News

Data-sharing project DRIVE aims to help publishers build digital revenue

2022-04-13. German press agency dpa and consultancy firm Schickler launched DRIVE, a collaborative project with a focus on digital subscriptions and joint experimentation to allow publishers to learn from each other.

by Simone Flueckiger simone.flueckiger@wan-ifra.org | April 13, 2022

Organised by German press agency dpa and consultancy firm Schickler, the DRIVE (Digital Revenue Initiative) project was launched in 2020, and now comprises 20 independent publishers.

“The basic motivation to start this initiative was that we all felt that we are lagging behind in building significant digital revenue,” said Meinolf Ellers, Chief Digital Officer at dpa, Germany, during WAN-IFRA’s recent Digital Media India Conference.

“Here in Europe, Scandinavian publishers like Schibsted and Bonnier were first in saying to ‘get rid of digital advertising strategies because they are only good for the platforms. We are not generating the revenue we need to monetise and refinance our newsrooms.’ So they made a big strategic shift to rediscover reader revenue, build the paywalls and start to convert people into digital subscribers. We are following the same path.”

In the first phase of DRIVE, Schickler’s data team built a data warehouse, which stores the anonymised usage data of the different publishers. Using an open source tracking tool integrated on the publishers’ website, usage, content, and subscription data is being recorded and fed into the warehouse, which currently contains some eight billion data points.

“This, of course, is a fantastic basis to train algorithms in order to analyse this data,” Ellers said.

120 dashboards to help benchmark performance

Participating publishers have access to some 120 dashboards, enabling them to benchmark their performance against the entire group or individual partners.

“This is very important because if you are lagging behind in certain aspects, you will see who is leading the group there and you can copy them, and learn from the others who are doing better,” he said.

Highlighting this spirit of collaboration and experimentation, he cited one example from a publisher who started releasing articles that summed up the events of the day in a certain region, and generated very good results. Other publishers then started introducing the same format step by step, and also found success.

“We do A/B testing, such as, ‘Is it better to present a lead paragraph or just the headline?’ And we can directly compare which format is doing better and everyone is sharing their learnings and can use them to optimise their performance,” Ellers said.

Pushing for cultural change

Naturally, participating in a project that encourages publishers to move away from print-centred workflows and establish a mindset that prioritises digital revenue and subscriptions requires cultural change.

“If you don’t get into the minds of the editors that they embrace data-informed publishing as the new driver of their work, you will fail,” Ellers said.

What helped the organisers of the project in pushing for cultural change was the creation of a key metric called Media Time, which, simply put, measures the cumulative time spent on a publisher’s website or app by a single user in the past seven days. It also contains a range of other metrics, such as number of visits, session length, and correlates with articles read.

“We first thought about using conversions as the key metric,” he said. “But conversions and how to translate usage into subscriptions is very far away from the daily work and the thinking of editors.”

dpa is working off of the proven hypothesis, according to Ellers, that higher Media Time increases the chance of converting users into subscribers, while reducing the likelihood of churn for those who have already taken out a subscription.

Print DNA prevents publishers from reaching digital goals

In order to increase conversions and Media Time on articles published, participants in the project are now also working with personalisation. In cooperation with Schibsted in Norway, dpa found that only about 20 percent of overall content produced drove traffic, while the remaining 80 percent attracted minimal usage.

“We basically fail to reach our goals with most of the stories we produce, so this mismatch is a big headache,” Ellers said.

“We believe that has a lot to do with our print DNA. Newsrooms are still very much driven by filling the pages and daily agenda setting. And both are not helping to meet our digital goals,” he added.

To counteract this, dpa is implementing step-by-step personalised user experiences. Currently, they are training algorithms to support them in achieving personalisation, which is only possible if publishers understand the user and user preferences, Ellers said.

As an example of personalisation, users may be presented with a small number of stories based on their preferences and interests, in addition to a handful of signature, brand-defining stories that will be shown to everybody.

The importance of trust

The fact that publishers participating in the DRIVE project are sharing sensitive usage and user data with each other has often prompted the question as to why they’re willing to do that, Ellers said.

“It has a lot to do with trust. The dpa is the national news cooperative of Germany and all these publishers are our shareholders, and that helps us to start initiatives in order to help the publishing industry. We are accepted as the honest moderator and the neutral institution, and the same is true for the Schickler consultancy. They have worked for all these publishers so I would say we are all members of the circle of trust here and this is very, very relevant.”

On May 24, WAN-IFRA’s Data Science Expert Group, an international ​community of practice ​​for news media professionals working in data science, data analytics and data engineering, will be hosting a virtual meet-up focused on learnings and insights from the DRIVE project. Click here to register.