A priest, a rabbi and a data scientist walk into a bar.
The bartender says, "What is this, some kind of a joke?"
The data scientist says, "Why the long face?"
The bartender says, "Because I'm a horse!"
(Wait, did I not mention the bartender is a horse? The bartender is a horse.)
I forget the rest of the joke -- I'm not very good at telling jokes -- but I'm reminded of it because of a recent post at AdAge.com titled "It's Time to Bring Data Scientists on Sales Calls," by Christopher Hansen, president of Netmining.
"Online media companies," Hansen wrote, "historically sold networks and sites the same way someone sells watches: silver, gold, big face or small. But now, when display advertising is becoming commoditized and algorithms are potentially the only differentiation, the tech world needs a data-versed seller who can get down in the weeds while making it all seem not so scary."
I liked everything he wrote up until that "not so scary" part. Personally, I'd like to see data-related meetings of the minds get more scary. Let's put the fear of God into some of the rising purveyors of media out there -- particularly social-media platforms.
I want to see not just media sellers, but media buyers, bring their data scientists to sales calls. Heck, why not just lock them in a conference room together and see who comes out alive? That can be the algorithm for the new new programmatic buy: May the best math win.
Since I first started toiling in the web trenches in the late-'90s, the false promise of digital has always been measurability and accountability. I still remember the first time I was confronted with a yawning gap between the internal server numbers I was seeing on a website I built and the traffic an external measurement firm said I was getting. I thought my head was going to explode.
At least, we could reason, we were all more or less all on the same (flawed) page.
But then came the era of social media, with a whole new set of making-it-up-as-we-go-along standards regarding metrics. Whereas the old Nielsen-esque mentality applied to the early web was "This many people are seeing my content," the new discourse became about things like sheer scale (ubiquity, à la Facebook, as a self-evident measure of success), "engagement," "shares," "social reach" and so on. A new sort of mushiness began to prevail.
Even worse, there's this tendency of everybody more or less accepting self-reported claims from social-media companies. Especially the cheerleading tech press, which seems all too willing to accept that 1+1 = 3 (or, hell, even 11) when it comes to new platforms.
For instance, we've all pretty much accepted the MAU -- monthly active user -- as a meaningful metric. Facebook, for one, self-reports 1.19 billion MAUs. I don't know about you, but these days I'm using Facebook less than ever (and I was never a heavy user). Even if I spend just five minutes on Facebook every few weeks (which I likely do simply because Facebook stalks me so with emails subject-lined "Simon, you have 3 photo tags, 114 messages and 6 event invites"), I count as a MAU. (Meanwhile, as my colleague Cotton Delo reported in December, "Facebook Admits Organic Reach Is Falling Short, Urges Marketers to Buy Ads." Um, yeah.)
As I've noted before in this column, last year Twitter bought and absorbed two formerly independent social-media measurement firms, Trendrr and Bluefin Labs (both supplied data to Ad Age via editorial partnerships I set up), so we won't be seeing those companies comparing social-media platforms anymore (publicly, at least) -- which is, frankly, depressing.
Meanwhile, one detail buried in the war between hackers and white-hot disappearing-messaging app Snapchat over the past couple of weeks has been about basic math. If you haven't been following the story closely, just in time for Christmas a company called Gibson Security declared that, thanks to a security flaw, Snapchat was putting its users at risk for having their phone numbers publicly exposed and associated with their account names. And then, sure enough, a week later 4.6 million Snapchat user names and phone numbers were leaked online.
The mess not only shined light on Snapchat's lax security, but called into question claims about its size. As Violet Blue of ZDnet's Zero Day security blog wrote, "It now appears that anyone who reverse-engineered Snapchat's API" -- as was done by hackers in the process of exposing this exploit -- "could have written a script to register false accounts in the tens, if not hundreds of thousands. It's impossible to know what percentage of Snapchat's accounts are valid."
In that piece, it's worth noting, Blue pointedly referred to Snapchat's "alleged 8 million users" (emphasis mine).
So, yeah, as long as everybody's going to be bringing their data scientist to sales calls (and VC pitches), let's also invite a hacker or two.
And, for good measure, a priest and a rabbi. Might as well, right?
Simon Dumenco is the "Media Guy" columnist for Advertising Age. You can follow him on Twitter @simondumenco.
As the role of programmatic buying and selling in digital advertising continues to grow, issues surrounding viewability and verification are moving to the forefront. This white paper looks at the current state of and future prospects for programmatic in a digital ad industry increasingly defined by viewability and verification. Brought to you by RhythmOne.Learn more