a monopoly. Through monopoly plays (buying or killing competition) Facebook built a network of 1.6 billion tin cans and wires that provide for the easiest and least important relational communication available in human history. Other Guys love that simple and easy relational sharing. And, they will continue to do so as Peace and Prosperity permit more free time for more Other Guys to kill with pastimes like Facebook. But Facebook as news source or Facebook as persuasion instrument – fuggetaboutit!
Let’s pull all these strands together in summary. Facebook is a digital water cooler that allows Other Guys to give and take information within loose social groups. Facebook sells itself as an extraordinary communication form that revolutionizes persuasion, information acquisition, and relationships when it only allows Other Guys to more easily chit-chat about the eternal to-doing list in the mess of life.
While Facebook engineers are great at building that interface and maintaining a database of information, they are proven failures at any other kind of analytics or communication. At best, they merely provide another source of marketing information about Other Guys and even that unique feature requires support from your grandfather’s marketing Oldsmobile. Past the digital water cooler, Facebook cannot hit its butt with either or both hands and must resort to persuasion plays to build a curtain around their Wizard of Oz.
But, as the recent ad sales season demonstrates, the Children of the Night with ad money are waking up to that persuasion and beginning to see behind the curtain for themselves. And you know that because those money guys are counting the change with the TV guys, not the digital guys. Sure, they can run interesting combinations between the various platforms, but the money guys know that if they want to hit the TACT, the hammer of choice isn’t Facebook or Twitter or Google, but the real alphabet companies like ABC, CBS, and NBC.