Extra, extra bytes: How the USA TODAY Network is reinventing the news again, now with cloud
Matt A.V. Chaban
Senior Editor, Transform
Whether faster photo editing or personalized front pages, Gannett, the largest news publisher in the U.S., continues to innovate with the latest tech.
Forty-five seconds. Three-quarters of a minute. One lap around the track for the world’s fastest people.
That’s also how long it takes a typical photo editor at USA TODAY and its more than 200 publications across the USA TODAY Network to crop an image, ensuring it appears centered whether on a desktop, mobile browser, dozens of local and national apps, or even in print.
Or at least that’s how long it used to take.
“On traditional photos that editorial would use, they would have to go and manually crop those, find the center point of that photo to be able to then have that display correctly on all of our different products,” said Erik Bursch, the senior vice president for consumer product and engineering at Gannett, USA TODAY’s parent company. “It’s 45 seconds on every type of photo, but that adds up, when you're talking about the amount of content that we publish.”
With dozens of daily papers and websites publishing stories up to the minute, and republishing them to social media, all that image formatting accounts for thousands of hours of work each year.
In the fall of 2020 , Gannett began using a custom application built on Google Cloud’s Vision AI to automate the cropping process. This has freed up photo editors to focus on more high-value tasks like photo selection, multimedia production, and story development.
“When I look at technology right now, it's all about how we can understand the data, and the models, and create engines to then help our consumers, both internally and externally, either consume better or do their job faster,” Bursch said during our latest episode of the Transformation Debrief.
That new technology, combined with decades of journalistic tradition and innovation, are proving essential elements to Gannett’s ongoing success in one of the most challenging businesses there is.
Every enterprise is looking for an edge these days, especially with the pandemic, climate change, labor disruptions, and global uncertainty among the many issues weighing on the economy and corporate bottom lines. In the media industry, though, and the news business in particular, the challenges are both more acute and more long-term, given the obstacles presented by the world’s digital shift.
Uncertain times call for innovative measures
From the outset, media organizations have been digital pioneers. They were among the first enterprises to launch websites, embrace online video and social media, and reinvent subscription models and event businesses.
Monetization of all that influential work has been less straightforward, and competition has been fierce and pervasive.
In the past, your local paper would have shown up on the doorstep, along with a pile of magazines, and maybe you tuned into the evening news or talk radio. That was pretty much it. Now, not only is it possible to consume virtually every news outlet on the planet, you can also find information — and diversion — on all kinds of feeds.
“Our print competition is very traditional-media-space-based,” Bursch said. “Our digital competition, it’s certainly our friends at The New York Times and Washington Post, but it’s also TikTok, Instagram, Snap, YouTube Shorts, right? It's all in the spectrum of — where are people spending their time and their consumption?”
Gannett has certainly learned to compete. A series of strategic mergers over the past two decades helped grow its stable of 200 publications into the largest news publisher by circulation. Through its digital and print operations, it reaches more than 136 million unique readers a month.
“A conversation that we've talked a lot about, internally, is the attention span of users, which has really narrowed aggressively in the last couple years,” Bursch said. “While there are still a lot of folks that are ready to sit down and consume a bunch of words in a story with some photos and videos attached to it, there's a lot of consumers that are really interested in seeing that in a very quick snapshot, and covered in that short attention span.”
“It's been a very aggressive route for us, from a product perspective at Gannett, to serve these different audiences,” Bursch added. “We're looking to release several things this year that help answer that.”
The 'dead tree' business is alive in the clouds
The speed and scale required for such innovation can typically now be found in one place.
“It’s agility that businesses need to have today, to be able to transform to the different types of competition that might be out there, the different business models that you might need to employ,” Kirsten Kliphouse, president of the Americas for Google Cloud, said. “Cloud is one of those great examples of how we actually help organizations start to be able to create that agility, and also the scale.”
When put that way, it quickly becomes obvious why traditionally ad-supported media companies would be so enthusiastic about the cloud: They need to innovate as much as anyone. In some ways it’s ironic, publications have been so disrupted by technology — from sites replacing classified ads for jobs and apartments to the dominance of certain online ad platforms — that publishers really have to embody the mantra of “technology first.”
