As I sat down to write the first draft of this post, I got an email notification from Constellation Research (I know, I should be writing distraction-free) with the boldly proclaimed subject line: “AI Delivers Mass Personalization in Commerce.” I didn’t go back and try to count how many emails I’ve gotten lately headlined with something about artificial intelligence or machine learning in marketing: how they’re the future; they’re disrupting marketing as we know it; watch out marketers because AI/ML is coming. (As an aside that I’ll save for another time: artificial intelligence and machine learning are not the same thing, but this article in Wired explains it well if you’re interested in the difference right now.)
Our incomparable director of customer success and I have been firing off Slacks to each other about this topic for a couple weeks, so we thought it made sense to post about it here and get more coherent about our points and solicit more input. It all came to a head when another vendor announcing a new function:
To be clear, we are not denying that:
- Automation in all its forms is critical to and constantly changing marketing
- Marketers need to rely on technology to keep up with their consumers and their competitors
- Algorithms can add sophistication and capabilities that a human can’t
What we are saying, and saying it emphatically, is it will take months and years (trying) to get all the data, segments, processes, budgets, campaigns, measurement, and people set up to then hand it over to a machine in the hope of automating the customer experience entirely. But here’s the rub: focusing instead on just getting started while being diligent in all of those areas will give you quick wins now, measurable achievements in the short-term, and transformative milestones over the long haul.
Marketers know what questions to ask; the automation helps you get faster, better answers.
In college, I took a micro-economics class (at my mother’s insistence) during which I came to terms with an inescapable reality: in order to have a free market, you needed a market, which requires property/goods, the ownership of which must be enforced, thereby necessitating some kind of non-market intervention right off the bat. There is no such thing as a truly free market, and in the same vein, there is no such thing as a purely machine-generated customer experience. Marketers need to do the heavy lifting to figure out the right metrics, segments, and levers to pull – and the right times to pull said levers. We can’t simply create an abundance of experiences and tell the machines GO; we need to find key areas to focus on, and identify things that people respond to… and yes, automation helps us get there better and faster. Operative word being “helps.”
Machine intelligence is decades away from matching human intelligence.
Even with strong machine learning in place, having human minds applied to the problem in the right ways will only improve upon the effort. Even Facebook’s own head of artificial intelligence knows this, arguing that machines are still years away from the capacity of the human brain. Furthermore, this isn’t only a question of intelligence, but also one of emotion and reason which machines sorely lack. Take the blowback against the content recommendations on many publisher’s pages that range from misleading to offensive. Built for clicks, that’s what these solutions deliver – without accounting for cost to brand reputation, for example. To borrow from the Dow Chemical campaign, what’s missing is the human element.
Algorithms are like an engine: they run, but someone still needs to turn the ignition.
Before you counter, “But aren’t self-driving cars the way of the future?” consider this: humans built the cars, will maintain the cars, and provide critical information to the cars. Marketers still need to construct the experience, ask the key questions to ensure optimal performance against business goals, and intervene to add more data sources or channels as needed. In other words, there’s still a lot for the marketer to do that all the hyped-up discussion of a mechanized “make it so” button obscures.
We’ve seen them (maybe yours!): companies that will fail because they either 1) neglect more deliberate progress in order to dedicate all their time and resources to get to an as yet unproven and undefined place OR 2) never even get started because the scope of the task seems too enormous for their current______ (fill-in-the-blank: team, budget, project list, etc.) It doesn’t have to be that way. As Arthur Ashe said:
Start where you are.
Use what you have.
Do what you can.
This might help, too.