The New York Times publishes around 250 pieces of original journalism each day, ranging from stories on climate to gender to arts to investigative journalism. With 3.6 million paid digital subscribers as of Q1 2019, The Times is the most successful digital news organisation in the world, and the company has a unique lens for understanding what matters to the world, based on what people read.

To better understand its readers and what’s important to them, The Times built a machine learning-driven data tool called Readerscope. It reveals who is reading what and where when accessing The Times digitally. Originating from the company’s efforts to help expand readership outside of the United States, the tool uses anonymised data to visualise which topics certain audience segments are interested in at different time periods and the top articles that are being read on the site for context.

As brands seek relevance in the world, they are eager for new angles of insights into their consumers and what matters to them. Where previous data classifications — such as keywords and sections — were driven by what a publisher thinks are relevant categories, Readerscope’s topic model uses natural language processing to group articles that are alike. Indexing algorithms further connect these topics and articles to the most relevant audiences and locations.

The knowledge gained from the tool can be used by The Times’ in-house brand marketing agency, T Brand Studio, in developing creative content ideas that are relevant. The knowledge also is helpful in targeting direct media campaigns by zeroing in on a brand’s target audience segment, such as Millennial women, to understand what they are reading. And it can help brands find the right audience or geography for a certain message by topic — such as human rights, philanthropy, or travel — and see which audience segments over-index for interests around those subjects.

Data matters, and Readerscope is showcasing the depth of Times data by generating insights that continue to shed light on topics that matter to readers. For example, the tool revealed a distinction between topics that occur frequently in our coverage, such as sports scandals, sports business news, and sporting event results. On the audience, it found that younger men tend to read more about the scandals, and older men the scores.

These insights are actionable.

When a pharmaceutical company wanted to align with Times content, but was unsure where to target ads, Readerscope revealed that “wellness seekers” not only read health articles, but also significant amounts of science and home coverage. This guided the company to a diverse and performative set of placements.

The Times’ advertising business sees insights as the next frontier. To make products worth paying for, it has to understand its audience with extraordinary depth and turn those insights into engaging products. The Times is now building a proprietary targeting capability that will target media against articles classified in a given topic, enabling advertisers to target a topic like “sports scandals.” It is also engaging with more and more advertisers to ingest their first-party segments to provide meaningful intelligence about their consumer base using the tool. Readerscope is a launching pad for insights and targeting.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .