How The New York Times’ AI-driven data insight tool is informing ad campaigns

From the personal to the less personal

0
87
New York Times (streetscape of building)
New York Times (streetscape of building)

For publishers, AI-driven tools have largely been used for editorial purposes, to write articles on themes like sports scores, weather forecasts and real estate sales. For example, Reuters uses an AI-based tool called NewsTracer to sift through millions of tweets in real time, to flag potential news stories, for its journalists. UK-based Reach does something similar, using an AI-based tool called Krzana to monitor 60,000 online sources to alert journalists to breaking news, and Forbes has tested an AI-based tool to draft stories for contributors. Publishers have also leveraged artificial intelligence to power content recommendations, to edit homepages, and for translation.

But, publishers are finding new uses for algorithmic-based tools, like offering AI-driven data insights to their advertising and marketing partners and clients.

A few years ago, The New York Times launched Readerscope, an AI-driven tool that summarises what segments of The Times’ audience is reading, visualizing who is interested in which topics and where they are. The tool can be used for content strategy, for branded content, or campaigns, to help advertisers and marketers understand their desired audiences better.

Readerscope has two separate functionalities, according to Kendell Timmers, VP of advertising analytics at The New York Times. One function is that it targets topics, creating clusters of topics by processing a corpus of Times’ articles going back four years, she explained. This created approximately a hundred separate topic clusters which highlighted distinct content and storylines frequently published by the Times. “The second function of Readerscope is that it creates a way to properly index and surface the most interesting relationships between those topics and our audiences in our DMP, between those audiences and location, and between those audiences and directly to articles,” she said.

Kendell Timmers, vice president of advertising analytics at The New York Times
Kendell Timmers, vice president of advertising analytics at The New York Times

While the importance of personas have been central to marketing and advertising campaigns for decades, offering insight into what a target audience identifies with and captures their interest, AI-based data can offer additional insight to an extent that wasn’t possible before.

These days, advertisers and marketers are increasingly craving data insights to guide their content strategy, and this is where Readerscope is making a difference. “Any way that you can fill in that picture a little bit more, give a little more color to help them imagine who this person is, who is in their focus target group, is really helpful for them,” Timmers said.

Readerscope data gives advertisers and marketers a broader image of their target audience, honing in on areas of interest that might not always be apparent. The tool has broadened the scope of places advertisers are going within the Times site, Timmers said. “I think it’s a little bit of a crystal ball for an advertiser to be able to say, tell me about this group that I’m interested in, tell me things that I wouldn’t already know about them.”

Part of the beauty of the tool for the Times, Timmers explained, is that anybody within their advertising group can access this information. So, if a salesperson has a call to make on a client, they can check ahead of time and get some insights to share when they go and see their advertiser, as opposed to having to talk to the data group for a lengthy analysis or talking points.

Using Readerscope, a sales rep would be able to go to a potential advertising client and explain topics of interest for target audiences, but also where they’re from, according to the data.

“With the idea being, if somebody who’s really interested in fashion comes to our page, they’re still reading some of the election coverage or they’re still reading some of the opinion stuff. And some of the topics that come up, high up on the topic list are some of the more serious topics that maybe a fashion company wouldn’t have originally thought of targeting.”

Privacy

Tools like Readerscope and targeting methods like topic targeting are ways to avoid using personal or even anonymous browsing behavior, Timmers explained. What Readerscope and topic targeting are doing is, if an advertiser is targeting by topic, it targets just based on the article that is being read right now.

“We do take great care to make sure that there’s enough sample or enough people in each of these groups that we never get anywhere close to the ability to personally identify somebody,” Timmers said. “It is anonymous and aggregated to a point where there are thousands of cookies in each one, so there’s no way to get granular in a privacy invasive way.”

Readerscope as an analytical tool is showing you based on third party data, what people are reading, but it’s sort of allowing you to make that bridge from something that may be more personal, to something that is less personal.”

Reprinted from FIPP World newsletter.