I’d like to rant about something only spuriously connected to topics my group covers (finding the single version of truth among document versions and data, threaded discussions, tagging and hyperlinking, searching).
Every news story is a thread. It connects backward through history and will be updated with new information over time. Most news websites ignore threading and treat each story as a self-contained entity, frozen in time. At best, one gets links to a tag that yields a bucket of other stories related to a broad tag like “technology” or “Africa”. Consider a narrow story that will only produce a few articles, such as a missing boy story on CNN, a lawsuit involving a celebrity or official accused of doing something they deny, or an amazing breakthrough in cancer research that promises to yield new treatments. You often read the “introductory news” story, think to yourself you’d like to know how that turns out, and then never see the resolution because the next articles in the thread escape your attention. I suspect news outlets are partly to blame since they see the accusation or possibility itself as the story and then don’t care about the ending. After all, a possibility is what gets people talking around the water cooler. But for me, I’d like to just subscribe to that story and see the resolution when the ending is known.
When it relates to a titillating accusation that turns out to be false, this loose thread causes a great deal of harm. Studies of retractions have shown that almost no one sees or remembers the retraction, leaving the introductory news story as the sole impression on readers.
If news sources provide metadata-based ways to subscribe to threads they have a better chance of keeping readers who may otherwise set up a Google alert using keywords that will catch stories in competing publications. This is a little more work, but not a lot. Journalists have to know the last story their publication published on their topic and would have no trouble pointing to it. The resulting timeline of articles on a story thread would be trivial to track and allow users to navigate with a little metadata.
This doesn’t just apply to current stories that I want to follow in the future, but also to stories in the past that come up in searches with no connection to their storyline. If I find an introductory story that’s 5 years old, how do I find out what happened next? My only chance is to wade through the noise in a full text search and hope I find it.
The closest I’ve found to this is The New York Times website. They show the historical thread as “past coverage” and the broader tags as “related searches”. The only omission is that one can subscribe to tags, but not the story thread, and that means they do not show up in “My NTY” as an increasingly personalized set of stories one is interested in. Here’s what it looks like – this is the bottom of a story called “Corzine Proposes Steep Rise in Tolls“.
So, news organizations:
- Understand that stories are threads
- Add metatags that help your journalists and the readers to follow the threads over time
- Enable searches to find the most recent update on a story, not just a list based on a relevance calculation of how often the search terms were used in each article or a list sorted in reverse by time but including all the noise a full-text search tends to include
- Do periodic audits on loose threads to identify stories that need to be completed. Don’t leave readers hanging with just an introductory article written based on their potential interest
- Allow RSS/ATOM subscriptions to stories (specific stories, not category tags) so readers can follow them. Maybe readers will be less likely to cancel a subscription when they know how many story threads of personal interest will be cut. And they will be more likely to return to your publication to find out what happens rather than finding its continuance in your competitor’s publication.
- A side benefit of connecting articles into a story thread is that if a mistake is made (such as giving the wrong name of someone convicted of something – oops), you have a better chance to properly informing your readers and minimizing the damage (to the innocent person and your company in the inevitable lawsuit) if you make it very easy for readers to find follow-up stories which may contain corrections
Michael Sampson pointed to an interesting article in the Venture Chronicles on The Future of RSS which hones in on a fault of RSS (or “opportunity” depending on if you’re a glass half-full kinda person). According to Jeff Nolan:
Basically the entire RSS market has been built around a use mode of subscribe-then-read, and that is likely to continue as an exclusive model for many users or in parallel to other use modes. The weakness in this approach is that you only know what you know, as in you have know about a feed before you can subscribe to it… and I generally work off the approach that it’s far more likely that the best content on any keyword is not necessarily found in my OPML.
There are an increasing array of companies that are working on a next generation of feed consumption use model, built not around the explicit subscribing of feeds and chronological consumption of content. In order for RSS to get to the next level of mainstreaming we have to think in terms of behavioral filtering of content and discovery of new content sources based on explicit preferences or inferred preferences derived from behaviors.
I’ll second that. While I think RSS can (although not always) be better than manual methods for reading through a lot of information, it’s not the silver bullet for attention management. I often use it as an example of something that can fall out of an enterprise attention management gap analysis, but it’s just one example and piece of a much larger puzzle.
