Data-driven marketing requires a lot of data for optimal results, so much so that we often overlook how much. Artificial intelligence (AI), or what marketers sometimes mistakenly refer to as machine learning, requires orders of magnitude more.
AI is beyond machine learning and potentially more useful to marketers. Bernard Marr, writing in Forbes, explained the distinction quite succinctly: AI is the broader concept of devices executing tasks in a “smart” way. Machine learning, on the other hand, is based on the idea that machines will learn on their own given access to data.
Put another way, AI is centered on devices designed to act intelligently. Machine learning is about complex calculations leveraging huge volumes of data. Yes, machine learning can unearth changes in customer behavior, allowing marketers to adjust in real time but that’s only the initial steps.
What if we reimagined marketing science to test messages, channels and timing strategies and tactics on AI machinery that portray the profiles, interests, buying habits and changing preferences of our clients partners’ customers? Would we trust it as much as we trust testing on people? Should we trust devices that act intelligently?
Powerful technology companies that spend millions on brand and message testing, such as Amazon, Facebook, Google and Microsoft are as much in the business (or more) of collecting customer data than in selling their products. In some cases, they may think of their products and services as vehicles to collect data for marketing development as well as building beneficial tools.
Among marketers machine learning has become a popular buzzword because it enables teams to incorporate new data quickly and help businesses to speedily react or create processes on the fly, far faster than people can.
In the long view, however, data is embryonic, a step but not the end goal. Data that feeds machine learning is like coal was in the Industrial Revolution era, wrote Neil Lawrence, a professor of machine learning at the University of Sheffield and part of Amazon’s AI team, in a recent article in the Verge. While the technology giants mentioned above may have data in abundance, their machine learning systems are largely inefficient, he says.
The problem isn’t the volume of suitable data but instead about making our deep learning systems more efficient and able to work with less data with more insightful and applicable results. That’s why AI is the next frontier for marketers — the ability to use sparing amounts of data for more precise testing, conclusions and results.
By Curtis Thornhill