Yeah, participation from any kind of passive teasers or pushed content is a lot lower than you'd think. If we launch a new project and post a teaser about it on our very homepage, it's a coup if we get a 10-15% click rate. Think that low clickage number is pretty standard across news sites, and well, sites in general.
Anyhow the part that's interesting with facebook is that there is no homepage on facebook. Everything's fairly temporal, so you only have a brief period of time to capture interest. On top of that, it's a hit/miss possibility that you'll start a chain reaction of interest, so I'm amazed when there's a 5:1 or 10:1 or 20:1 ratio of shares to pageviews. It means that facebook as a content platform is totally plausible, which I don't like.
Think my niche is programming in news. Something between a programmer, an editor and a reporter. I don't work on the core CMS because we have dedicated teams for that. What I specialize in is all the things the CMS can't do, won't do, or isn't designed for, like interactives. Or this http://projects.voanews.com/ebola-tracker/ ; the tracker data is updated manually because it has to be reviewed (sudo vi ebola/data) but an automated watch script alerts me when the World Health Organization updates their ebola stats every few epidemiological weeks. I like using RaphaelJS for making graphics.
Currently, working on 3-4 editorial projects. First one is a bit dry, a metrics dashboard done in Node.js that gets top pageviews from our articles and then scrapes titles, number of article comments, facebook shares, tweets. Some real surprises there - an article can have 10x number of fb shares than pageviews. As in, the teaser photo + summary is enough for people to share and not read in entirety. I like node. Also (probably unsurprisingly) but facebook is faster than our own sites for scrapes, even though we use akamai / CDN.
Next up is a parallax-y report on fourth of July for our Learning English division. It's a longform writeup laced with some cinemagraph-style looping videos and embedded quizzes explaining the constitution and a bunch of Americana to our international audience. The internal tool we generate these projects with (tool separates content and programming/design) are a mix of PHP / Smarty templating that I want to convert to node with some realtime collaboration features, but once baked the final reader-facing stuff is HTML/JavaScript. Looks kind of like this http://projects.voanews.com/central-african-republic-diamond... project. Funny thing is the internal tool is called "timeline editor" and it does everything except timelines. It should be called interactive editor.
Also, oh boy, I made that map in the article and am so proud of it. It remixed a bunch of complicated and overwhelming data points and simplified it for the reader. Had to do some lat/lng to pixel conversions, some point-in-polygon checks, got the CAR shape polygon list from a UC Davis site and then used inkscape to simplify the shapes because it had millions of points. So guess this kind of programming is a niche specialization for news agencies, and it's worked well for me, going on 7-8 years now.
Anyhow the part that's interesting with facebook is that there is no homepage on facebook. Everything's fairly temporal, so you only have a brief period of time to capture interest. On top of that, it's a hit/miss possibility that you'll start a chain reaction of interest, so I'm amazed when there's a 5:1 or 10:1 or 20:1 ratio of shares to pageviews. It means that facebook as a content platform is totally plausible, which I don't like.