If you read a preview or a recap of a game featuring Group of Five or FCS schools last weekend, there is a good chance it was written by a machine. “I wouldn’t say [automation] is ubiquitous, but it has become something on the road map of most” sportsbooks, sports media companies and sports publishers if they have not already implemented the capabilities, Data Skrive CEO Brad Weitz said. That is because in addition to helping companies save money, the technology is now boosting top lines. “By being able to produce useful content and more content with automation, we have been able to grow revenue,” said Barry Bedlan (global director of text and new markets products, Associated Press).
JWS’ Take: Historically speaking, media companies and publishers saw automation as a means of saving money. The logic was, “It is cost-prohibitive to cover every game or player. Let’s use machines to do that,” Weitz said.
The trend started in the financial reporting business, where companies sought to write stories about every stock’s earnings report. By the early 2010s, sports media companies and publishers began using automation to write game previews and recaps. Like business content, sports content “is very data-driven, making it ripe for automation if the goal is to tell more stories,” Bedlan said.
But early adopters quickly learned that automating sports content was far more challenging than the financial documents they were accustomed to summarizing, leading many—including one-time sports market leader Automated Insights—to leave the vertical in favor of more structured, formulaic types of writing (think: sales pipeline reporting). “If you think about a financial report, it is pretty predictable and pretty static,” Weitz said. “[Companies] report earnings every quarter. Their financials are pretty well-understood. The [stock price] goes up and down, and that’s what [the story] is.” By contrast, sports are unpredictable. Stories originate in “real time, [outcomes are] highly variable, and there is a lot of context required and storylines of interest,” he added.
Data Skrive, which recently closed on a $7 million Series A round, is trying to fill the void. The automated content company works with a host of media outlets, rights owners and publishers (see: Associated Press, ESPN, USA Today, NASCAR, Catena Media and Sports Illustrated) to not just save money, but to broaden their coverage base, optimize their content and ultimately drive new revenues.
For the AP, automation was never about a reduction in workforce (in fact, the company added employees to support its efforts). “For us, [automation] has been a sweetener,” Bedlan said. “We have contracts and licensing agreements that are contingent upon having a level of content that we did not have prior to implementing the technology. So, if we had not leveraged automation, we either would be spending a lot more, or we would risk losing revenue.”
When it comes to sports, the AP primarily uses automation to create “preview capsules” for every major U.S. sporting event (think: what to watch, injuries, betting lines)—content they did not previously offer but that has become critical in a post-PASPA environment. “Because of sports betting, the expectation is that not only will we have recaps and photos from every major sporting event, but we’ll have some sort of a preview laying out what is at stake,” Bedlan explained. Remember, the AP is both a B2C and a B2B company, licensing its content to media outlets for their platforms with the intention of growing advertising and affiliate revenues.
For reference purposes, Bedlan suggested the company produced about 10,000 automated previews within the last year.
The preview capsules proved valuable during the Oct. 2 Army-Ball State game. The matchup drew little national attention, but Data Skrive software wrote about it. So when Ball State wide receiver Justin Hall returned the opening kickoff 99 yards and search interest in the game spiked, Weitz said the AP’s clients who had broadened their coverage base with it “were able to capitalize” because the content existed and was indexed properly.
The NFL Draft provided automation users with another long-tail opportunity. While countless outlets focused on the players selected in the first round, companies using Data Skrive technology were able to have “a story written, published and indexed in Google and ranking within three minutes [of every selection being made] for all seven rounds,” Weitz said. Organic long-tail content can help to reduce companies’ reliance on an affiliate model that requires them to “pay a bounty for leads,” he added.
It is not difficult to envision automation helping newspapers that have cut the breadth of their high school and prep sports coverage pick it up again.
In addition to broadening the coverage base, automation technology helps sportsbooks and media companies effectively index content at scale. “In order to optimize you need to put the right ads in, link to other articles, promote it through SEO and ensure the call to action is above the fold,” Weitz said. “All of those things require more bodies than just [the person] writing the article. If [a company] is doing it for 10 articles, it is one thing. If [they are] doing it for 100,000 articles, it is very different.” Naturally, those bodies come with additional costs.
There seems to be little downside to using automation. So, it’s fair to wonder why holdouts remain. Weitz indicated that some legacy players operating outdated tech stacks are incapable of leveraging the technology. That could be changing, though. As he noted, “The Knight Foundation and Lenfest [Foundation] are starting to do grants to try and enable newsrooms to be more ready to take advantage of technology.”
It should be stated that the increased use of automation within sports media does not mean a machine is going to be writing JohnWallStreet anytime soon. “Automation can’t conduct an interview,” Weitz said. “Automation can’t do investigative journalism. It can’t opine.”