These days, it’s not just the traditional publishing community that has reason for leveraging the content syndication model. As more and more companies across vertical sectors themselves become content providers, syndication makes sense for them, too.
NewsCred has a new – and semantic – take on content syndication, with content partners ranging from Reuters to The Guardian to The Economist. Recently-added customers that leverage the service’s fully licensed text, image and video content include traditional publishers such as the New York Daily News (and NewsCred is in talks with it about becoming a content provider, too). But other recent customers point to the importance of quality content to the consumer and corporate brand market: For example, insurance provider Zurich recently signed on. NewsCred also just closed a deal with Johnson & Johnson to be a subscriber of its syndication services for content related to the health care products and pharmaceuticals space.
Brands, says NewsCred CEO Shafqat Islam, are responding to consumers getting smarter and more demanding. “They have so much access to information that brands are starting to realize they can’t just sell products or services anymore,” he says. “They need more authentic, engaging conversations with their customers and the best way to build these authentic relationships is with highly-engaging, trusted, high-quality content.”
Islam says he hadn’t expected that consumer and corporate brands would become such a big part of his company’s business a few years ago. A company like Zurich not that long ago wouldn’t think it needed its own content strategy team, but as such organizations start to make this shift, “ we are very well-positioned to serve this big sea change,” Islam says.
Here’s how: The technology platform NewsCred has built to help connect publishers and brands with the world’s best journalism – content licensed from 750+ publishers – is responsible for ingesting some half million articles a month. That’s a job that involves normalization to reconcile the different formats in which publishers provide content, and then, using its proprietary semantic and natural language processing technology, organize it all. “We want to understand semantically what every article is about. About 90 percent of contextualization is done using our algorithms,” says Islam, with humans taking on the rest of the curation work. “We think that makes for the best user experience,” he says.
In fact, for brands it may be even more important. For instance, while its algorithms can help weed out content that references a brand customer’s competitors, it’s helpful to have humans at NewsCred double-check some of that, even if the brands themselves have their own in-house curatorial teams, as they often do. NewsCred also has some tech proxies – looking at certain keywords or phrases – to help understand the sentiment of a piece of content that could be appropriate to a brand to use but that might also have some less positive connotations. Yet humans step in here, too. “I purposely didn’t say sentiment analysis [technology] because I don’t think the technology is necessarily there yet in terms of where we are on an individual article basis,” Islam says.
Its algorithms can go pretty deep, though, in terms of categorizing articles, extracting entities and adding metadata to the content. For instance, in another brand-space example, NewsCred is behind Orange Telecom’s World of Entertainment portal and mobile-optimized site, which required adding to its conceptual metadata store more structured information about movies. “We built the metadata layer to be almost infinitely malleable,” he says. It syndicates the content out to this and other customers through APIs. “ Why just dump content on the plates of customers? We deliver it through a powerful API so they can pull the content they need about the topic they are interested in,” he says.
When the company started a few years back, it thought about using OpenCalais for entity extraction, but decided to build its own technology in-house. “OpenCalais is great for generic use cases, but we deal with news content, so we needed domain-specific algorithms to work well in news,” Islam says. That said, he sees a lot of great stuff out there on the Web that NewsCred could leverage even more, such as Freebase. “For certain types of topics we do take advantage of some of the work they did to augment the data quality in our system,” he says.
When it comes to servicing the more traditional publishing community, NewsCred also has on its 2012 agenda doing more in terms of niche content focusing. “We’re building out collections that are at a slightly more granular level than the broader category. For example, say the category is lifestyle – we have a lot of them. But within lifestyle we can build a wine and food collection and get those blogs and publishers,” he says. “We already have a lot of them but we want to expand more, to higher-quality blogs.” Its recent deal with The New York Daily News was an example of going niche, in the sense that NewsCred is powering a sports site targeted very specifically to South Asians in the U.S.
Islam also is tracking developments such as the rNews standard. “There’s definitely, long-term, no doubt in my mind there is a really good fit. I think it’s going to become even more important as a lot of publishers try to look and see when is it worth doing it to get ROI,” he says. He thinks NewsCred may be able to step in and help there – “we think about doing [for them] all the semantic tagging and helping tag up with rNews-specific metadata for all our content providers and doing it on their behalf.”
Forbes, for example, gives NewsCred all its content – “what if we mark it up in rNews and give it back? If we can do that for all 800 providers we can get a whole industry tagged up and ready to go.”
That could be good for its business, but good for the industry at large, too. “If data is tagged then we can start seeing interesting implementations – does that improve search or [make for] more innovative front-end experiences?” Islam says. “Until things are tagged we won’t see those applications. So it’s about enabling some of that innovation to happen.”