Behind one of the most popular new ideas people are debating right now in marketing is a shift that almost everyone is ignoring.
Augmented and artificial intelligence is the hot topic everyone wants to talk about: how marketing can use AI, algorithms and other machine learning to improve targeting, messaging and delivery. To date, the conversation has been focused almost exclusively on the supply side of marketing. But if a chief goal of marketing is to impact consumer purchase decisions, shouldn’t more attention be paid to the fact that an increasing number of these demand-side decisions are also made through augmented and artificial intelligence?
Whether it’s through explicit augmentation (like recommendation engines) or implicit augmentation (like filter bubbles and content algorithms), we make more and more of our choices about products and services with the help of machines. And it’s a trend that’s only likely to increase. Mobile penetration means that the most common decision-making unit will be person plus smartphone (or maybe that it already is). And in complex categories like financial services, agents, bots and other types of filtering and recommendation services are already shifting the majority of the thinking over to machines.
At the same time, the latest thinking in experience, loyalty and service design seeks to eliminate decision fatigue by making interactions as simple and straightforward as possible. That means a reduction in human decision-making wherever it can be part of making an experience easier – which is just about anywhere.
“Hey Siri, find me the best dog food.”
In the end, what it means is that we might be consuming more, but we’re deciding less.
Unsurprisingly, marketing literature has not kept pace. Decision-making theory is almost entirely consumed by human nature and human biases. Very little has been written about the impact of machines and artificial intelligence. (Google has, perhaps, come the closest with its ZMOT thinking. But that’s largely a model that examines when decisions are made, not how.)
So, the question of how to account for the influence of artificial intelligence in purchase decisions is murky. Technology continues to evolve and, as it does, the way we interact with it evolves on a parallel path. That means it could be a while before we can create new models that accurately depict and predict machine-aided decision-making. In the meantime, we need to look at traditional marketing models, those that are built upon a foundation of 100% human decision-making, with some healthy suspicion.
Some of these models and assumptions are easy to spot, like the the oft-quoted rule that 70% of CPG brand decisions are made at point of purchase. But others may be harder to uncover. For example, the Ehrenberg NBD model, the first successful model of repeated buyer behavior, is almost exclusively based on data from human purchases. Sure, there’s no evidence that it’s wrong (yet), but there’s also no guarantee that it’s still right.
In the future, we’re going to need a new and better way to think about decision-making and how technology is changing a space that used to be entirely human. If you’re working on it, get in touch. I want to hear about your approach.