Posted by G. Coco
MCT SmartContent is the mastermind behind some serious brawn.
You won’t see it mentioned anywhere on this global news leader’s digital products, but MCT SmartContent is powering their categorization.
That means that MCT SmartContent is sorting and ordering stories so that readers and their own interests come first.
Why did this heavyweight pick SmartContent? With a century of collective experience behind them, editors at MCT SmartContent know content.
We collect, classify and distribute about 2 million stories a year. Our stories come from nearly 600 trusted U.S. and global news. MCT SmartContent’s ontology, or classification system, has more than 18,000 topics — enough to serve a global news provider, backed by years of editorial expertise.
But MCT is not just for the big guys.
The good news is that you, too, can syndicate MCT SmartContent’s topical U.S. and world news — or allow MCT to process your own custom content through our ontology and feed it back to you via RSS or FTP with all the metadata, including topics, that you’ll ever need.
Don’t take our word for it: Go to www.mctsmartcontent.com and build your own preview feeds to see what you might be able to offer you and your customers.
Posted by Michael Airhart
Much of the conversation about smart news content these days dances around two obvious questions: What does smart news actually look like? How do I use it?
The programming and editorial skill may involve a bit of rocket science, but for the end user, smart news content is really quite simple.
Let’s use the following local news story as an example:
Every day, McClatchy Tribune scans thousands of stories such as this one for relevance to specific topics. After scanning the story, we rank each potential story topic on a confidence scale from zero to 1.00. Finally, we deliver the story to SmartContent clients along with the story’s topical metadata.
Within the metadata, we include the confidence score for each topic. Any score above 0.48 meant there was modest confidence that the story is relevant to a given topic; scores above 0.60 reflected high level of confidence.
Here’s how we ranked this particular story by topic:
|Labor unions and disputes||0.69|
These prime topics are intended to aid the user in grouping the story with similar, topically focused content. If numbers make your eyes glaze over, here’s what they mean for the relative strength of each topic choice:
But MCT didn’t stop there. For each story, McClatchy Tribune also provides the user with an array of related and parent topics and their confidence scores, so that users may optionally link the story to a wider variety of related articles. Here’s how we scored the story for parent and related topics:
|Animal health and medicine||0.55|
|Careers and workplaces||0.69|
|Economy, business and finance||0.55|
|Health and medical||0.65|
|Labor unions and disputes||0.69|
|Restaurants, bars and food||0.94|
Based on these confidence scores, MCT SmartContent customers grouped this story primarily with their other content relating to food safety — or if their news library lacks a food-safety section, then with stories relating more broadly to agriculture, food, veterinary medicine, or labor unions.
Additional MCT metadata informed the user that the story relates to a government agency, the U.S. Department of Agriculture; and that the story relates to the Petaluma and Santa Rosa, Calif., geographic areas. With that additional information, SmartContent customers were able to target regulatory and Northern California markets.
By constraining the article’s potential audience, contextual metadata increased the article’s value and its likelihood of reaching those who most want to read it, while sparing uninterested readers the chore of sifting through unwanted stories.
And by using the metadata to optimize relevance, clients boosted the visibility of the content in semantic searches.
In upcoming blog posts, we discuss in further detail how SmartContent clients use smart content to optimize both contextual relevance and search-engine page rank.