Blog Post

The NYT is doing something smart by using Twitter trends to target ads

We spend a fair amount of time criticizing traditional media players like the New York Times for not being innovative enough, relying too much on paywalls for revenue growth, and other perceived failings — so I think it’s only fair when we recognize things that they do that are innovative and interesting, and a new ad product it has launched definitely falls into that category. In a nutshell, the newspaper is selling advertisers the ability to target their ads based on whether a story is trending on Twitter.

Michael Zimbalist heads up the NYT’s research and development arm, and described the new advertising venture — which is called “Sparking Stories” — at a conference on big data put on recently by According to Zimbalist, it came out of a new unit of the NYT R&D lab that is designed to commercialize the research done there, and is based on a social-media analytics tool the Times has been using for some time called Cascade.


In a nutshell, Cascade is the newspaper’s in-house version of Twitter’s trending topics: it allows the NYT to track which stories on its site are being shared heavily through social media, information that is used by both editors and reporters as a way of seeing how much traction their content is getting with readers. The new “Sparking Stories” advertising product gives advertisers the ability to place their ads inside those specific stories. Said Zimbalist:

“We developed this tool that lets us see what stories are trending, so now we’ve created an API that lets our ad server talk to this tool, and we’ve created an ad product called Sparking Stories, where an advertiser can come in and buy a package of stories that are trending right now on Twitter, irrespective of section or context or topic — just the stories that are really breaking through right now on social media.”

New York Times Building Day

Zimbalist notes in his interview with an analyst from Forrester Research that the original head of the NYT’s digital arm, Martin Nisenholtz, had an advertising background before he joined the newspaper, and says he “realized the promise of this medium was going to be targeting based on data,” something the paper has always done through fairly traditional means: namely, separating registered readers based on gender, zip code, age, etc.

But what Spark allows the newspaper to do is much more interesting: it focuses not on age or gender or zip code — things that may be irrelevant (or at least less relevant) for some advertisers who want to target specific content topics — and instead focuses on what content is being talked about the most. Although Zimbalist doesn’t say what the paper charges for this product, that ability could theoretically be much more valuable than a standard display ad.

At a time when the advertising business is undergoing a tremendous upheaval, forcing media companies to experiment with everything from affiliate links to sponsored content, it’s nice to see the Times focusing on how it can use technology and data to push traditional advertising into the future a little bit. (Note: We are going to be discussing new forms of monetization for content at our paidContent Live conference in New York on April 17).

Images courtesy of Shutterstock / noporn and Flickr user Robert Scoble

5 Responses to “The NYT is doing something smart by using Twitter trends to target ads”

  1. Harry Hawk

    Someone asked if Sparking Stories equated to News-Jacking.. Here is my answer.

    I don’t. This is simply a real time tool to sell real time ads, some buying models require rapid and high volume of both reach and/or frequency and this would provide that (at a price). However, this could hurt those who buy based on equitable rotation.

    If there is enough transparency within the buying process, such that a buyer could see what exact stories are trending and if creative could also be designed on the fly and run on a story by story basis, then this tool could be used to engage in news-jacking but the topic of the story would have to relate to the ad (e.g., a black out and then Oreo’s: you can dunk in the dark).

    The implication of the story is that you are buying a bundle of what’s trending and there is no indication that you will be told which stories those are nor any promise of how much or little the bundle will change throughout the day/night.

    If stories about ladies shoes, college tuition and grass fed beef are trending and you use Sparking Stories to buy those stories via a bundle and promote hot water heaters, that is not news-jacking IMHO.

    My primary area of academic research was in the area of future of advertising and new media infrastructure (circa 1993).This is exactly the type of real time ad market that was envisioned, which will allow marketing and other professional communicators bid for higher valued “eye balls” and allow media properties to escape from the lock-in of an annual rate card.

    Conversely you could use other tools to guess at what is trending and then craft creative to news-jack that topic(s) then bid on/buy ads through Sparkling Stories and hope/pray that you guessed right.

    I can see the NYT selling this in a variety of ways, A) Way one is you lock in to a bundle of stories that is fixed for the buy.B) A second way would be to buy a bundle that changes over time to maintain a specific level of “trend” (e.g., engaged viewership). C) Finally, they could sell ads on a story by story basis. Given A, B, & C, A would be the most affordable and C the most expensive.

    I assume they are not also selling ads on a equitable distribution model which would assume that every ad that is running within a section or in the totality of the paper get’s an equitable share of the trending and non trending stories. That is the real issue here. If you are running for two weeks in NYT online, in a political section (because it’s close to the election) — basically you buy an equitable rotation of section at a time you hope will be busy and trending, Etc. and your “best” stories could be sold away via Sparkling Stories.

  2. This product, when scaled out to the Internet, will be a multi-billion dollar business (target your ad to what people are reading NOW). Today, there are three players who hold this information: Google, Facebook and Twitter.

    The challenge is how much of this would impact buying behavior. That is where Pinterest scores. But the numbers for Pinterest are rather sad.

    Which is why, like many things, this remains a tough puzzle.