People can use RSS readers to narrow down their view of what news channels they will pay attention to and ignore the rest. Even if someone follows 200 feeds, at some point that list will become stale. While you’d probably notice something outside of your feed set due to the magic of linking (someone you follow must be smart enough to notice things outside your periphery, especially people that do link and quote-heavy blogs), at some point new centers of gravity can emerge that go unnoticed for too long. It’s like picking your set of friends and then never going to parties to meet new ones. Or just listening to music your friends recommend without ever listening to the radio to see if you’re missing anything.
I like the idea of leveraging more of the EAM architecture by adding rules, filtering, profiles, and proactive discovery to the RSS model rather than using it “as-is”. I hope lots of vendors and users start experimenting with this and working the kinks out (decreasing type I and type II errors) so that in five years or so even late adopting organizations can start benefiting from this technology.
Well, it’s that time of the year when the top 10 lists take over the front pages. Those of you who read this blog regularly (yes, both of you) know that I tend to focus on communication, collaboration, and content technology and, sure enough, I’ll be bringing this all around to that at the end.
A quick scan shows that Time magazine published 50 (fifty!) top 10 lists here: 50 Top 10 Lists of 2007. Hmmm – that’s just a categorized top 500 list, isn’t it? I don’t have time to get through that much – let me know if they publish a “top 10 ’top 10 lists’ ” and I’ll take a look. Wired’s homepage today has The Top 10 Heartbreaking Gadgets of 2007, The 10 Best Gadget Ads of 2007, and Top 10 Scientific Breakthroughs of 2007. Perhaps the best of all is The Onion’s What the Hell Just Happened?
I think there are two kinds of readers who enjoy these lists. The first is people that follow the subject in question and want to see how the author’s list (supposedly some kind of expert) jibes with theirs to validate their views, give them something to gripe about, or point out a few things they may have missed. Movie buffs love to see the “top 10 films of the year” list to see if they should brag to their friends about their good taste or slam the critic as obviously out of touch.
I want to focus on the second kind of reader doesn’t follow this subject and likes the top 10 lists because it provides a year’s worth of news in a handy capsule. For these readers the top 10 list acts as a filter to all the noise that occurs during the year. If you are stuck in with kids all day and don’t get out to the movies, the list is a handy way to fill up your Netflix queue for next year (after a 6 month lag or so for the DVD to come out).
Now, wouldn’t it be handy if, rather than once a year, that filter was always in place? I could subscribe to this filter and instruct it to alert me only when a top-10-worthy film, or classical CD, or news story comes out? And to remove the noise by not bothering me with the lesser films, CDs, or news in the meantime? It’s hard to guess what will exactly equal 10 by the end of the year, but I’d accept say 15-25 items and a dial to increase or decrease the sensitivity if I’m getting too many/few each year.
I’m bringing this up because I see the “top 10 list” phenomenon as a good analogy to what a slew of technologies at the intersection of portals, RSS, and social software are trying to do: filter out all the noise and just bring me the important information, encapsulated, all in one handy spot. It is a commonly recognizable form of attention management.
The process for assembling this is the same whether it’s Time coming up with a top 10 list, a blogger filtering news to find just the important stuff worth posting about, or the rules engine for an enterprise attention management system that is trying to find important events and pull them forward into the user’s focus. The process consists of:
Integration: Connecting up with all the event streams, information sources, and data
Categorization: Determining what subject the event falls into
Rating: Prioritizing this bit of news. This is probably the toughest part of the process at the moment, but attempts have been made in the form of social ratings engines (Digg) and attention profiling (APML).
Personalization: Lining up the category against the set of subjects that you are personally interested in, either through explicit declaration or implicitly.
Display: A UI that presents the user with capsules on each of the items and allows the user to notice, track, and manage the information
This process is even more important in the enterprise, where the stakes are higher than missing a good opera CD. How do you create your own “competitive news critic”, “financial event critic”, or “sales critic” to pick the most important information for you and how do they encapsulate this information and display it for you? It could be the head of each of these departments flagging important news and alerting others to it (hopefully not just through email). It could be through social ratings of important events. It could be through automated alerting mechanisms that work off of triggers or rules. No matter how it’s done, having an enterprise Roger Ebert to pick the best (and worst) as it happens and a good display channel (like Roger Ebert’s newspaper column) to present the information is as useful in a noisy enterprise environment as it is in a noisy entertainment environment.
With everyone focusing on top 10 lists, I’m hoping this “angle” helps an evangelist for RSS, portals, social software, or attention management to make their case in a way that will resonate with business partners and executives during the New Year’s season.
