To Automate or Not to Automate (Everything)

A couple of years ago, I was perplexed that my son didn’t want to practice driving so he could get his driver’s permit. I pressed him and his reply was not what I expected. “Why do I need to learn to drive? Cars will do the driving for us in a couple of years.”

Kids say what kids say, but it’s a fascinating response when you think about it. His belief that driving was going to be automated took away any interest on his part of learning how to drive himself. Driving a car would be one less thing for him to ever worry about.

For shock value and context, imagine a trafficker saying they don’t need to learn Excel because Machine Learning will do all that work in a few years. I mean, in this commercial, Common isn’t selling spreadsheets for Microsoft:

Crazy right? Ad ops without Excel experience? *shutter at the thought*

More recently: I had a conversation at the Miami Publisher Forum about training. The person was trying to figure out how to balance outsourced trafficking with a training program for new hires. Outsourcing was great, but it stole an opportunity for junior people to get valuable experience.

Another conversation: another publisher frustrated with trying to get non-ops people to pull reports instead of asking operations to do it for them. The month-end would come around and ops was left to do all the work because everyone else wasn’t comfortable doing it.

Third conversation: a recent consulting engagement asked me to plot out what ad operations will look like five years from now. The expectation on the client’s part centered around automation and trying to figure out how few people they would need in the future.

Operations should always be looking to automate and no one wants to be stuck doing “manual” work (manual is in quotes because pulling reports ain’t picking fruit in the fields). However, we as an industry have not reached a point where we can leave it all to the machines. Seriously folks, the math of digital advertising is still half baked. An impression is an impression except when someone decides it’s not. Besides, we’ve talked about the limits of automation before.

That’s why you’ll see me roll my eyes at blockchain. I actually believe blockchain has huge societal impacts, but not sure why you’d start with digital advertising. If the numbers are bullshit, I’m not sure what making immutable bullshit does for anyone.

So I’m preaching patience. I’m advocating that you not lose sight of how things actually work. Keep some processes manual enough that you don’t forget how things operate. There may be additional costs with not outsourcing or automating everything, but what is gained or retained is understanding how things work. Manual processes provide a good testing ground for the next generation of ad ops leaders.

I imagine Tom Shields, who just left ad tech,  getting a call to come back and fix programmatic because no one remembers how it actually works. Let’s not make Tom a Space Cowboy:

FYI – Fast forward to the present day and my son is about to start delivering pizzas for Dominos. He isn’t driving stick, but at least he’s getting experience driving a car.