Hi! This is Readers First, a newsletter for INMA members on reader revenue innovation. I’m researcher-in-residence at INMA. E-mail me at: grzegorz.piechota@inma.org 

Was this e-mail forwarded to you? Sign up to this newsletter.

1. PROPENSITY MODELLING: Targeted paywalls outperform the most popular paywall models

Two experiments carried by mid-sized publishers in Europe and in Asia provided evidence of promising performance for paywalls targeting users based on propensity modelling.

The results of the experiments were shared exclusively with INMA by Deep.BI, an analytics platform provider, and presented at the INMA-supported GNI Subscriptions Lab last Friday.

• During the AB tests, carried out between January and May 2020, the visitors of both Web sites were split into two groups in the proportion of 70 to 30.
• Group A was a control group treated with an existing paywall solution: a freemium model in case of a European publisher and a metered model in case of an Asian publisher.
• Group B was a test group treated with a paywall targeted based on a propensity to purchase model trained with the data of respective publishers.
• The European publisher with the freemium model observed a conversion rate lift of 166% and the Asian publisher with the meter — 77% lift. 

Based on my 2020 study of the most popular paywalls, only 5% of news sites in 33 countries were hybrids that combined different models such as a freemium and a meter, or had a dynamic paywall targeting readers based on behaviours and other attributes.

The freemium model is by far the most popular among news brands — 47% of sites tested in 33 countries had some content free to all users and some premium content accessible only to subscribers. The meter model, in which readers can read a limited number of stories for free, is followed by 12% of brands. 

Targeted paywalls stop readers selected with advanced analytics techniques, such as data mining, statistics, modelling, and machine learning.

• In a nutshell, data scientists collect the historical or live data from different sources: for example, online traffic data, content data and customer data.
• After cleaning the data, they analyse it algorithmically to identify patterns, such as the behaviours of readers who subscribed to the site in the past.
• Then they can deploy statistical models to predict future outcomes based on similar patterns, for example they can calculate the likelihood for a new reader to purchase the subscription.
• The segment of prospect subscribers can be then fed to the paywall or e-mail management software to target the offers.
• According to Michal Ciesielczyk, head of AI engineering, Deep.BI uses an ensemble of machine learning algorithms such as Gradient Boosting Machine and Random Forest in its predictive analytics.

Being a registered user to a news site is found to be universally the most predictive feature indicated by five propensity to buy models that Deep.BI deployed at mid-sized news publishers, local and national, in Europe and Asia.

• The ranking of the predictive features is followed by interactions with the paywall such as clicks on an offer or the frequency of being stopped.
• Then come temporal variables such as a loyalty score (time since the first visit ever), frequency of visits, habit score (regularity of visits), time spent on the site, as well as a short- and long-term changes in the engagement score called RFV (measures recency, frequency of visits and volume of articles read).
• Other features found predictive are: whether a reader signed up for a newsletter, the number of shares on social media, and the number of images clicked and videos played.

According to Ciesielczyk, the biggest challenge in building successful propensity models is the amount of behavioural data that can be attributed to a single reader. It is easy when the reader is registered and logged in, as her every action can be tracked. It is much harder when she is anonymous and tracked only with cookies.

Deep.BI found a minimum of 1,000 transactions and six weeks of traffic data is required to train its propensity models. 

Share your case study with INMA peers: Have you deployed predictive modelling? What have you learnt so far? E-mail me at: grzegorz.piechota@inma.org 

2. COVID BUMP: The first churn benchmarks suggest newly acquired subscribers stay 

News subscribers acquired during the COVID-19 pandemic retained better after the first month than those acquired before the crisis. This comes as a surprise, as many feared these customers were attracted by the news on the pandemic and might cancel once the interest wanes.

