AI can help news publishers personalise while scaling up

By Sarah Schmidt


Brooklyn, New York, United States


Artificial Intelligence dramatically alters all facets of the media business and has vast potential for digital marketing. Its most powerful use may be helping publishers personalise experiences and reach individual readers in unique ways.

The challenge is honing in on the most practical and efficient ways to use it, said Rajkumar Venkatesan, marketing professor at the University of Virginia’s Darden School of Business.

Venkatesan is an expert in marketing analytics and studies the evolving role of AI in understanding consumer behaviour. During INMA’s Media Subscriptions Summit on Wednesday — sponsored by AdvantageCSChargebee, Chartbeat, Google, FT Strategies, Piano, United Robots, WordPress VIP, and Zuora — he shared key takeaways from his recent work that can help publishers take advantage of AI.

Venkatesan's three top messages are that data is still king, privacy regulations can actually help customer-centric businesses, and human intelligence will always be relevant and necessary. 

Marketing is really about predicting what a customer wants and then delivering the right, personalised experience, Venkatesan said. AI can tremendously help: “The goal of AI here is to personalise more effectively.” 

Venkatesan likened this task to expanding a small local chain in Charlottesville, Virginia, United States, called Shenandoah Joe. Customers loved the original shop because the owner, Dave, was popular and friendly and knew all his customers. 

“He knows your name, he asks about your kids, and he knows what kind of pastry you like,” Venkatesa said.

The coffee was good, but that was beside the point. When Shenandoah Joe added new locations, the company risked looking more like another Starbucks as they could offer less of what everyone loved about the chain — it was friendly and made people feel welcome. When it expanded, it needed to do so in a way that preserved that character.

In the age of AI, almost every company faces a similar challenge.

“How do I keep that same level of personalisation while scaling up? AI can help us with this,” Venkatesa said. 

News companies should personalise how customers engage with brands at all stages — from acquisition to retention to growth and advocacy. 

A framework for growth

Venkatesa, along with co-author Jim Lecinski, created a framework for this process in their book AI Marketing Canvas (2021) by looking at brands that had figured out how to do it successfully — a diverse array of companies that included Google, the Washington Post, McDonald’s, and John Deere.

They found a consistent pattern that included five maturity levels that can be applied to all stages of the subscriber journey: acquisition, retention, growth, and advocacy. 

Collecting quality first-party data provides the foundation on which publishers can build.
Collecting quality first-party data provides the foundation on which publishers can build.

The first foundational stage includes collecting good quality, first-party data. The rest of the framework depends on this.

For publishers, the initial subscription step is the key. Good quality data collected during the subscription process while obeying privacy laws will allow publishers to deliver a deeply personalised experience to each reader that keeps them engaged. 

At the experimentation stage, companies can learn from the data, finding what works for their brand and customers. After that, they can expand and broaden the scope of personalisation, eventually transforming to a more customer-centred model to build upon.

The reward — monetisation — comes at the next stage. At this point, successful companies can even start providing other companies with information about what they have learned. 

The process is not necessarily linear, though, Venkatesan explained: “It’s more of a puzzle you’ll fill in as you go.”

In the time that’s passed since researching and writing his book, Venkatessan said the lessons he learned held. But he also wondered what it all meant for the future and kept returning to three central questions that he called “the whatabouts.”

3 key questions answer “what about?”

1. Is data still valuable?

ChatGPT is by far the dominant AI platform, meaning almost everyone uses it. “If everyone’s using the same algorithm, how can companies differentiate themselves? In my view, the differentiation comes from data.”  

Ventkatesan used Netflix and Disney as examples. They may both have access to the same AI capabilities, but they each have data about their own customers.

“We know our customers better than anyone else. That will give us the advantage,” he said. AI provides an additional layer of data from algorithms to personalise platforms for each customer. Good data means you can generate value. 

2. What about privacy?

Like most people, Venkatesan feels a bit spooked by the potential of AI: “That spookiness comes from worrying about the invasion of privacy.”

And if effective personalisation depends on data, should companies worry about privacy concerns and privacy regulations?

Venkatesan decided to study the issue by looking at the effects of the European Union’s General Data Protection Regulation (GDPR). He compared U.S. companies that acquired European companies to those that only acquired other U.S. companies and, therefore, had no exposure to the GDPR. 

Although overall AI valuations of GDPR-exposed companies were reduced, that difference disappeared when he narrowed the comparison to customer-centric acquisitions. 

Venkatesan discovered that valuations of customer-centric AI acquisitions were unaffected by GDPR.
Venkatesan discovered that valuations of customer-centric AI acquisitions were unaffected by GDPR.

“When you focus on customer value, privacy regulations are your friend,” Venkatesan said.

This also suggests that privacy regulations improve customer trust, which can go a long way toward improving the customer experience. 

3. What about humans?

Venkatesan also explored “algorithm aversion” and the widespread worry that AI will replace humans in harmful ways.

He looked at the research, including a study from the Pew Research Center, which showed that people are particularly uneasy about AI’s potential in healthcare, including surgery and mental health. At the same time, he found that people held AI to a higher standard for accuracy:

“We don’t think machines understand us, and they don’t know our feelings. We also don’t tolerate mistakes from AI, but we tolerate them from humans.”

The lesson to be learned here is that human intelligence cannot be left out of the equation. In fact, he learned a similar lesson when interviewing journalists and marketers at The Washington Post for his book.

When The Post began using AI in news writing, it enlisted journalists to train its engine. It learned that AI can free journalists from tedious tasks and allow them to do the important work of real journalism that only humans can do, including investigative work.

“We are still relevant. Not only does AI need us for training, we can only use AI to the extent that it allows us to do what’s really important and that only we can do.”

About Sarah Schmidt

By continuing to browse or by clicking “ACCEPT,” you agree to the storing of cookies on your device to enhance your site experience. To learn more about how we use cookies, please see our privacy policy.