Digital Innovation Top Stories
4 mins read

Mood analysis can be a powerful revenue generator for publishers, NYT’s “Project Feels” shows

Does your state of mind impact your purchasing decisions?

Conventional wisdom suggests the answer is yes… you might click on certain kinds of ads when you’re happy, or be more prone to check out some different products when you’re sad, and even make some rash buying decisions while you’re a bit tipsy.

Now The New York Times is getting in on the action.

The Advertising team at The New York Times asked a question: could we accurately predict the emotions that are evoked by Times articles? If so, we could empower advertisers to place ads more relevant to the context in which they are shown.

To explore this idea, The Times’s Data Science team launched Project Feels, a project to understand and predict the emotional impact of Times articles.

Alexander Spangher, Data Scientist at the New York Times

The Times has over 4.3 million subscribers worldwide, with 3.3 million of those paying for its digital products. A vast data trove to run such a test on. And as the publication confirms, the experiment was quite a success.

“We built prediction algorithms with large amounts of data collected via crowdsourcing,” said Alexander. “Our predictions made sense qualitatively, and we ran successful experiments demonstrating that readers’ emotional response positively correlated with engagement on articles.”

According to the publisher, this approach is called perspective targeting, and was one of the first data products launched by nytDEMO, a new initiative aimed at helping advertisers place the right marketer stories with the right articles.

nytDEMO—which stands for data, engineering, measurement, and optimization—is a collaboration among members of The Times’s data, product & design, technology, and advertising groups.

“We’ve created nytDEMO, a first of its kind team, to transform our vast audience understanding into ad products and tools, inviting brands to tap into our knowledge to drive results,” says Allison Murphy, VP of Advertising Innovation at The New York Times.

“Project Feels” has now generated 50 campaigns, more than 30 million impressions and strong revenue results.

Rick Edmonds, Poynter’s media business analyst

“So it’s a hit,” declares Rick, “and the sort of advertising innovation based on data science, artificial intelligence and algorithm that the Times expects to scale up — and that other news organizations are exploring.”

“It’s an exciting frontier,” said Chris Wexler, SVP of media and analytics at the ad agency Cramer‑Krasselt, “because we’re looking for people who are open to our message, and emotional state is a key part of it.

“Demographics based targeting is better than no targeting. Behavior-based is better than demographics. If you put it in order of importance, mood is above behavioral.”

While this may be the next level of targeted advertising—raising concerns about its influence on editorial decision-making—The Times is quick to insist that “this is an advertising project and was done without coordination with the newsroom; its findings will never impact our news report or other editorial decisions.”

We see insights as the next frontier for The Times advertising business. We have to make a product that’s worth paying for, which means we have to know our audience with extraordinary depth, and turn those insights into engaging products.

Brands that work with us crave insights to guide their strategies, and The Times can play a critical role helping them understand what audiences care about now, and will care about next.

Allison Murphy, NYT’s VP of Advertising Innovation

The Times has also launched Readerscope, an AI-driven data insights tool that summarizes what The Times audience is reading using anonymized data to visualize who is interested in which topics and where they are.

According to the publisher, Readerscope can be used as a content strategy tool to develop creative ideas for branded content or campaigns by searching a brand’s target audience segment (e.g., millennial women) to understand what they’re reading, either as topics or as representative articles exemplifying those topics.

“These data products grew out of our internal innovations using machine learning to understand our readers better,” said Chris Wiggins, NYT’s Chief Data Scientist. “Project Feels and Readerscope are just two of the data-empowered algorithmic tools we’ve developed as data products that can help The Times’s advertisers.”

The good news is that this new ad technology isn’t dependent on dropping cookies or compromising reader privacy. While NYT insists the tech is benign, experience has taught us that every innovation brings with it both the good and the bad.

So while this new breakthrough is exciting, it’s also a cause for concern, since in an era when developing content to incite outrage has become almost an art form, it’s hard to believe that it will not be eventually weaponized by malevolent players.

But progress is inevitable. For now, it’s a welcome thought that publishers have yet another creative way of collaborating with advertisers to generate revenue, especially in a manner that doesn’t involve privacy violations.

Download WNIP’s comprehensive new report—50 Ways to Make Media Pay—an essential read for publishers looking at the multiple revenue opportunities available, whether it’s to reach new audiences or double down on existing super-users. The report is free and can be downloaded here.


Related posts