Mediahuis, headquartered in Belgium, set itself the goal of implementing a fully automated data-driven customer journey with the aim of increasing revenue and reducing churn.
With this in mind, the company wants to target its potential and existing customers with the best registration offer, the best sales offer, and the best retention journey in an automated, efficient, and personalised manner.
In recent years, Mediahuis, has rapidly developed and acquired several media groups, such as NDC mediagroep in the Netherlands in 2020 and Ireland’s Independent News & Media in 2019. The company is now active in four countries, counts some 4,300 employees, and says it brings in 1 billion euros in revenue per year.
“The main question among all our entities, at least from a data science and business perspective, is how do you acquire new subscribers, keep current subscribers happy, and retain those subscribers, while doing this in the most efficient, personalised manner,” said Jessica Bulthé, Data Science Business Partner at Mediahuis during WAN-IFRA’s recent Digital Media Europe conference.
“The answer? A data-driven customer journey.”
Why retention policy deserves priority
Using a churn propensity prediction model, Mediahuis is targeting a selection of existing customers who seem likely to churn but worthwhile to keep, and who could potentially be convinced to stay on.
“In an era of relatively scarce small data, for instance, one CRM database is more or less manageable to be followed up by human assessment,” said Jan-Bart Vervenne, Data Scientist at Mediahuis.
“But in the big data era where tracked digital reading behaviour and service centre interaction data is available as well, this task has become more and more complex. Computer algorithms can be of help to support humans in this process.”
He outlined the steps Mediahuis took to build a churn propensity prediction model:
Step 1: Collecting raw customer data from company databases (subscription data, 1st party data, contact centre data, webshop transaction data, digital reading behaviour data)
Step 2: Equating whether customers remained loyal in the past against a selection of customer behaviour variables derived from those company databases.
Step 3: Letting the computer derive patterns that differentiate between loyal and non-loyal customers in the past (the training part).
Step 4: Comparing patterns with current customer behaviour. Based on the degree of pattern correspondence, current customers are scored on their future loyalty.
“By combining churn score information with other customer information, a churn model can be a pillar of a more systematic retention policy,” Vervenne said.
The results of the churn prediction model show that it is able to predict loyalty at an accuracy of 91 percent, able to predict that 60 percent of those that effectively churn will churn, and that 75 percent of predicted churners effectively churn.
Bringing churn propensity prediction into retention policy
For Mediahus, making use of this model for its daily retention business is the most important step, which currently remains a work in progress.
This process involves continuously updating all customers’ scores with the most recent data on a regular basis, as well as defining, testing and scaling up audiences and inbound and outbound treatment scenarios based on churn scores and other KPIs.
Inbound treatment scenarios, which have yet to be defined, refer to scenarios that are pursued when customers initiate an interaction with Mediahuis, whereas outbound refers to scenarios where Mediahuis takes the initiative to contact customers.
In addition to its churn prediction model, Mediahuis started working on a propensity to buy model a few months ago to be able to predict whether or not someone will buy a subscription and what type of subscription they will buy in the next month.
In terms of initial results, Bulthé said they looked promising for both logged-in and anonymous users, with the model correctly predicting in eight out of 10 cases whether somebody will buy a subscription in the next month.
“That’s a good first step towards adding propensity to buy and the next best action modeling for propensity to buy,” Bulthé said.
The post Mediahuis’ plan to create a fully data-driven customer journey appeared first on WAN-IFRA.