In our recent blog we wrote about how big data and AI machine learning will affect content personalisation. In this article we want to delve further into the topic of content personalisation. Read on to know why your media publishing brand needs to implement content personalisation APIs onto your website.
What is a personalisation – and how does it work?
Some of the biggest publishers in the world use personalisation to suggest content that a specific reader might be interested in, offering this tailored experience is designed to boost the time a user spends on the site and the number of page views they have.
This comes from the deep need to have more pages views and a more engaged audience to improve the rates they can charge advertisers for sponsorship and programmatic advertising exchanges.
Personalisation works by having a small analytics package collect information about what a reader is viewing and how long they are staying on the site. Data from analytics is then used to compare users and interests, using machine learning. It suggests the best content for that reader to be shown next, based on what other users with similar profiles have read.
Machine Learning is a new field and can cause issues if used poorly. It is vital to understand what machine learning is, how it works and what your responsibilities include should you decide to use it for any purpose.
NVidia defines Machine Learning like this: “Machine learning at its most basic is the practice of using algorithms to parse data, learn from it, and then make a determination or prediction about something in the world.”
Issues – collecting data, processing data, the dataset defines the output.
If you have bad data you get bad results. This potentially causes problems when designed in a monoculture and it reflects the opinions and assumptions of its creators.
Is Personalisation worth it?
The beauty of content personalisation APIs being powered by machine learning, is that no human participation is needed. On a basic level, machines analyse information from a browser’s search patterns and recommends content for the reader, all without a human lifting a finger. Modern computing power means machines can create data and predictive modelling much more accurately than a human can.
Collaborative or social filtering is a personalisation method which uses the recommendations of other users to make suggestions to readers. This type of personalisation is based on the power of social proof in marketing, people are always likely to try something new based on the recommendations of others with similar interests or needs. It’s this that makes reviews so powerful online.
Item based collaborative filtering is when a machine finds similarities between lists of articles on a publisher’s site.
User based collaborative filtering is where a machine finds similarities between a user’s browsing history and other articles on a website.
In collaborative filtering if user X and user Y both read the same article, machine learning can take another article read by user X and suggest that user Y might like it based on their preferences.
In content filtering a user reads an article and machine learning recommends similar articles that it predicts the reader will connect with.
Netflix uses this personalisation method. Using social filtering Netflix filters content by popularity with its ‘Trending Now’ category and by location with its ‘Top 10 in the UK Today’ category. It also suggest similar titles to those you’ve watched and enjoyed: ‘Because you watched …’. This keeps users on the platform longer and reduces their discovery time, which has a strong knock on effect to platform loyalty.
Amazon is another well-known brand which gets personalisation right. Amazon offers recommendations on product pages for items which might accompany what you are considering buying by offering a feed of compatible products, ‘Frequently bought together’.
It also makes recommendations based on its ‘Wish List’ feature as the company knows these are items the user is especially interested in. Because of this it can confidently recommend users products that they will be interested in purchasing.
Personalisation APIs for publishers
One such API which has proven beneficial to publishers is Parse.ly/api. Parse.ly offers immediate content recommendations and personalisation to publishers like Wired and Slate.
Their API automatically reads and measures a publisher’s content archive across their website, meaning their AI technology can present tailored content to users quickly.
To show that personalisation can impact a publisher’s bottom line, Slate received a 105 per cent uplift in subscriptions after implementing Parse.ly, and WIRED were able to get their marketing messages in front of many more eyes after a 400 per cent increase in newsletter subscribers since using the tool.
Parse.ly’s smart algorithm is customised for performance. It takes into account user browsing patterns, topic relevancy and popular content to make accurate recommendations. Brands can perfect recommendations based on user demographics, and company goals to boost performance and hit KPIs.
Slate is an online magazine covering politics, business, technology and the arts. It has a humorous take, while providing insightful commentary on current events and politics.
“Slate has relied on Parse.ly’s recommendations API to grow time-on-site and reader loyalty through several iterations of our site.”
“It has powered everything from our most popular modules to personalized recirculation links to our queue of infinitely scrolling articles.” David Stern, VP of Product and Business Development, Slate
Machine learning means accurate recommendations
API machine learning content personalisation tools mean we can make recommendations with greater accuracy than ever before, with less effort needed.
Content suggestions are made by thorough analysis not hunches or guesswork. This encourages greater user satisfaction and a real boost to a publishing brand’s bottom line.
Get in touch with us today and we’ll discuss how content personalisation APIs can open the door to achieving your brand’s goals and improved customer experience.