Beware the Zombie Profiles

Will data devour insight?

In 1965 Gordon Moore made a prediction about how computing power would evolve over the coming ten years. Fifty years later, he’s been mostly right and the application of Moore’s law has transformed technology.

An interesting thought experiment about Moore’s law was to imagine the automobile if it followed the same trajectory. Former Intel CEO Brian Krzanich explained that if a 1971 Volkswagen Beetle had advanced at the pace of Moore’s law over 34 years, “you would be able to go with that car 300,000 miles per hour. You would get two million miles per gallon of gas, and all that for the mere cost of four cents.”

That didn’t happen with cars, which is really a shame.

But remarkably, advances in technology brought enormous amounts of information into the corporation, including massive opportunities to better understand consumers. Whether it is big data, micro data, machine learning, or artificial intelligence, all have radically changed business intelligence and it’s so exciting.

Machine learning can help us develop customer segments but that segmentation is only meaningful if we know enough about the customer to know why they engaged in a behavior that put them in a bucket. Several years ago I conducted an ethnography for an automotive manufacturer to help them humanize an owner segmentation that machine learning had created. The people were grouped by degree of loyalty/risk and while the algorithm got it mostly right there was one segment that the machine saw as at risk of leaving the brand when they were actually immensely loyal. Conversely, there was another group the algorithm thought was loyal when in fact they were only engaging as often as they were because they were looking for deals. It turns out the latter was the group most at risk of defecting. We could only have known this by talking to them. The ethnography brought a level of human insight about their behavior which enabled the algorithm to be boosted to analyze not just what, but why.

Sophisticated machine analysis without genuine human insight risks creating zombie customer profiles — they look human, sort of, but they lack reason, we can’t communicate with them, and we definitely don’t want to have them over for dinner.

Now, about that Moore’s law car, can someone get on that?

According to IDC worldwide revenues for big data and business analytics will grow from nearly $122 billion in 2015 to $187 billion in 2019. That’s an increase of more than 50 percent over IDC’s five-year forecast period. Nearly 70% of Global Marketing Leaders in a 2017 Salesforce survey say they are using machine learning for marketing automation, social listening, and customer identity and access management. Meanwhile AI is already being used by 51% of survey participants. Clearly, data and their tools are taking over. Yet, at the same time half of digital marketers expressed “data exhaust” in a 2018 survey conducted by Digital Element.

The last point –about data exhaust– is something we need to keep an eye on. If we are not careful we risk losing human insight in our mountains of data.

Having technology that can do the grunt work is terrific. It frees us up to focus on the insight while we let machine learning and AI help to process the data. We can use it to track words, movement, and behavior. It can listen to social content, create segments and simple content, and done right, it can even present a surprisingly convincing chat-bot. But, if what we are saying or selling requires the need to know why, or taps the human drive to connect with beauty, creativity, love, humor, or joy — AI won’t be able to get us there. That requires intuition, not programing.

If you don’t believe me, have a look at these hilariously AI generated Halloween costume ideas.

Moore’s law has doubled processing power every two years for the past fifty years and that has radically transformed data which is great. Despite this stunning increase in power, the human side of intelligence continues to be where the living, breathing, inspirational insight is found. Let’s keep an eye on that lest we all end up wearing AI generating halloween costumes next year.