Online News Site

If it is implemented properly, personalization at scale can increase total read time and watch time in case of an online news site, which is a function of how many users there are, how often they come and how long they stay, thus more advertisements can be displayed, while having positive impact on customer experience.

In addition, also conversion probabilities to subscribe and churn can be calculated, the models are updated in real-time, therefore they do not get obsolete. The higher the read time and the watch time the higher the probability that a subscriber will continue to be paying the monthly fee, and that an ad-hoc user will convert to subscriber.

Real-time content personalization

Content recommendations


The personalization engine can automatically suggest relevant content (articles, videos, etc.), based on data about the visitor, or feeds data to another application for purposes of making changes to the presentation layer. These are additional elements on the site where one or more content pieces are displayed. The contents displayed and their order are defined by the personalization engine.

Use cases

  • Main page – Recommended
  • Main page – News from the world
  • Video article page – Others
  • Article page – Others
  • Article page – Check also
  • Article page – Most read
  • Category page – Most read
  • Main page – recommended from other domains (cross domain recommendation to increase traffic between different domains of the publisher)

Personalized content listings


Personal ordering of complete list of content displayed on the site according the selected filtering criteria (e.g. contents from selected category).

On-site personalization


On-site personalization includes the ability to present personalized, on-screen content or messages to a visitor based on specific data gathered about the visitor. This also includes the functionality to adjust elements of the page based on visitor characteristics, and it applies to any page of the site, whether it’s the landing page or subsequent pages.

Use cases

  • Personalized carousel, which means selection of the most relevant carousel image for the given user in the main page carousel
  • Personalized category ordering on the main page
  • Personalized section ordering, which means modification of the order of main page sections, based on personal preference
  • Personalized behavioral pop-ups, which means to display the most relevant pop-ups across the whole domain based on past behavior after a certain time period. E.g. popup to promote subscription package,
  • Personalized welcome messages, which means displaying relevant welcome content on the main page or any landing pages,
  • Personalized banners on the top or bottom of the page, which means to display banners depending on user behavior

Mobile Personalization


All the modules: content recommendation, personalized content listing, and on-site personalization is recommended for mobile apps as well. Probably some adjustments are necessary due to the different form factor and due to differences in the UI, but the main features should be the same. 

Personalization of marketing campaigns

Personalized email marketing


Email marketing can include transactional emails (like application confirmations, purchase receipts, account creation), relational emails (like new subscriber welcome, gated content delivery, blog articles, survey / review, social update, contest announcement, referral request) and promotional emails (like promotional content, sale announcement, new movie release, upgrade offers). Personalization engine can help to increase the efficiency of several of these emails by personalizing the content, or the offer, or the sending time.

Personalization engine is capable to apply this functionality to other channels as well such as SMS or push notification.

Use cases

  • Personalized article selection in newsletter (even multiple times per day, e.g. morning newsletter, evening newsletter)
  • Thematic weekly newsletters
  • Topic recommendation, meaning recommending the most relevant newsletter topics for the reader
  • Email subject line personalization, which means to use the (short) title of recommended product in the e-mail subject line
  • Email sending time personalization, which means sending time personalization based on the recipient’s response rate on past e-mails

Triggered emails and messages


This functionality is a category separate from email marketing, because it requires a specific trigger to trigger the email. It enables sending marketing messages to customers in real time as response to customer activities. Messages can be triggered by any incoming events if the event is synchronized with the personalization engine. Messages can be delivered by 3rd party applications through multiple channels like email, SMS, push notification. This functionality can be combined with personalized email marketing to personalize the content of the triggered emails.

Use cases

  • Triggered message sending by online events
  • Triggered messages based on external events

Next best recommendations

Probability to subscribe and churn


Besides real time content personalization, calculation of probability to subscribe is also included in the offering.

  • In comparison to pure personalization which is a ranking task, we do not provide just the ranking but also calibrated probabilities;
  • In comparison to just an offline probability to buy estimation, our models are updated in real-time, therefore they do not get obsolete;
  • The previous campaign setups in the personalization engine can be reused for the next campaigns, as only the target needs be redefined.

Probability of churn can be estimated as well.

Estimations for an offering by touchpoints and each client


Next best recommendation probabilities can be estimated for several other attributes, as follows:

Use cases

  • Sending time of the messages;
  • Email subject line;
  • Different design elements, for example, background color;
  • Variants of the same text, for example, discount in % or absolute number;
  • Next best action;
  • etc.

We are happy to open up "the black-box machine learning modeling" for online news platform companies