You take a handful of data sources and some smart algorithms, build a few applications and feed them with lots of data – this is what Badische Zeitung (BZ), a newspaper published in Freiburg on the edge of the Black Forest, developed in the past few years.

We call this system Data Machine. Data is its most important resource. We deploy this machine to systematically:  

1. Automate production processes. 
2. Personalise digital products. 
3. Segment our audience using marketing automation.
4. Build new habits. 
5. Curate content.

Data Machine is a multifunctional tool. We collect its data with several external services like Piano, Smartocto, or Chartbeat. However, this data gets its real power and magic by interacting with our own applications and algorithms. Consider the following five examples.

1. Automation

On social media, some of our content is published automatically. Our main Twitter account, @badischezeitung, is fueled by engagement data powered by Piano in real-time. This data acts as a trigger. As soon as an article reaches a threshold, an algorithm of Data Machine fires a tweet, generating keywords, and a paid content label if necessary.

Automation had a big impact on our Twitter reach. We increased our referral Twitter traffic by more than 100% compared to the same month of the previous year.

2. Personalisation

If you listen to Spotify, you’re familiar with Discover Weekly, a playlist tailored to your personal music taste. Badische Zeitung’s Die Woche (The Week) is inspired by this service. Based on individual reading habits, The Week recommends stories published in the previous week that a reader hasn’t yet read.

Important to note: Our algorithms only show the most engaging stories of the week. To select these stories, we add an engagement score to any article. This engagement score is provided by Smartocto, a content analytics service. Thus, The Week only recommends highly engaging content; Piano provides the personalisation.

3. Marketing automation (segmentation)

Badische Zeitung also deploys marketing automation tools provided by Piano. One example is a series recommending hiking tours in the Black Forest that was published in late summer 2020. One major purpose was to acquire new digital subscribers. In Web site article views, we integrated Piano’s personalised content recommendations, showing even more hiking tours stored in our archive.

Only readers who read an episode of the current hiking series saw this personalised module. We called it The Hiking Lasso. Even our paywall was personalised for this audience segment with an atmospheric photo of the Black Forest. The result of this campaign: Almost 150 new digital subscribers. 

4. Build new habits

In March 2020, some days before the first COVID-19 lockdown in Germany, Badische Zeitung launched its new newsletter service called BZ am Abend (BZ in the evening). The newsletter is published at 7 p.m. Monday through Saturday and is distributed to more than 100,000 readers.

The magic: The newsletter runs almost completely automatically. Exactly one hour before distribution, an algorithm selects the most popular content published in the past 24 hours, considering engagement data provided by Piano.

We are able to assign most of the articles to local communities, reaching newsletter subscribers who share their ZIP code when signing in. By this means, our newsletter has a very local flavour. 

The Data Machine algorithm creates more than 200 variants of local newsletter editions automatically. Several minutes before distribution, an editor takes a final quick check — there’s no further human interaction necessary. The newsletter has an average open rate of 30%. It’s one of the most important services of BZ to habitualise readers and build new rituals. 

5. Curation

Most articles published in our printed newspaper are also distributed digitally on our Web site. But the newsroom only optimises a small percentage of these articles for the digital channels by adding an extra headline or a special photograph. However, our editors know that if they optimise newspaper content for the Web, they can boost reach and engagement.

Therefore, the question is: Which of the more than 300 articles published daily should get a channel-specific optimisation?

Our Data Machine supports the newsroom by making this decision. A smart bot analyses engagement data of the printed content after it’s published online and gives recommendations via e-mail as to which articles should be optimised. By this means, Data Machine reduces the cognitive load of our editors and saves them time. 

These examples show that a strict focus on data can leverage value and efficiency. Even before the COVID-19 peak, Badische Zeitung demonstrated a strong digital presence with almost 34,000 digital subscribers and an exponential growth rate. For the first time in 10 years, we saw a positive growth rate of our total subscriber base (print + digital). Above all, this is the merit of the newsroom and our editors. Data Machine helped them to do their jobs. 

This case study originally appeared in the INMA report, The Guide to Smart Data Strategy in Media. 

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .