Sorry, we could not find the combination you entered »
Please enter your email and we will send you an email where you can pick a new password.
Reset password:


Plus Report - By Thomas Baekdal - April 2019

Machine Learning is like black-magic for publishers

I recently published a Plus article about the importance of measuring time as publishers. We looked at the difference between viewing, watch, activity, and dwell times, as well as the problem with data-overload, and how using machine learning could help fix that.

But the problem with machine learning is that we often don't understand it. For example, machine learning is very good at identifying things like churn, but it doesn't explain it. We get this big list of people that the algorithm thinks won't renew their subscription, but it doesn't tell us why.

So, in my last article, I said that we need to improve machine learning to get a result, and to have it explained to us humans.

I quoted a revised version of the Three Laws of Robotics, to include a 4th law by James Bridle, saying this:

This, of course, is not as easy as it sounds. One big problem is that the machine learning algorithms don't actually know what the data means.

So, in this article, we will have a deeper discussion about this. I'm going to show you what the problem is, and talk about ways that could solve it.

Please note: Like in the last article, I will talk about technical things, but I will not include any code. This article is mainly for editors or media executives, designed to help you better understand this from a strategic perspective.

Hey, computer. Draw me the number 8

The amazing thing about machine learning is that you can get it to look at a very large collection of data and identify patterns we humans cannot see for ourselves. But, after it has done so, it has no idea what it just identified.

For instance, today we can use machine learning to identify cancer cells in human bodies at a higher level of accuracy and much earlier than what a human doctor can do, but if you then asked the computer what a cancer cell looks like, it doesn't know how to answer that.

Just think how insane this is.

This 30 page report is exclusive for subscribers. (login)

Try it free for one week

Register to try out Baekdal Plus completely for free for one week.

for just...
for just...
You get two months for free


Baekdal Plus is your premium destination for trends and analysis for the media industry. Every year you get 25 reports about the future media trends, business and editorial strategies, monetization analysis and insights about how to use analytics specifically for publishers.

As a subscriber, you also get full access to all the Plus reports (more than 200) published over the past 8 years, as well as the ability to share what you read.

I'm a company, can we pay via an invoice?

Yes, of course, please write to and I will send you a regular invoice that you can pay via your bank. I will need your company name, address and VAT number (if within the EU). Also, please note that due to this process being manual, this will be for an annual subscription only.

Is there an Enterprise Plan?

Yes, please write to for details. But for 25-99 users: the price is 20% off the subscription price ($79/year per user), 100+ users is a fixed price at $5,000 (for all combined).

Can you create a report just for us?

Yes, please head over to Baekdal Media to read about consulting where I can help you with strategy reviews, trend and strategy reports, and strategic guidance for you media company or a specific publication.



The Baekdal Plus Newsletter is the best way to be notified about the latest media reports, but it also comes with extra insights.

Get the newsletter

Thomas Baekdal

Founder, media analyst, author, and publisher. Follow on Twitter

"Thomas Baekdal is one of Scandinavia's most sought-after experts in the digitization of media companies. He has made ​​himself known for his analysis of how digitization has changed the way we consume media."
Swedish business magazine, Resumé


—   analytics   —


Creating a propensity model for publishers


How my focus on analytics has changed as an independent publisher


How can publishers measure trust and other editorial metrics?


A guide to analytics for independent journalists


Why producing less news leads to a boost in subscriptions


GDPR: How publishers can track things without tracking people