• This insight is based on the study of 295 paywalled news sites, released by Piano, a publishers’ business platform.
• Cancellations of monthly subscriptions acquired in March dropped an average of 17% compared to subscribers acquired in January and February.
• According to Piano, European sites retained better than the United States — they reduced churn by an average of 34%, while in the U.S. churn was flat.
• “Given the big increase in acquisition, even flat churn rate is impressive,” believes Patrick Appel, director of research at Piano.
• In March, the European news sites sold on average 146% more subscriptions than in February, and the U.S. sites sold 60% more. 

Last week, after three months since the COVID-19 outbreak in Europe and in the United States, weekly online subscription sales continued to be above the pre-pandemic levels.

• In the week of May 24, the new starts in Europe were 85% up versus the weekly average of January and February.
• The new subscription starts in the U.S. were 25% up comparing to the pre-COVID averages.
• “The European increase is driven by promotions, so I wouldn’t read too much into the divergence,” Appel explained to INMA.

3. BUNDLING: Readers of Norwegian Amedia local sites can upgrade to get access to news in other regions 

In the middle of the pandemic, Amedia asked subscribers whether they were interested in news from other regions in Norway. The results were so promising, the company launched in May a paid upgrade offer.

• The new digital-only bundle is named simply “+Alt” (“+Everything”).
• Since May 18, subscribers to any of more than 70 local and regional news sites, published by Amedia, can upgrade to the Everything bundle priced at US$25.90.
• The additive cost of an upgrade depends on the rate they currently enjoy — readers of a small local newspaper Lier Posten pay US$11.30 monthly for the digital-only plan, so they need to pay more than double the rate to enjoy everything.
• Readers of a larger regional newspaper Drammens Tidende US$22.80 monthly, so for them the upgrade costs just US$3.

In the interview with INMA, Executive Vice-President Pål Nedregotten revealed the company has been playing the idea of a multi-title bundle for many years: “We thought a lot how we can increase engagement of our subscribers.”

In 2018, it made the first experiment: A selected group of subscribers to their newspapers in northern Norway got an e-mail with an offer to pay for access to other news sites. Those who expressed an interest got a message: “We don’t have a product yet, but here’s a free access to other news sites for six months.”

“We wanted to test a concept in the market without spending a penny on development. We learned this trick from the Financial Times,” Nedregotten explained..

In the autumn of 2019, Amedia launched a number of experiments — testing different selling points and prices — planning to scale the project in 2020.

The right time came when the pandemic hit Norway in March, and the local newspapers observed a spike in traffic.

“We stopped the tests, and just offered a Norwegian audience an opt-in free trial until May 1. It was a resounding success: Out of 600,000 subscribers, 175,000 signed up for the trial,” Nedregotten said.

All news sites of Amedia use the same identity solution called aID, a cornerstone of the company’s first-party data strategy, so an upgraded reader basically uses the same login to access any Amedia site.

At this point, there is no aggregated news feed, Web site, or an app for Amedia content across the regions. The company has created such a Web site for video streams from the local sports games across the country.

“In this phase, we focus on making the ‘+Alt’ offer work. In the next phase, we will explore how to aggregate and present content, for example in the form of personalised news feeds and newsletters,” Nedregotten said.

Although he didn’t share the results for the first fortnight, he said that the uptake has been high: “We had an ambitious goal that we wished to reach by summer, and we are reaching it in a week.”

Share your case study with INMA peers: What are you experimenting with? What have you learnt so far? E-mail me at: grzegorz.piechota@inma.org 

About this newsletter 

Today’s newsletter is written by Grzegorz (Greg) Piechota, Researcher-In-Residence at INMA, based in Oxford, England. Here I share results of my original research, notes from interviews with news publishers, reflections on my readings. Previous editions are archived online. 

This newsletter is a public face of a revenue and media subscriptions initiative by INMA, outlined here. E-mail me at grzegorz.piechota@inma.org with thoughts, suggestions, and questions. Sign up to our Slack channel.

Banner image courtesy of Phil Shaw from Pixabay.

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