How AI and mathematics are affecting business media websites. The impact of AI and math on business media.
Data never sleeps. Each click, scroll, and share is recorded into a machine which learns faster than any editorial staff. Today, business media platforms are not only content hubs, but real-life ecosystems of algorithms, models and numbers that work in real-time.
The numbers behind the curtain
The math isn’t read by most readers. It’s ubiquitous, however.
Data is processed in real time on platforms like Bloomberg, Reuters and financial news aggregators, with millions of data points processed each hour. Read the following: Thresholds are paywall thresholds that are dynamic. Even the headline you will see could have been selected for millions of clicks in the past.
Why math became the editor-in-chief
Instead of editorial decisions, A/B Test was introduced. Instead of engagement scores, readership was used. More than three-quarters of the leading digital publishers have picked up on algorithmic tools for making editorial choices in a 2023 study by Reuters Institute, whether that’s to identify topics to consider or timing for publishing.
Short attention span. Competition is brutal. There is no shot in the dark when it comes to publishers. So, they do measure what they do.
Algorithms that learn what you want
It is a widely used component of any modern platform, a recommendation system.
Netflix has cited about a $1 billion annual savings in churn as a result of its recommendation engine. This is the same in the business media. Truly, if someone reads a blog about fintech and is interested in fintech, they read more fintech. The algorithm recurses around the loop and the loop keeps people subscribed.
Personalization at scale
One article. Ten million readers. 10 million different experiences.
How can this NOT happen without math? It’s what modern media outlets do, though. User segmentation models are created based on user behavior, geographical location, device and reading habits. Content delivery goes with the adaptation of it.
This is the rationale that Spotify uses for music. It’s utilized in journalism, e.g., by Axios and The Financial Times. Personalization has become a standard requirement.
Solving the complexity behind the scenes
What is the best time to publish a story? Math gives her the answer to that too.
Predictive models analyze the historical engagement data and determine the best time to publish. Mon (7 AM) is not the same day as Thu (2 PM). The best platforms do not publish, they publish on a probability basis.
Advertising, pricing, and real-time bidding
There is some of this math that is extremely difficult to do by hand.
There is some of this math that is extremely difficult to do by hand. Platforms make use of automatic tools for number crunching. For example, an effective image math solver can take care of complicated monetary formulas, statistical designs, or even relationship projections with just a few clicks. Ultimately, math solvers free up analysts’ time to concentrate on analysis and also communication instead of computation. In industry, these little tools have a great deal of power regarding big decisions.
In programmatic advertising, the bidding system that bids for ad space is referred to as real time bidding or RTB. Just a few of the dozen or so variables in each Auction, User Profile, Content Context, Bid Floor and Viewability Score. eMarketer estimates that programmatic makes up more than 90% of all US digital display ad spend.
Revenue modelling and subscription economics
The subscription business has changed the lives of publishers with regard to the ads.
All tier decisions are based on lifetime value models, churn prediction scores and pricing elasticity analysis. It is no guess: “$9.99 plan vs $14.99 plan” is a calculated gamble, with cohort data to support it. Publishers who do this do so by slowly and stealthily stealing subscribers.
AI-generated content and its limits
AI can write. It can’t do that without context though.
A number of tools were available as early as 2015, like sports recaps and automated earnings summaries. Thousands of financial reports are generated by the Associated Press, a user of automation. The weapon’s benefit is speed. Nuance is still the human edge.
What comes next
AI is taking over multimodal. Multimodal is getting caught up in AI. Audio, video and texts are converging.
Among these, platforms are already testing AI real-time translation and video highlights & podcast summaries on a scale never before seen. The more the formats increase, the more complicated the math will become. But the purpose is simple: the correct reader, reading the correct thing, as long as he or she can.
This content is provided for informational purposes only and is not a substitute for professional advice. AFP editorial staff were not involved in the creation of this content.
Support AFP
Share