When people say that “every company is a tech company now,” that couldn’t be truer then when it comes to the news business. Whether it was the printing press, radio, newsreels, or television, and now the internet and social channels, newsmakers were always among the first to embrace new mediums (part of why they call it “the media”). The cloud is just the latest step in that evolution.
“We couldn't do what we do now, moving in a lot of different directions and scaling on breaking news, events, and otherwise, where we were — and even in our earlier cloud journeys,” Bursch said.
There’s the big, audacious implementations of cloud solutions, like AI-driven personalization; or empowering reporters with database research looking for patterns they’d never see otherwise, say in political donations or health code violations; or the aforementioned photo editing.
At the same time, the nuts and bolts of running a daily collection of sites for tens of millions of users is just as important. Simple, reliable performance has become a hallmark for every site within the USA TODAY Network. If it’s slow to load or a feature crashes, especially during a major news event like an election or natural disaster, it’s as easy as a click or swipe for users to go to a different news source, or a different diversion altogether. The news has to be there, where and when the audience wants it.
“We've really pushed aggressively, from a product perspective, on all things performance, and being the most performant media site in the U.S.,” Bursch said. “A lot of the metrics that we hold to, from both a web perspective and a native perspective on our products, we judge on that user experience, because I'm a big believer that performance relates to engagement. And we couldn't do that if we didn't have strong infrastructure.”
Three things matter: Personalization, personalization, personalization
If performance is table stakes, then personalization is the big bet.
One of the best ways to keep an audience engaged in a busy and boisterous media and entertainment landscape is to know the news they want as well or better than even they do. AI and machine-learning tools like Recommendations AI go a long way toward learning users' preferences overtime, whether through data collected on the digital publications or elsewhere.
“We have just cracked the tip of the iceberg, with how we would want to go and use a service like Recommendation AI,” Bursch said. “For our users, whether they're consuming one article and we want to lead them to the next that’s most relevant, or maybe everybody is getting recommended a full frontpage that’s just for them to use. That's certainly on the forefront of what we’re doing.”
Gannett is already developing services so that if a subscriber is traveling, even if they log into their familiar publications, like the AZcentral or North Jersey apps, they could see local stories from their destination in the feed. Or, if they’re a big sports fan, and the Phoenix Suns played a game against the Bucks in Milwaukee, they could get not only a write-up from the columnist for the Arizona Republic on their page or in their inbox but also one from the columnist at the Milwaukee Journal Sentinel and the basketball beat reporter at USA TODAY.
We couldn't do what we do now, moving in a lot of different directions and scaling on breaking news, events, and otherwise, where we were — and even in our earlier cloud journeys.
“The amount of content that we publish on a daily basis across our network is massive,” Bursch said. “We really want to know what the users want to consume. And that might not be what is necessarily one of the top stories on USA TODAY right now, at the given moment.”
Projects like this, both from an experimental and an infrastructural perspective, make cloud essential to what Gannett is trying to do. Without the scale, speed, security, and connections, these projects won’t be stable, and they won’t be possible.
“Everyone's sort of making the move to the cloud, and the last couple years it's accelerated for sure, but the opportunity yet ahead for cloud is even so much bigger when we look at where we're going to be in the next five years, with the majority of the business transacting in the cloud,” Kliphouse said. “It's really now about not the entry phase, just getting it there to the cloud, it's now — how do you leverage the information that you do have in the cloud to make different decisions, or innovate differently, or have new services or new products?”
Bursch and Gannett are as thrilled to have gotten there as for what comes next.
“We now have data, we have so much data to consume and analyze and deliver,” he said. “And we're now at that sort of second or third phase of saying, ‘Okay, how do I bring that together? How do I make it meaningful and actionable for ourselves internally, rather than just creating a massive time sink for different teams, to be able to try and dig for the right data?"
“It's going to be a fun time, as we progress down that journey.”