Should Marketers "Go All In" on Algorithmic Marketing?

One of the most profound changes in marketing over the past two decades has been the explosive growth in the development and use of marketing technology. Today, technology touches virtually every aspect of marketing, and our use of marketing technologies continues to grow.

Big data, predictive analytics, and "data-driven marketing" are now among the hottest topics in marketing circles, and most thought leaders say that companies are only beginning to scratch the surface when it comes to using data and analytics to automate marketing. Some thought leaders envision a not-too-distant future where computer algorithms direct many of the interactions between companies and their customers or prospects without human intervention.

For example, in The Marketing Performance Blueprint, Paul Roetzer, the founder and CEO of PR 20/20, writes:

"Imagine an algorithm-based recommendation engine for all major marketing activities and strategies. The engine will use a potent mix of historical performance data, industry and company benchmarks, real-time analytics, and subjective human inputs, layered against business and campaign goals, to recommend actions with the greatest probabilities of success. If built or acquired by marketing technology heavyweights, these tools will add algorithmic marketing strategy to the automation mix."

In an article for Marketing Land, Mr. Roetzer expanded on his view of the future:

"Natural language processing, hypothesis generation and dynamic learning are core components of the technology that will transform the marketing industry. Rather than simply automating manual tasks, artificial intelligence adds a cognitive layer that infinitely expands marketers' ability to process data, identify patterns, and build intelligent strategies and content faster, cheaper and more effectively than humans."

The support for data-driven marketing is strong and growing, and many enterprises are already using data and predictive analytics to automate some interactions with customers or prospects. I contend that marketers should approach automated algorithmic marketing with a healthy dose of caution and make sure they fully understand its limitations, as well as its potential benefits.

One important limitation is that algorithmic marketing relies mostly on behavioral data. Virtually all of the data about customers and prospects that falls under the rubric of "big data" is data describing behaviors and actions - the digital footprints that we leave behind as we use digital devices and channels to consume or exchange information. As I wrote in an earlier post, the problem with behavioral data is that it can tell us what someone has done (and often when and where he or she did it), but behavioral data alone doesn't tell us why someone took a particular action or behaved in a particular way. In many cases, behavioral data reveals little about customer attitudes and motivations, and these factors play a huge role in successful marketing.

We also need to be cautious about algorithmic marketing because it can make us overconfident. The vast amount of data that we can now access and analyze, and the growing power and sophistication of predictive analytics software can easily lead us to think that algorithmic marketing is more effective and reliable than it actually is. In reality, algorithmic marketing makes us susceptible to a version of the McNamara Fallacy.

The McNamara Fallacy was named for Robert McNamara, the US Secretary of Defense during the early stages of the Vietnam War, and it relates to his approach to managing the war effort. The term was coined by the noted social scientist Daniel Yankelovich, who expressed it in the following terms:

"The first step is to measure whatever can easily be measured. This is OK as far as it goes. The second step is to disregard that which can't be easily measured or to give it an arbitrary quantitative value. This is artificial and misleading. The third step is to presume that what can't be measured easily really isn't important. This is blindness. The fourth step is to say that what can't be easily measured really doesn't exist. This is suicide."

Like all humans, we marketers have a strong tendency to base our decisions on the evidence that's easily available to us, and we tend to ignore the issue of what evidence may be missing. Psychologist Daniel Kahneman has a great way to describe this powerful human tendency. He uses the acronym WYSIATI, which stands for what you see is all there is. My point here is that it can become easy for us to believe that the data we can track, collect, and analyze is the only thing that matters, and therefore that predictions and recommendations based on that data are inevitably correct. But, it's just not that simple or straight forward.

I'm not arguing that marketers should ignore or avoid using big data, predictive analytics, and algorithmic marketing. These tools can be immensely powerful,so the key is to use them wisely and to remember that they're neither complete nor perfect.

Image courtesy of Daniel Morrison via Flickr CC.