Case Study: Introducing Personalised Newsletters at the Atlanta Journal-Constitution

The Atlanta Journal-Constitution (AJC) is a major daily newspaper published in Atlanta, Georgia, and is a merger of the Atlanta Constitution and Atlanta Journal. The newspaper covers local, national, and international news, and is the flagship publication of Cox Enterprises. 

Newsletters have been playing an important part in driving reader engagement at AJC and in 2022, AJC decided to join the JAMES Experimentation Program. Through the program, AJC could test the benefits of automated and personalised JAMES newsletters.

Fast forward to a year later, AJC now uses JAMES for 4 different newsletters, showing positive results. In today’s blog, we share insights from the latest analysis presented to AJC.

You can learn more about how other publishers use JAMES at the Digital Growth Summit in Brussels on September 26th. Seats are getting filled up, get your tickets today.

JAMES Experimentation Program

JAMES is a Twipe technology that drives activation and habitualisation for readers. It enables publishers to automate and personalise the distribution of newsletters to subscribers, through interest based, one-to-one personalisation based on the JAMES Hybrid algorithms.  

The Hybrid algorithms have initially been developed by Twipe & The Times of London and further fine-tuned by Twipe’s team to combine individual reading preferences and trends of the day.  

The JAMES Experimentation Program consists of 5 stages: 

  • Setting up data streams and algorithms; 
  • Defining which newsletters to personalise; 
  • A/B testing personalisation on selected cohorts vs control groups for 3-6 months; 
  • Analysing, collecting insights, defining recommendations; 
  • Rolling out to all audiences. 

To generate the most impact, JAMES is best used to personalise newsletters where the pools of content are large enough for personalisation to be meaningful (e.g. > 40-50 articles) and where interactions are of high frequency (e.g. daily) in order to drive enough individual data for the algorithms to process.  

There are 2 options for starting with personalised newsletters:  

  • Personalising an existing newsletter
  • Starting a completely new personalised newsletter.  

In the past 5 years working with various publishers to set-up personalised newsletters, we have performed both depending on the publisher’s strategy. 

A/B testing personalisation for the Evening Update Newsletter

At the kickoff of the JAMES Program at AJC, we identified 18 different newsletters where personalisation & automation could be meaningful. Together with the project team we selected the first newsletter to use JAMES for: the Evening Update.  

The Evening Update is a daily subscriber newsletter sent at 9:30 PM with a selection of articles from the day. During the experimentation program, an exposed cohort of approximately 40,000 people started receiving a personalised version of their Evening Update instead of the regular editorially curated version. All other subscribers continued to receive the same editorially curated version with a selected group being labelled as control group.

Via JAMES, the Evening Update contained for each individual a selection of articles they haven’t read, and are personalised to their individual reading history which is based on a unique mix of interests and trends of the day. 

Higher click rates for JAMES exposed groups than control groups

One of the metrics tracked during the A/B testing was email interaction with more emphasis on click rates. After 3 months of A/B testing, the average click rate for newsletters based on JAMES was 1.5 percentage points higher than the average click rate of the control group which continued to receive the editorial selection. 

Increased activation and habitualisation of readers on JAMES exposed groups compared to control groups

Aside from email metrics, there are 2 other key metrics to the JAMES Experimentation Program: 

  • Habitualisation – the increase in the % of readers who had 10 or more days of interaction with content in the past 30 days;
  • Activation of zombie subscribers– the decrease in the % of readers who had no interaction with content in the past 30 days. 

These metrics were tracked for AJC throughout the duration of the experiment on both JAMES exposed cohorts and the editorial control cohorts. After 8 months of testing, positive results were seen for both activation and personalisation. 

While an increase in habitualisation has been observed on both groups, thanks to the many initiatives AJC takes to continue engaging readers, the increase on the JAMES exposed groups is 11.8 percentage points higher, which translates into more habitual readers. These are people who are becoming more frequently engaged, which has been highly correlated to a lower chance of churn.

💡In the analysis done at The Times of London during the JAMES research project, JAMES exposed readers had seen a 50% lower churn than the non-exposed readers.

Similarly in terms of activation, an increase was observed on both groups with a higher impact seen on the JAMES exposed cohorts resulting in more zombie subscribers becoming active.

The results of the test were promising so after the initial testing phase, personalised JAMES newsletters have now been rolled out to all Evening Update subscribers.

Similar results were seen at other publishers using JAMES such as NRC, DMG Media, and Keskisuomalainen. You can meet them all at this year’s Digital Growth Summit in Brussels.

Expanding JAMES to 3 additional newsletters

Based on the results observed with the Evening Update, and following the recommendations and benchmarks from the Twipe customer community, AJC expanded the use of JAMES automation and personalisation to 3 other newsletters: morning, weekly and monthly. 

The Morning Update

For an initial cohort of 37,000 readers, JAMES populates a list of 3 recommended articles. This experiment has also proven to be successful with 2.2 pp higher click rates observed in the newsletters with the recommended articles than the ones without. 

The Monthly Update

At the beginning of each month, AJC sends a reactivation email to less active readers and lapsed subscribers. We looked at expanding the use of JAMES to the reactivation email, and this produced good results. Currently the email uses JAMES for all its recipients.

Introducing the new Weekly 10

This summer we have started a completely new personalised email on a weekly basis – the Weekly 10. This is a selection of 10 articles that the readers may have missed from a selected content pool of the previous week’s content.

Images source: AJC Articles

Stay tuned as we will share our results in the upcoming month, or join us at the Summit to hear directly from JAMES users about the performance of weekly emails. 

Join us at the Digital Growth Summit

To know more about how JAMES can help with your business growth, join us at the Twipe Digital Growth Summit on September 26th in Brussels, Belgium.

It will be an exclusive day of learning, inspiration and connecting with industry colleagues. Reserve your spot today while tickets are still available!


Daniel Tee

Get insights on Digital Publishing direct in your inbox

Subscribe to Twipe’s weekly newsletter and receive insights, inspiring content and event invitations directly in your inbox!

Subscribe to our Future of News newsletter