classification

RECENT POSTS

Developing categories

Image courtesy of Flickr user OZinOH.

We put a metric ton of thought into the best ways of organizing content on the Argo sites, considering a vast variety of approaches and taking into account the content plans and user persona information that stations provided us. We knew early on as we started talking through this material that we generally wanted to target several discrete groups of users who’d be interested in different dimensions of the topic.

Many topics, for example, have what I’d call an industry community – people whose livelihoods are directly related to the subject. Many also have an academic community – a crowd of wonks and researchers studying related matters. Then there’s the legislative and political community – the folks making and advocating for policies. And of course, what I’d call the cultural community – those whose interest in the topic takes a more personal or social bent. This isn’t one-size-fits-all – these communities break down in different ways for different topics, and some topics have an altogether different set of interests. But as we realized we had these diverse audience needs, we decided we should keep high-level navigation that addressed them. Hence, categories.

I tell this story because I think these are the first questions our bloggers should answer as they develop their content plans for their site: Who are the four or five main audiences that you’d like to convene on the topic, and what dimensions of the topic are they most interested in?

The answer to these questions will form the spine of a solid content plan. Every day, we’ll want to be producing and curating content and sparking discussions that target each of these audiences. Each week, we should be planning at least one post intended for some viral pickup among these communities. And these four or five audience needs will drive the high-level navigation for the site, e.g. “Business,” “Politics,” “Research,” “Culture.”

How to choose your taxonomy terms

I’ve written before about the reasons for being thoughtful about classification on our sites. The Argo editors have taken a couple of stabs at sketching out what their site taxonomies might look like, and now we’re asking for a starter list of real terms – at least 30 – that we’ll use to create the initial set of topic pages for each site. I thought I’d write up some thoughts on how to develop that list.

Each term should reflect an audience interest.

In the Argo universe, every term we apply to content will generate a topic page. So as a baseline filter for what constitutes a good term, ask yourself whether you think your users would value seeing a page of posts about this subject.

I stress this point because we sometimes start out by structuring our sites around the subject rather than the audience. If we approached global health or the cruise industry from a subject-oriented perspective, for example, we might create topic pages for “Diseases and conditions,” or “Types of cruises.” But an audience-oriented approach might cause us to question that decision. Is there a genuine audience interest in reading content about “Diseases and conditions” or “Types of cruises”? We can imagine audiences would want to read about specific diseases or cruise types, so a page with posts about malaria or senior cruises would make a lot of sense. But “Types of cruises” makes more sense as a collection of topic pages rather than as a collection of posts.

That said, it will be useful to capture concepts such as “Types of cruises” that describe a class of topic pages. We might create a page or module like this, for example, to serve as navigation to different topic pages. Or we might use that information to surface related topics: if you’re reading about malaria, perhaps you’d be interested in reading more about TB.

Start with the low-hanging fruit.

The easiest thing to wrap our minds around in the taxonomy universe is topic pages. And the easiest topic pages to build are pages for entities – individuals and organizations – specific, proper nouns that an algorithm could search for and bring back relevant results. So I’d start with those. Who will you probably cover most frequently? The Gates Foundation will undoubtedly loom large on a site covering global health in Seattle. A site about the changing face of DC will probably feature many items about the city’s polarizing chancellor of public schools, Michelle Rhee. Generate a list of names like these that are central to the topic at hand.

Think about your various audiences.

With many of our topics, we want our community to include a few distinct types of audiences for the topic. In several cases, there’s:

  • an audience of industry insiders whose livelihoods are directly tied to the topic,
  • a governmental, activist or legal audience around the topic steeped in the relevant legislative issues,
  • a coterie of scientists and researchers who’ve studied various dimensions of the topic,
  • and a broad universe of people potentially affected by the topic, many of whom might not regularly seek out related subject matter, even though it concerns them.

Thinking about the topic this way might help generate ideas for categories – broader types of content bundles that require human judgment to define. For global health, these audiences might suggest categories such as “The business of global health,” “Law and politics,” “Science and research,” and “The impact of global health.” All the Argo editors have developed sets of personas typifying the different users who might arrive at the site. Feel free to revisit these personas as you develop your list of topics.

Consider your content plan.

Another set of terms might relate to how you program the blog. Tara Parker-Pope’s Well blog, for example, has several series of posts on topics like diet and nutrition (“Eat Well”), running (“Run Well”) and pets (“Well Pets”). These posts are collected into topic pages. Make sure to include these types of regular features in your list of terms.

Aim for relevance, not breadth.

We need to start off with these terms before the bloggers are hired, because they’ll help us determine how to build a topic infrastructure that can accommodate different types of terms. But we expect that these lists will change and develop a lot when your blogger comes on board. Your goal right now is not to capture every term that might touch on your topic’s universe, however tangentially. Right now, limit yourself to topics you expect will be most relevant to your coverage from day one. Keep 30 terms in mind as a minimum number of terms, but if you find yourself straining beyond that, don’t fret.

Continue reading

What a great taxonomy could do for you

The Argo team talks about taxonomies. A lot.

This is an unusual thing. Until very recently, saying the word “taxonomy” in a newsroom would elicit looks suggesting you’ve just spoken an exotic strain of Urdu. As topic pages have caught on with news sites, the word’s become a bit less rare. But it’s still strange to find a team in a news organization that so relishes some of the geekier bits of library science.

All the Argo stations have been thinking of taxonomies from the beginning – lists of categories and sub-categories were part of the initial topic proposals. And since, we’ve kept asking them to refine and extend their thinking about how their site’s content will be classified. Some folks might start to wonder about our obsession with descriptive metadata. So here’s an attempt to explain (rationalize?) that obsession.

At minimum, taxonomy will power one important component of all the Argo sites – the aforementioned topic pages. I don’t want these to be afterthoughts. I’m on record as saying that every time a Web surfer runs across a topic page that’s essentially a gussied-up Google search, an angel loses her wings. The topic pages should reflect the thought we’ve put into creating collections of content that will be valuable to users. And they should hook users coming from search engines into serendipitous streams of relevant related content.

But a beautiful classification schema holds promise far beyond topic pages and search engine optimization. Continue reading