Is It Fair to Claim Regression As "AI"?

← Back to all posts

Is "AI" as it relates to marketing analytics really something new, or are people merely applying a new spin to an old concept? Descriptions like " AI can help marketers make data-driven decisions and optimize their campaigns..." are common. How is that any different from how database marketing practitioners have described their work over the past, say, 25 years? Or longer? It isn't really different, and it is difficult to find anyone that accurately describes just what makes AI unique in the area of marketing analytics.

A quick search uncovers several sources that lay out what are considered to be examples of how AI is implemented in the area of marketing analytics, and one site I found actually provided a handy list of "applications of AI in marketing analytics." What was on the list?

1. Predictive Analytics
2. Customer Segmentation
3. Personalization
4. Campaign Optimization
5. Sentiment Analysis

Other lists include things like forecasting and recommendations.

That suggests you need to go deeper to find the difference between AI and advanced statistical techniques, and doing so finds that many AI proponents are claiming predictive techniques including regression, both linear and logistic, random forest, neural networks, deep learning and more, as well as classification techniques like decision trees and cluster/segmentation, now qualify as examples of AI analyses.

Is it fair to lump these all under the AI umbrella? These applications and analytic techniques all existed long before the current burst of AI into the collective consciousness. In fact, this is very similar to the list of capabilities that my previous company (and many others) offered way back in the dark days 15-ish years ago or earlier! (Maybe not sentiment analysis, although that became a much more common topic when social media companies like Crimson Hexagon and Radian6 came to prominence in the early 2010s.) We never thought of describing what we were doing as "artificial intelligence" but given that the general definition of AI is along the lines of "the ability of a computer to do tasks that are usually done by humans because they require human intelligence" maybe we would have been ahead of our time!

It seems there is still some work to do if we're to define what exactly makes "AI in marketing analytics" different than "marketing analytics". In my mind, right now, the benefits of AI start with improving the data we are inputting to analyses, especially in the areas of process automation and data prep, including hygiene and transformations. And you'll notice I didn't even touch on generative AI, that is a topic for another time...

View original on LinkedIn