Stop creating self-fulfilling prophecies: How to apply AI to small data problems

December 24, 2021.

d3sign (opens in a new window) is an image.

The digital revolution has given us a lot of data. This is exciting for a number of reasons, but mostly in terms of how artificial intelligence will be able to further change the enterprise.

In the world of B2B, the number of transactions is vastly lower than in B2C, which is why we are still experiencing a shortage of data. In order to deliver on its promise of revolutionizing the enterprise, it must be able to solve these small data problems as well. Thankfully, it can.

The problem is that many data scientists turn to bad practices, creating self-fulfilling prophecies, which reduces the effectiveness of artificial intelligence in small data scenarios.

Data science practices and avoiding bad ones are the keys to applying artificial intelligence correctly to small data problems.

In the world of data science, self-fulfilling prophecy is simply described as "predicting the obvious." Sometimes companies find a model that predicts what already works for them and apply it to different scenarios.

A retail company can determine that people who fill their cart online are more likely to purchase than people who don't, so they market to that group. They are predicting the obvious.

They should use models that help convert first-time buyers who don't already have items in their cart. This retail company will be much more likely to impact sales and acquire new customers if they solve for the latter.

The process you should follow for applying artificial intelligence to small data problems is below.

When you don't have a lot of data to work with, the first thing you should do is enrich it. External data can be tapped into to apply look-alike modeling. Thanks to the rise of recommendation systems, we see this more than ever. Even if you only have a few purchases on Amazon, they can make predictions on your next purchase based on the information they have about the products and people who buy them. If you are a B2B company that uses a single dimensions to categorize your deals, follow the example of Pandora and divide each customer by the most detailed degrees. The richer your data gets, the more you know about it. Powerful prediction and recommendation models can help you go from low-dimensional data to high-dimensional knowledge.

The process you should follow for applying artificial intelligence to small data problems is below.

I reached back out to 10 marketers we have met through our Experts program and asked them to share their predictions for the new year and reflect on some of the trends we have seen.

Socio economic status must be part of the diversity conversation.

We made it together. It's another year in the books. Was it normal for us? No. For many, it was a step in the right direction. Maybe two steps in that direction.

Microsoft is the latest company to join the list. The software giant just became the latest big name to announce that it will no longer be around.

As we get closer to the end of the year, we are running more stories that look back at topics we examined in depth over the last several months, and several that offer well-informed predictions for...

You have to help create the environment that you want to thrive in.

In a proposed class-action lawsuit, Candie Frazier said that she has screened videos showing violence, school shootings, fatal falls and even cannibalism.

We unpack the numbers and nuances behind the headlines on Equity, a show about the business of startups. It is just about Christmas, so you may be stuck.

Osayi Izedonmwen took leave from his company to explore an idea that he had toyed with for a long time.

The relationship between the tech sector and global politics is examined in the project.

The NHTSA is investigating in-dash gaming while cars are moving. The Guardian has learned that an update will disabling passenger play.

If a fashion brand gets that demand wrong, they have to deal with the waste, because of the unpredictable nature of the industry.

Over 2200 companies are confirmed to participate in person at the Consumer Electronics Show in Las Vegas. Our focus is onconv...

Over 2200 companies are confirmed to participate in person at the Consumer Electronics Show in Las Vegas. Our focus is onconv...

The magic black box on the table sometimes doesn't track my hands. My three-year-old doesn't seem to mind at all. He is laughing so hard.

What is the meaning of 2022, for marketers?

Capital is a commodity, the platform is the most important, but working with the right people may have the most profound impact.