Happy New Year!
At the presentation by Sam Weber of KnowNow at our Catalyst conference, I was struck by the similarities between the drivers of RSS and portals. I’m new to RSS, but have covered the portal market since its inception, so those drivers sounded awfully familiar.
At a high level, the driver is the same: reducing information overload by centralizing access to information from dozens or hundreds of sources. So RSS and Atom are doing for content what portals did (or promised to do) for applications – aggregating a diverse set of information and providing a common place to look for it. But where portals worked on a screen real-estate metaphor and unified applications, XML sydnicators aggregate news items as lists of headlines. This is a more efficient way of keeping appraised of all the news content one is interested in than gluing rectangular windows from a content provider’s site into a new page. And of course, many applications are now producing XML feeds to alert readers to updated information they should view rather than strictly news stories.
One difference is that portals tend to include some form of personalization model that can expose users to applications and content they didn’t know existed. Users of XML syndication expect to be exposed only to sources they have explicitly subscribed to.
What comes out of this for me is a design criteria for use of XML syndication within portals:
Expose news-type content through an RSS/Atom portlet within a portal (rather than links or screen scrapes of the web sites of news sources) and non news-type content (content without headlines or requiring multi-line presentation) and applications through separate portlets.
When I spoke at the Catalyst conference I mentioned 3 technologies that I consider to be the most important for enterprise attention management. I wanted to take this opportunity to elaborate a little bit.
By “most important” I mean they offer the greatest opportunity to improve the attentional characteristics of enterprises since they have experienced recent advances in their capabilities that many organizations have not taken advantage of yet. Therefore they can contribute significantly to improving an organization’s ability to improve the efficiency and decision making quality of its information workers.
- XML Syndication: Sam Weber of KnowNow spoke on RSS right after my EAM presentation and made it clear how RSS was developed as a response to the information overload problem that EAM addresses. In my EAM model, RSS is a textbook example of an attentional technology for pulling desired information forward. This technology is here today, so overloaded organizations have no basis on which to throw their hands up and say nothing can be done.
- Presence: I see presence as being an important opportunity for three reasons. First, getting information workers to honestly and accurately use their presence indicators can help sigificantly with interruptions, on which I’ve written before (see an assembly of my interruption science entries here). Second, if one adds a discretionary function on top of presence that can make it message aware (not just am I unavailable to receive a message, but am I unavailable to receive a specific message from a given sender at that time), it could represent a major leap forward in enterprise attention management. That’s a big “if”. It requires someone to build the rules on top of the presence engine (probably not part of the presence engine itself, even in “rich presence” incarnations). And it could run afowl of all the problems associated with rules (e.g.,false positives).
- Search: While search has been around a long time, there has been a lot of forward movement recently on enterprise search, which can now aggregate search across many more content and data sources as well as include social searching capabilities. Improvements in search, particularly when used as saved searches, can help pull information forward that would otherwise be lost in the noise.
If you haven’t seen the Times Reader and are interested in news readers you should take a look at it. (Note: You’ll need .NET 3.0.)
There are several good ideas in here that result in a very readable experience. I like the leverage they get out of the metatagged content, such as a “pictures in the news” view (which does a slideshow of the pictures from each article but hyperlinks to the text if you click to learn more about the picture), a custom-assembled page around a metatag, and the “all navigation” view that shows each section of the paper with light/dark grey boxes. That last features allows the reader to see at a glance which stories they didn’t get to and hover over them for a quick preview.
I found the font very readable and the full screen view allowed me to focus on just the paper and not distracting OS menus and browser interface buttons.
Still, there are a few issues:
- No personalization
- No community
- An insular content experience (pure NY Times content with only a search to external content) unlike the doomed Encyclopedia Britannica which provided its own content in context with other content
- Some random bugs in the interface and syncing
All in all, a good view into a forward-looking news reader.
May 2013 M T W T F S S « Aug 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
Analyst biz Attention Management BEA Blogs Browsers business case collaboration communication Composite Applications Content Management Economics email Enterprise 2.0 Fun Gaming Globalization Google Governance IBM Information Work interruption science Intranet knowledge management Lotusphere2008 Microsoft Microsoft SharePoint Mobile and pervasive computing Office Oracle portals presence Recession RSS social software Uncategorized usability User experience virtual worlds Web 2.0 XML Syndication