top of page

Bringing AI to Publishing: Pelcro Announces Its Newest Partnership with Protégez-Vous

Updated: Jan 19


Pelcro is pleased to announce its newest partnership with Protégez-Vous.

The well-known consumer magazine recently partnered with Pelcro to take advantage of the platform’s machine learning capabilities and will be using the technology to predict subscriber likelihood and decrease churn rates. Protégez-Vous is a non-profit that tests thousands of products in specialized facilities in order to offer consumers expert advice on consumer products. Subscribers can access content from their magazine, website, and special series guides. In order to remain objective, the publication is ad free. We look forward to this partnership as an indication of AI’s bright future in publishing.


Machine Learning for Media and Publishing

Pelcro’s solution was created with the media and publishing industry in mind. Currently there is very little available in terms of AI paywall technologies created specifically to address publishing needs. Publishers either have to work with third party services such as Amazon Web Services and then collect and model the data themselves, or hire a consultant. Neither of these options is very time or cost effective.


Pelcro’s platform allows publishers to easily leverage machine learning to increase subscriber revenue without sacrificing user privacy or security. The technology collects the data, models it, and then processes the predictions, automating the entire process.


Subscriber Likelihood Prediction: When someone visits a website there is a lot of insight that can be gathered from simple interactions. With machine learning you can model the data and understand which patterns of behavior lead to subscription, using this information to predict subscriber likelihood and tailor subscription offers.


Something unique that Pelcro brings to the space is its ability to differentiate between content that is and is not behind the paywall, offering dynamic content recommendations based on this information. If the information reveals that users generally subscribe only after the tenth visit, the dynamic paywall will not recommend articles behind the paywall if they’re coming on for their fifth or sixth time. Only after the tenth visit will it begin recommending premium content.

Churn Rate Scores: Just as there are pathways to subscription there are also patterns for churn. Looking at statistics on web engagement, email engagement, and a host of other factors reveal patterns of behavior that can indicate a high likelihood of churn. Machine learning can then model the evolution of customer behavior and create a customer profile based on these behavioral trends. For users with a high churn score, the platform will automatically offer content recommendations tailored to a particular user’s interests, allow you to start an email workflow to re-engage users, or offer special incentives.


With machine learning you can build multiple models for different types of churn (such as downgrades versus lost customers). You can also track cost of churn. Often a false negative (predicting a customer will stay when they actually churn) is more expensive than falsely predicting a customer will leave and offering them an incentive. In the first case you lose a customer and have to pay all of the associated costs of acquiring a new one, in the second you are still continuing to build customer loyalty. Having this insight empowers publishers to understand their user behavior and more accurately address issues before they arise.

With the industry increasingly shifting towards reader revenue, predictive analytics and machine learning are the next phase in understanding how to provide readers with the best user experience while still conforming to modern data protection regulations. 

Comments


bottom of page