Hi there. I need you to know that I am going down a rabbit hole on personalisation. It’s actually more of a burrow and a very interesting one at that. Although it can be a complex subject, I’ll simplify the personalisation process, variables, and dependencies as much as I can as we investigate this further. So buckle up because I’ll be using the next couple of newsletters to cover what I have and continue to learn.
Why would you personalise?
The more I look into this, the more I am convinced that personalisation isn’t just a way forward — it’s necessary. Gone are the days when we can talk about an “average consumer.” Our customers have specific needs and problems that need addressing. In order to delight customers with outstanding problems, we must address those needs.
As this is the product initiative, let’s take a product approach and look at personalisation from a customer problem perspective:
Customers can’t easily see the most relevant article to them. Some articles they would enjoy are buried deep, so they get missed.
When customers know where to find the most relevant articles (or think they do), they often need to bounce around the app/Web site to find them, causing a lot of friction.
Customers feel that other platforms know them well: Netflix highlights TV shows relevant to their interests right away, Spotify curates lists based on their preferences, Facebook is a highly personalised feed. But your site doesn’t.
The knock-on effects of these problems are numerous. Here are a few examples:
New users may not see enough content that would make them sign up/subscribe.
Customers may get frustrated with friction and move on quickly, meaning your content isn’t being read/viewed as much as it could be.
Content that is high value for a low number of people is rarely seen by the right people.
Your products and content are not being optimised for your business needs.
Let’s take a real example: Aftenposten in Norway. They looked at their different cohorts of users, saw how different they are (and therefore how diverse their habits are), and realised that curation for the “average” user doesn’t work. Or at least not as well as it could. Editors have always decided what goes on the front page based on what the most important stories are. But that’s subjective. It’s always exceptionally difficult, actually impossible, to do that for all audiences.
Before the above paragraph is misconstrued or oversimplified, this does not mean the home page should be moved out of editorial control or that we should immediately personalise our home pages with out-the-box recommendations (please, please do not do that). Editorial is absolutely key in determining the rules and guardrails of any automation. In fact The New York Times appointed some editors to work with the product team on “providing a more valuable experience.” For personalisation to work, it must be a true partnership between product, editorial, design, and technology. We’ll delve into these roles, and dependencies, in future posts.
Personalisation: 3 dependencies to consider
Now you’re convinced that personalisation is the way forward. How do you actually start? From the conversations I have had so far, there are three main dependencies:
Technical capabilities. The good news is you don’t need to build a personalisation recommendation engine internally. You can if you want, of course. It may give you more control, and there are plenty of open source software options to help this. But most of you are not tech companies and will likely prefer out-of-the-box solutions such as Google Rec AI.
CMS. Presuming we are mostly talking about content personalisation, you will be limited to the capabilities of your CMS system. What is tagged? How is it tagged? Taxonomies of content are vitally important.
Ability to test. Personalisation is a long game. It will need constant tweaking from your learnings. You’ll need to test, tweak, test, tweak over and over.
Once you have taken stock of your main dependencies, you need to decide who you are optimising for. Some media organisations do it for all users, hoping to engage new and less frequent visitors more to move them down the funnel and engage them more.
For example, Mediahuis Netherlands found that if users engaged with a particular topic, they were more likely to sign up if you show them multiple articles on the same topic rather than show the breadth of your content. And Schibsted in Norway saw a 20% uplift in subscription sales by using their conversion maximiser ranker.
Many news organisations offer personalisation as an additional benefit to registered users and/or subscribers. You could optimise to build daily habit or for types of content. If you wanted to go deep, you could also optimise for business reasons such as showing more profitable content such as video, which carries higher CPM than text articles.
Once you have been through these exercises, you can take your consumer problems, what you are optimising for, and your dependencies, and use these to create your strategies. Here is an example of how Schibsted in Norway has put these together to create their four main strategies:
Sometimes we see personalisation as one thing. But actually there are many different types, as this tweet states. And, as with everything product related, this isn’t a one and done, it’s an ongoing process.
- Interview with Yang Sui when he was director of product management for personalisation at CNN.
- If you want to go deeper on personalisation, take a look at INMA Knows: Personalisation.
About this newsletter
Today’s newsletter is written by Jodie Hopperton, based in Los Angeles and lead for the INMA Product Initiative. Jodie will share research, case studies, and thought leadership on the topic of global news media product.