Is It Time to Ditch Demographics?

When it came time for my responsibility to participate in the United States 2020 Census, I was amazed by how limited the questions were, and worse, how little the government was interested in anything about my family beyond our race and how many children and elderly we have living with us (our son is sure to tell you that "my Daddy is old" in case they asked him). After all, we live in the age of #bigdata, where software and machines can make sense of all kinds of data and information that humans alone cannot. I know the federal government wants to measure certain trends and aspects of the population, often on either precedent or for good reason.

My wonder on profiling audiences is more finite than the general US population, though. I'm in the business of helping brands understand their customers, both known and prospective, beyond email opens, ad clicks, social likes and shares, and purchase behavior. Working with clients that are either products or services and others that are agencies, I'm concerned about how our media networks - broadcast, audio, out-of-home, digital - still call for demographics in planning and serving advertisements.

I've long believed that as the world is ubiquitously digitized (by smartphones, at least), marketers need to focus on customer activity and behavior rather than segmenting audiences, creating messaging, and buying media based on demographics. I've learned that there's nothing homogenous about "non-Hispanic white males over 40," and I would wager the same rings true with other vertical demographic views.

Attempts to define personas and customer profiles based on age, ethnicity, and gender are not always necessary (exceptions lie in healthcare, movie tickets, alcohol, and other verticals, sure); and, because media networks and DSPs are dialed up by such coordinates, I hold the use of demographics in communications greatly responsible for much of today's societal ills and disparities, with misinformation and bad data. I'm digging deeper to prove that I'm right with that view, and I'll post what I find here, or here.

This brings us back to big data and machine intelligence.

It's been said that it is "garbage in and garbage out" when it comes to artificial intelligence and machine learning. This is especially true when it comes to customer data and demographic profiling -- and, because agencies, brands, and DSPs have been so reckless with audience segmentation based on what we look like, our age, and worse, the Federal Trade Commission is now looking to leverage old controls and new policing to ensure #AI doesn't make matters worse.

If you are either building or selling software, or your company is investing in automation, AI, and all the wonderful possibilities such math may enable, bravo!

Still, it is past time to closely focus on how we may and/or how we should not leverage demographics and the decisions we teach our machines to make for us.

What are you doing to ensure that you and your algorithmic wares are fair to you, and your customers?