Buddhist Marketer

Let’s predictively model the next article that will annoy me

I often forget that the decade I’ve spent spelunking in the bowels of online marketing has resulted in quite a bit of knowledge regarding how the industry works and how the internet works. I’m eyeball-deep in the stuff, and I forget that just because other folks are surfing the web all day doesn’t mean they have the kind of wisdom that I do about what’s going on behind the scenes.

It’s jarring to hear someone ask, “What is a link?”

I get confused when I overhear a conversation about logo size vis a vis ad space real estate. This is a discussion point, I think? People don’t just know that you run an ad flush with your logo so that your brand doesn’t suck up monetizable pixels?

And, equally unfairly, I get frustrated when I read articles that I perceive as dunderheaded in publications that are perceived as respectable. Take “Technology Gets Personal” as published recently in The Wall Street Journal:

Marketers have been doing a low-tech version of [computer-aided personalization] for years, of course, using focus groups and surveys to extrapolate general appetites. But they’ve never tailored their efforts so specifically to the individual, and, more important, they’ve never used such sophisticated methods. The new personalization isn’t just focus groups on steroids — it’s focus groups at math camp, employing esoteric stuff like regression trees and support vector machines to figure out what we want before we want it.

Let’s set aside the fact that the author starts his article with examples of Netflix and Amazon, a.k.a. marketers, thus already disproving his point. But clearly he knows nussing about online advertising and behavioral targeting. Hey, if I want to serve an ad for some fruity-rooty vegan smoothie supplement to the people who have:

• Visited nikeplus
• Visited such-and-such websites that focus on vegetarian recipes / eating / etc.
• Live in the states of Colorado, Wisconsin, Connecticut, New Hampshire and Vermont

I can do that. Sure, maybe it’s not ultra-sophisticated maff, but it’s not focus group either. Today’s marketer is increasingly a sophisticated wielder of rows and rows of data, mightily parsing individually exhibited behaviors to identify new triggers that will more effectively push product.

So, okay, this guy clearly knows jack about today’s marketer. Fine. But the rest of article simply underscores lazy writing, which is celebrated on blogs, but grating on The Wall Street Journal:

Can technology improve to the point that listening to computer recommendations instead of to friends and critics will make us, on the whole, happier with our choices?

Gawd. He acts like these algorithms are based on ether inputs. What does he think all these predictive models are based on? Prostituting capuchins in a Yale Med School lab? No, dodo; the technology is built off of the recommendations and ratings of your existing humans. Likely many recommendation engines do some kind of weighting, like 20% hoi polloi, 30% critics / experts, 50% your friends.

I recognize I am being kind of snobby here. I will annoy myself just a little more with one last blockquote:

It may be in this combination of technological prowess and human intuition that the answer — and future — lies… Computer programs are indeed powerful, but when it comes to taste, they need to pair up with a kind of human wisdom if they hope to be successful.

Um, I’m sorry, this is somehow news? Got news for ya, buddy: this is already happening.

The future is already here, but clearly this portion of it hasn’t been evenly distributed to the article writer’s brain.

/me-being-a-punk