Overview of this campaign
As the global economy has suffered from the unstability in the world, it's become harder to recruit and retain subscribers for Amedias 130 Norwegian media brands - of which nearly all of them have a business model relying on subscriptions.
In this climate, ensuring that our products feel "need-to-have", not "nice-to-have", becomes crucial. And as always, when Amedia go about solving our issues, we turn to data.
For years, we've built a data platform consisting of several integral parts. Clean, accurate user profiles remain a key part of our strategy and product development. At the moment, we've got more than 2,5 million adult Norwegians registered through aID, our sign-on solution, able to visit all of our media brands through the same profile. In this way, we've learned a lot about our users, their behaviour and what drives their engagement.
On the other hand, we've spent significant resources building ACS - Amedia Content Store - which is both our content storage and a library brimming with metadata.
When we paired robust, precise user data with advanced knowledge about our content, we realized we could create powerful, personal news experiences. And so we set out to do.
Our objective for our latest push in personalization was, at it's core, easy:
To create a better, more personalized experience for our users by providing them with content that's tailored for them. The end goal is to build loyalty by ensuring that our users spend more time consuming our content, which we know is the biggest loyalty indicator.
After a year of testing, iterating and deploying personalization throughout Amedias media brands, we can safely say we're thrilled - and slightly surprised by the results and findings.
Results for this campaign
Results overview:
- 4,4% increase in daily usage among subscribers YoY
- 14% increase in the number of well-read stories
- A personalized web site teaser performs 81% better than regular, editorial teasers
- We've grown from 440.000 digital subscribers at the start of the year, to 466.000 digital subscribers
For years, we've experimented with customized user experiences for individuals. The results, though, have been a mixed bag. But after building our content store, we ensured that we could categorize our content properly and keep track of it's metadata, while at the same time making all this knowledge available for all of our product teams. This was a game-changer.
At it's core, Amedias personalization engine receives information about our content from the content store, enriches this data with information about every single user that's visiting our news sites, and customizes what content each user is provided with when visiting our sites, using all of this data. This isn't limited to editorial content: The same engine handles all of our content, including commercial content from our subs or ads teams.
One of the things we realized, and that surprised us, was that the best way to personalize isn't through advanced algorithms, or, for instance, trying to leverage AI to do the sorting for us.
If you've got precise data, the best way to drive engagement proves to be through the use of a simple, logical and rule based system. And the best way to improve your results, is by iterating the system - one step after another.
Since we've got 110 media brands, it's easy to conduct controlled A/B-tests. This has been crucial to our process. Our product development loop always contains small scale testing, then deplyoing at scale. At the same time, it's been important to be transparent about what we do, and ensuring that our newsrooms and editors understand the process. It's crucial that we deploy systems that ensure editorial control.