Using Analytics to Write Stories That Actually Matter
November 12, 2014


Editor’s Note: Andrew Montalenti is the chief technology officer at Parse.ly, a company that offers an analytics system allowing digital publishers to measure digital readership and social media sharing.

The Gutenberg Press revolutionized the world’s information exchange, empowering both authors and readers. This invention enabled mass distribution of printed material and allowed independent publishers to create a free press, a cornerstone of our democracy. Publishing technology changed dramatically in the six centuries that followed, with the most rapid change coming in the web era of the last 15 years.

Publishing and distributing words at scale is no longer the challenge – quite the opposite. Today, content management systems such as WordPress and lightweight distribution platforms including Twitter make publishing to global audiences a snap.

Now digital publishers face a new technology challenge — digital audience development. This seems as intractable now as mass-scale print publishing seemed 600 years ago, and potentially just as transformative. Putting an audience – the reader – at the center of decision-making is a new possibility for publishers, one enabled in part by the growing availability of digital data about audiences. New possibilities are opening to those who understand the paradigm shift occurring from the old top-down publishing model to a newer, audience-centric model.

As the Internet accelerated content production and distribution, it also gave publishers a gift – an abundance of information about what topics hold their readers’ interest, where they discover content, how they share, and more. This first-party data is available directly from the source, the people who are reading and interacting with content. To put readers first is to prioritize the information that they give a publisher, both consciously and through their behavior online. This can inform decision-making across a publishing organization — in sales, product, and yes, even editorial.

What This Means for News

To see this data in action, consider a report we released at Parse.ly earlier this year. We discovered a trend that could help inform both product and editorial decisions for news organizations. Visitors who came to news sites on desktop computers did so during normal work hours, but tended to use their phone or tablet to read news articles well into the night and throughout the weekend. This suggests that the digital equivalent of a “weekend edition” product could be a valuable mobile or tablet experience for a publisher to provide readers.

Editorial guidelines can be just as thoughtful in the digital era as common-sense guidelines were in the print era, but now they can be backed by data, rather than gut instinct. For example, The Atlantic discovered the importance of headline keyword choice through Parse.ly data. It found that in terms of content discovery and performance, using the word “Republican” in a political headline is a much better choice than “GOP” because it generates more clicks. This is likely because the term “GOP” is more familiar to political insiders and, though synonymous with “Republican,” may be met with indifference or confusion by those not in the political loop. The Atlantic staff might never have realized this if they hadn’t analyzed their audience data.

In another discovery that was helpful to both editorial and analyst teams, The Globe & Mail learned that a well-written and well-packaged political column on a pertinent international issue (in this case, Scotland’s independence vote) can perform just as well online as less serious fare that is often seen as having a digital advantage. This kind of insight can influence future editorial decisions.

A Key Goal: Attracting Loyal Readers

Yet debate continues in many newsrooms over the value of audience data and how exactly it should be used. One topic to trigger much public discussion recently is the impact audience measurement systems and algorithms have on how people read and share news.

One narrative suggests that technology alone may become the primary driver behind how, when and where readers stumble upon their next read or share, as social networks such as Facebook and various mobile news apps are becoming important platforms for aggregating and curating the news. Some publishers fear that this technology will drive down the value of original content by promoting aggregation over originality.

Lost in this narrative, however, is the fact that news organizations also are using digital technologies to create meaningful information that is actually worth reading or sharing. Audience analytics platforms such as Parse.ly, in-house tools (such as The Guardian’s), and even technologies from legacy providers like Adobe Analytics help publishers put their audiences directly at the center of their publishing strategies.

Parse.ly’s software is built for publishers who continue to take pride in the creation of original content. It seeks to deliver information straight from the audience to every stakeholder within a news outlet, so they can each make informed decisions off the same data. This can drive revenue, quality and even morale throughout organizations that have been hit by tough market conditions in the web era.

We often hear from individual reporters and publishing teams that they feel like this is all driving them toward ceding editorial control to algorithms and analytics. With all the buzzwords around engagement, multi-platform distribution and “Big Data,” it can be difficult to maintain an editorial stance that isn’t in some way technology-driven. But data-informed decision-making is also about making judgments, not blindly following the latest real-time performance indicator. A news source becomes valuable – and sustainable – not by producing a steady stream of clicks, but by driving long-term reader loyalty and engagement. To obtain the latter two, which are key for any publisher that hopes to stay in business, analysis of online audience behavior is a necessity.

Respectfully capitalizing on the data that readers share with publishers is the only way that these organizations can compete against the online news curators, aggregators and filtering platforms that are siphoning mainstream consumer attention (think Prismatic, Flipboard, even Twitter and Facebook). At Parse.ly, we believe one key to success is consistent data used across an organization, so that the people in sales, product and analytics have the same access points into audience data as the editorial team does. This allows for shared insights into how to provide an enhanced experience for actual readers, from content characteristics (topic, length, publish time, etc.) to native apps to sponsored content.

Analytics Should Guide Meaningful Journalism

My personal belief is that technology shouldn’t crowd the role of publisher or journalist. Rather it should guide them by helping to drive a reader-centric vision. There has never been a time where readers have demanded a more direct relationship with their information providers, or where publishers are better positioned to deliver the goods in that relationship.

Ultimately, analytics platforms are only successful if an editorial team makes the cultural decision to see today’s available audience insights as an aid to their craft, rather than a hindrance. As with many things, attitude can be a huge deciding factor in success. The use of data in the news publishing industry is fast becoming “conventional wisdom” thanks, in part, to its prevalence in the leaked Innovation Report from The New York Times. The frustration felt by editorial teams is that audience data can also be abused, for example, to trivialize or gamify the relationship with readers. This is a valid concern that requires open and spirited intellectual debate, just like any editorial guideline. But the fact that some publishers abuse data does not imply that the data is not valuable, especially when deployed in the reader’s interest.

I empathize with journalists who feel that some media tech companies aggregate the output of their work to create short-term financial gains, often for people very distant from the hard journalistic efforts of storytelling and original reporting that goes into that work.

However, a new wave of tech companies is helping journalists charge headlong into this digital future. It’s in the best interest of technology companies to not just provide tools, as shiny and powerful as they may be, but also to guide stakeholders on how best to use those tools towards meaningful editorial objectives, helping to ensure the web has a healthy ecosystem of original journalism.  Because in the end, what good is a decentralized and free web if it can’t support the stories that actually matter?

The modern equivalent of the Gutenberg Press in the web era isn’t a content management system. The enabling technology for a new generation of digital publishers is not the ability to print, but the ability to understand – specifically, their audiences, with diverse interests and observable habits. What other discoveries might come from the data they readily share? How can publishers encourage deeper digital engagement with their work?

Coming to this understanding, in real-time and over time, can make all the difference to a publisher’s demise or survival.

Comments
  • GlobalEd

    Sooo… a keyword swap that brings more clicks (GOP < Republicans) turns that story into something that actually matters? There is a serious discrepancy between the narrative and examples in this article. Journalism this, journalism that and then it all comes down to getting the clicks.