Earlier in the month, I dined under the space shuttle Endeavour with some of the best minds in marketing science.
One mind remarked:
“That’s why I bring a glossary with me, oh, you want to do supervised learning? Oh you mean regression? Oh, okay, now we can talk… We’ve been talking to managers about these methods for decades, but it’s just suddenly sexy because it’s all machine learning and deep learning and reinforcement learning.”
A lot of the math that underlies much of machine intelligence and artificial intelligence is indeed remarketed marketing science. And, hipsterism aside, the annoyance is understandable.
Marketing science started out a bit of a revolt against the Mad Men. Some of the early stories feature post-war quants going to AAA conferences and really sticking it to advertising agency presidents. And over time it has become a bit of a distinct branch of information management, operations research, and management science with a particular edge to it.
To have a bunch of valley types come up and rebrand their entire discipline as something else would be pretty annoying.
Barbarians at the gate.
The marketing science community has its own traditions, distinct language, and culture. They’re in multiple sub-disciplines within a sub-discipline within a discipline of management. Each has their own set of axioms and accepted paradigms, built up like coral over the decades. There’s a lot of competition for limited journal space and limited funds. There are cliques that are all referential in their own paradigm, and other cliques that are quite different in their outlook and perspective. They’re all tortured in degree mills.
The community is quite dynamic.
There’s a bit of a clique emerging among technology oriented marketing scientists using well worn models and remarketing them as machine learning. Some of the more progressive grandees of the industry have been encouraging of the clique. Five people showed up to a machine learning session two years ago in Baltimore. Hauser’s talk at the Machine Learning II session was taken in by an overflow crowd of 40.
Data science also has it’s unique culture. It has its tool jockeys and its partisans. What’s so hot right now is Reinforcement Learning (RL). So hot right now. RL. RL isn’t new. Faster processors are new. The big investments are new too. If a given algorithm is proven better at solving a specific problem, they’ll gravitate towards it. Who knows what good stuff some people are keeping from us?
Sub-disciplines within data science
Arguably, if we don’t count 2009 as the origin of data science, there has always been sub-disciplines within the larger family. Hipsterism aside. Some of these lines trace their origins to computer science. Computer vision, natural language processing, and insurance arbitrage specialties all have rich roots and their own disciplines. As do the recommender systems kids and the morphing kids come from a mix of computer science and marketing science. The social contagion branch traces its roots to well before Granovetter. And then there are all the direct marketing and cataloguing folks that go way back to before machine readable data. What of them? Then there are all the engineering disciplines, with their specialization, mythologies, and biases. It’s all there.
It’s always been a very huge family and very loosely coupled.
And they differ from the marketing science community in that they aren’t housed in a handful of institutions and forced into churning out degrees from buyers. From the rotting lofts on Spadina, the innovation tourist zoo’s of New York, the lecture halls of Stanford, and the back table at a café in Paris, data scientists are everywhere. Most of the time they’re cleaning data, but after the sweat comes the good stuff. And is it ever a great time to be a data scientist.
The diversity is fantastic. And the petty jealousies and squabbles generate enough heat to pop the fluffiest of popcorn.
Predicting the future
Market forces will drive more marketing scientists further into sub-disciplines. New data sources are inventing new management problems, and even more fascinating marketing science questions, which in turn will drive a better society. The next MS Conference is in Philly (I believe), and it’s predictable that there will be even more machine learning tracks and talks. There’ll be a great rebranding of existing research to be cast into machine learning papers. And this is all fine and good. A rose by any other name.
Market forces will continue to drive more data scientists into sub-disciplines. There was a great convergence in the mid-2010’s when data miners, report mill rowers, and researchers all called themselves data scientists. This can’t last. Somebody, somewhere, will invent a brand new term for the next area, likely somewhere in intelligence, and it’ll take off. Isn’t it funny how marketing is applied directly back to the marketing science of data science?
So long as economic growth continues, the number of species of this larger family will diversify. That means a lot of choice and a lot of opportunities. It has never been a better time to be interested in technology and science.
And isn’t that wonderful?