The New York Times says it's getting ads to perform 40% better by targeting people based on emotion

Spencer Platt/Getty Images

The New York Times headquarters.

• The New York Times is expanding its emotion-based ad targeting after saying these ads outperform regular ads on its site.
• These ads are appealing to advertisers who are looking for new, better ways of contextual targeting as cookies go away in the age of privacy.
• But advertisers are skeptical that emotionally targeted ads will lead to better sales outcomes.
• Visit Business Insider's homepage for more stories.

For years, cookies were the basis of online ad targeting. But the rise of mobile and privacy measures like Google's and Apple's anti-tracking initiative is forcing marketers to find new ways to personalize ads to people.

A handful of publishers are trying to do this with sentiment-based targeting. The New York Times is at the forefront of this trend with Project Feels, a product that matches ads to people based on emotion, and now it's growing that portfolio of products.

"We thought we had a particularly big space to exploit because of the breadth and depth of our content, our really large audience, and ever-growing archive of content," said Allison Murphy, SVP of advertising innovation at the Times.

Read more: Apple's new single-sign-on feature has advertisers quaking over the marketing implications

Murphy said the Times has run more than 50 campaigns using Project Feels since introducing it in 2018, including ones for social responsibility and movie releases. According to Murphy, ads using this filter have gotten a 40% higher click-through rate than the average Times campaign.

To build Project Feels, the Times ran online surveys asking readers to read a Times article, then select from multiple choices which emotions it made them feel, like boredom, fear, hope, and sadness. Its data scientists then applied those results to similar articles, so advertisers could target their ads to or away from specific articles based on the emotions they do or don't want to be associated with.

Other publishers such as The Washington Post and Gannett's USA Today have built similar tools by scanning articles for keywords and tone and assigning sentiments to them accordingly.

Now the Times is expanding Project Feels by asking people not just how they feel but how likely they are to take certain actions after reading an article.

The Times is aggressively building its first-party data

The Times is also going deeper with its audience targeting. It has a tool called Readerscope that shows advertisers what topics audience groups like "business decision-makers" and "fashion enthusiasts" most care about. Then it lets advertisers target ads to those people across some 200 content topics using a tool it calls Topics Targeting.

The Times audience segments are built using data including how frequently Times readers visit a given section. But the Times wants to build its first-party data to make its targeting more accurate, by getting people to log in and also surveying them about their industry, salary range, and job level. The idea is that if the message is more relevant, it will work better.

The Times says it practices restraint in using people's data

Wait - isn't it creepy to exploit emotional and personal information for ad targeting purposes? After all, Facebook got a black eye when it had to deny news reports back in 2017 that it was ad-targeting users as young as 14 based on their emotions. And privacy is even more top of mind for companies today, with a wave of digital privacy concern sweeping the country.

• The emotion-based targeting isn't tied to individual readers.
• It won't let advertisers target people based on their political affiliation.
• It limits the data it lets advertisers collect on readers.
• In asking people about income, etc., it discloses that it'll use that information for advertising.

"We're forever practicing restraint in how far we would go," Murphy said.

Advertisers are skeptical of emotion-based targeting

Advertisers are intrigued by using emotions to target people as cookie-based targeting goes by the wayside. They're also skeptical that it can reach a big enough audience and sell more product, though.

Murphy said the Times is going to do brand lift studies this year to see if this targeting impacts how people perceive an advertiser. But she conceded it's hard to prove emotion-based targeting leads to business outcomes.

Michael Zacharski, CEO of EMX, the programmatic exchange of the agency Engine, said he's not buying advertising this way, but called it "interesting."

"Enhancing something like context and assigning emotion to it is brilliant in its simplicity, but will have to be proven out for buyers to see value and understand how it can scale," he said. "It should actually work with scale behind it in an understandable way that we can explain and measure."

Kristen Colonna, chief strategy officer at OMD, questioned whether emotion-based targeting can work as the Times describes it, given everything a consumer may be experiencing on a given day.

"Emotion is an extremely fluid thing," Colonna said. "One day I may be feeling carefree and open to new thinking and ideas, but the next day a negative news or personal event may have changed my receptivity."