At the Thursday meeting of the INMA European News Media Conference in Oslo, Eivind Hjertholm Fiskerud, head of data and analytics at Schibsted Norway, shared his thoughts on what it means to put Big Data analytics into practice.
Growing digital subscriptions is a strategic priority for Aftenposten and Schibsted’s publishers. Knowing if a user is ready to buy or not can help a news media company create a more relevant news experience and run more efficient campaigns on their own sites and in other channels.
The main objectives of the purchase prediction project have been to understand how user behaviour on the site relates to purchasing and to develop a prediction model to identify news readers likely to buy a subscription.
A key goal of the project has been to develop the prediction model in a site-agnostic way so it can be scaled to all publishers in Schibsted easily.
Aftenposten is the largest newspaper in Norway in terms of circulation and in terms of subscriptions sold. It has been in the print business for more than 150 years, and since 1995 it has been offering a free online version. Not so long ago, it started working with user payments on the side.
Aftenposten had the clear goal of growing subscription sales to 100,000 last year. By setting the new goal, the company wanted to learn how user behaviour leads to purchases. Not only did it reach this goal, but it is now getting close to 150,000. These numbers prove Aftenposten is very good at both producing and promoting its content.
Aftenposten has built a prediction model that looks at the behaviour patterns. It predicts the likelihood of individual users purchasing a subscription, based on their behaviour on the Web sites and apps.
To do this, the company trained a machine-learning algorithm on a data set of all logged-in users from a given observation period during which they did not have an active subscription. The algorithm learns the difference in behaviour patterns between those who do not purchase and those who do purchase during the target period. Some of these people went on to subscribe in the following target period.
Taken from 50 signals sent to the model, the company determined those most likely to purchase the content were users who read more articles, visited the site for more days, visited the site from multiple devices, visited on weekends, and visited from multiple sources.
Data analysis is about being able to take the output from the model and earn money. Behind the success of Aftenposten stands the multi-disciplinary team that is driving results. “We wanted to make sure we connected the engineering side with business,” Fiskerud said.
Aftenposten has led two experiments. One proved the users most likely to buy a subscription according to the model were both more likely to open the app to see the ad and more likely to click the ad if they saw it, demonstrating the potential of using such scores within the product. Another proved a six-fold increase of conversion rates in telemarketing.
Christian Lämmerer, managing director at Styria Marketing Services, presented that company’s approach to data-driven marketing. The application of smart data and predictive analytics to audience data provides an unprecedented level of intelligence to the quality of the relationship a publisher has with every one of his or her consumers. This is ultimately at the core of making strategic decisions across the organisation.
Styria Media works in three areas of expertise to understand the audience:
- Data science.
- Customer acquisition.
- Customer loyalty.
For Styria Media, Big Data allows the company to personalise its actions and design marketing campaigns. It considers the audience, not media.
The company’s approach to data-driven marketing is that an end-customer database is the foundation for all Styria brands in Austria. This central database is present among all the brands, and at all touchpoints such as e-shops, clubs, and competitions. It collects data whenever possible.
Klas Granström, managing editor and head of digital at Expressen, proved his organisation is probably the most data-driven company in the Nordics.
In January 2014, Expressen had low market shares digitally and no data strategy. The analytics team consisted of only four employees located far from the editorial and technology teams, working mainly with print analytics.
As a part of a digital growth strategy, Expressen set a bold objective to become the most data-driven media company in the Nordics and the fastest growing news site in Sweden.
Expressen is a liberal newspaper founded in 1944. A part of Bonnier Group, it has a total reach of 2.4 million people per day and a turnover of USD$140 million (2016). It has been profitable since 2003, and ad sales have been increasing since 2013.
Its digital core products are omnipresent, yet three print editions are still considered important. Expressen has also 12 digital sub-brands to monetise different verticals and touch on as many subjects as possible.
Granström believes data is exceptionally important, and Expressen’s goal is to become the best in the Nordics in that area. Expressen’s strategy is based on three pillars:
- Mobile content and distribution.
- Video and television as the core of news products.
- Social content and marketing.
Expressen realises this strategy by being a data-driven organisation and fostering cross-department collaboration. The Traffic Attack operational project, meant to create the largest possible audience through data-driven daily operations and operational audience development, has also been instrumental in reaching strategic goals.
Although Expressen has 12 data analysts, they are based throughout the organisation — some in sales, some in the newsroom, and some with the development teams -- rather than as one team.
Management presents all quality key performance indicators (KPIs) to the entire newsroom to keep everyone in the newsroom informed. Everyone tracks their own goals. The tools they use include those that allow the measurement of real-time traffic data and audience development, the comparison of different types of metrics against each other, and of different teams against each other.
Management tracks both numbers and quality through KPIs, looking at what is most read, and at text length, number of pictures, and whether video is included in an article. The way content is being consumed is also a measure of quality, and the company has its own article score for this, which is the mix of KPIs (including how far into the text a reader scrolled and how long he or she watched video footage).
Expressen has also a reporter tool. All team members have their own dashboard of KPIs and receive instant notifications about whether their content is not good enough or if they are exceeding expectations. Print operations also use data to track time spent and eye-tracking.
All of this work has resulted in a first-place Google ranking in Sweden, a first-place ranking on Facebook in Sweden, strong traffic growth, and an increase in total views of mobile traffic.