Project Description

PUBLICATIONS

PERSONALIZATION AT SCALE

Background

The Client, a publication house, wants to leverage their digital content effectively to drive engagement on digital channels to drive sales and loyalty in 4 countries.

By determining the relevance of every article to every individual reader, the Client would be able to show the right content to every individual to increase their perception of relevance and subsequently increase their engagement.

Challenge

There are thousands of articles to choose from. Traditional methods such as tagging articles to identify similarities, and simple algorithms such as “others have also read this article” are not sufficient to match the readers’ interests at scale. There is a need for an “engine” that inspires readers to engage with more content and build a habit of using the Client’s digital services and subscribing.

Solution

Determine an individual reader’s preference to the Clients content across brands and feed that into an automated decision process of who should see what content, in what order, using One Prediction.

One Prediction’s “Cognitive Match” feature builds a unique profile per reader, using AI and ML. The “Cognitive match” assesses the relevance of a specific piece of content for each individual profile.

The most relevant content items based on One Prediction’s scores are automatically merged into emails, using the Client’s existing Marketing Automation platform.

Results

The Cognitive Match features account for between 60% and 90% of the models outcome.
The models’ precision is higher than 90% (A.U.C).

Effects of the loyalty is pending as case is still ongoing