How, and where, to hyperlink within a news story

Late last year, I urged OJR readers not to forget the value of hyperlinking, to look for opportunities to link their stories to supporting information elsewhere on the Web. Today, I’d like to continue with that topic and write a bit about the thought process behind link content, that is, the decision about which text within an article to link, and to where.

Let me start by writing that I am not going to address issues about the design of contextual hyperlinks — issues such as whether to underline a link, what color to make it, etc. For those questions, I direct you to Jakob’s Nielsen’s outstanding Guidelines for Visualizing Links, which includes his usability guidelines for showing textual links.

How to decide what to link? At many news organizations these days, the decision is punted to a computer algorithm, which inserts links automatically into a story wherever it encounters a noun in the publishing system’s database of keywords or phrases. My wife passed me one such example which recently annoyed her.

It came from a Feb. 3 story in the Washington Post: “Sleepless in Amsterdam (And Munich),” about the orchestra conductor Mariss Jansons.

The offending link came at the end of the story, in a paragraph referencing the Pittsburgh Symphony’s final concerts with Jansons, in Carnegie Hall.

The hyperlink does not link to a review of these concerts, as the context of the sentence might imply. Nor does it link to Carnegie Hall’s website. Instead, it links to an automatically generated page of content about Carnegie Hall, from the Post’s site and elsewhere.

And that page is woefully thin, claiming, for example, “No results” for Carnegie Hall video or audio on the Web.

Oh, really?

The Post is hardly alone in relying upon an automated solution for hyperlinking. Indeed, this method is often used by webmasters to build massive numbers of inbound links to their pages on specific keywords, in attempt to vault those pages to the top of Google’s search engine results pages for those words and phrases. But littering your article pages with hyperlinks for search engine optimization purposes can cost a site audience share if it leaves its readers frustrated after they follow worthless links.

If you’re going to opt for automatically inserted links, you must revisit your algorithm from time to time, instead of adopting a Ron Popeil “set it and forget it” approach. Otherwise, you end up with a system like Yahoo’s, which today linked the first reference to the country “China” in a story about the upcoming Beijing Olympics to… a page listing China’s medal count from the most recent Olympics, in 2006. In Turin, Italy.

The three principle reasons for hyperlinking within an article are to cite an attribution, to provide context for an article, and to reward readers with an “easter egg.”

1. Attribution

Examples of this type of hyperlinking would be the court records I referenced in my Sept. 2007 piece for OJR, Don’t forget the value of hyperlinking, as well as research papers, census databases and other pages on the Web. (Like how I just linked my previous OJR story.)

Attribution hyperlinks provide readers with the opportunity to delve into original source information, at a deeper level than the writer provided in his or her article. That not only rewards the curious reader, it helps a writer and his or her news organization build credibility with a skeptical audience.

2. Context

Ideally, a news story will provide within its copy the context that a reader needs to understand the piece. But some stories are so complex that an author will not want to risk testing the patience of loyal readers by rehashing basic info they already know. In these cases, a link to background information can help bring new readers up to speed, while allowing more informed visitors to read ahead without distraction. (This is how Wikipedia’s built a ton of inbound links over the years, and why news organizations ought to consider more frequent use of standing reference articles on their websites.)

Contextual hyperlinks can links to the definition of an unfamiliar term (see, for some readers, the “easter egg” link above). They also can help explain gags that the author attempted but that some readers might not immediately get (see, for many *more* readers, the “set it and forget it” joke above).

3. Easter eggs

These, like the easter eggs hidden in DVDs and video games, are there just for the amusement of writer and audience alike. They defy too much explanation and analysis, as their purpose is simply to provide a little humor. Just, please, for the love of all that is holy, do not let it be a yet another Rick Roll.

Ultimately, the addition of useful hyperlinking within an online news story reflects the strong reporting of its author. If a reporter does not know of online pages with extra information relating to the story, he or she cannot link to them. But if you have that information, why not share it with those readers who are eager for it? In a hyper-competitive online news market, writers and their publishers need every advantage they can offer against other websites.

The test: When to link?

Though I promised not to tread on Nielsen’s turf, for clarity’s sake, two or more links should not bump against one another, leading to readers to believe that they are seeing just one link. Nor should linked text make up more than a small fraction of the text on the page. For that reason, online writer do better to link a key word of clause within a sentence, and rarely an entire sentence, when inserting a link.

To conclude, here is my four-question test for online writers to keep in mind as they consider how, and where, to link within their stories:

1. Does the URL to which I am referring the reader reward him or her with additional content that a reader of this story likely did not know, or know how to get easily?
2. Does the text I am selecting to link this text give the reader an obvious clue as to what the hyperlinked page will contain?
3. Am I using the shortest possible amount of text to provide that clue?
4. Would the content of the linked text, or the context surrounding it, reasonably mislead the reader into believing that the linked page contains something other than what it does?

If the answers to these questions are yes, yes, yes and no, you’re good to go with the link.

In defense of Facebook

One my student editors here at OJR forwarded to me at New York Times piece reporting the latest complaints about Facebook’s policy toward its users who wish to cancel their memberships and delete their profiles.

Facebook does not provide a “one-click” solution for leaving the site. Members may delete content they’ve submitted to the site, one item at a time. For active users, tearing down all that content could take dozens of hours. And even then, Facebook retains much of your basic contact information, making it possible for other members to contact you through the site.

The frustration, even anger, that many such users feel toward Facebook is palpable. The Times quoted several readers who had attempted to delete their information from Facebook, with varying, but never total, success.

Facebook began as a social network for college students, many of who believed that what happens on Facebook, stays on Facebook. But as the social network has opened membership to those without .edu e-mail addresses, it’s become a much broader community, with many professional organizations maintaining groups and contact lists through the site.

Put it this way: Imagine you are a journalism student and you join OJR’s Facebook group in order to connect with editors whom you might be sending your resume and URLs in a year or so. Do you really want those editors on OJR’s Facebook group to be one click away from pictures of you, drunk, at some party last semester?

(Not that I am trying discourage journalism students from joining our Facebook group, of course…. Heck, there are likely a good many photos from various industry conferences that I am sure that those editors would not want one click away from a journalism student’s eyes, either.)

As a website publishers and editor, I’d like to offer a few paragraphs in Facebook’s defense, however. Granted, my defense is not absolute. Facebook has done many things over the years worthy of its member’s criticisms. (Beacon, anyone?) But publishers do have reasons to limit their readers ability to control information published on a website.

Editorial integrity

Not every website that accepts and publishes user-generated content (UGC) is a pure social network with no interest beyond providing registered members a place to chat amongst themselves. Many websites rely upon UGC to power wikis, discussion forums and other collaborative publishing forms that are read by a much larger number of individuals than ever post to the site.

Empowering registrants on such sites to delete all their submitted content with a single click threatens the integrity of those collaborations. How difficult would it be for readers to follow a discussion thread where every fifth response, say, or even the parent post, was deleted? No, this is not a problem on websites where discussion threads have no archival value. But that’s not always the case. Publishers who are attracting fresh traffic, and advertising revenue, based on informative discussion threads have powerful incentives not to allow readers to destroy that content.

I’ve lost count of the number of discussions I’ve had with colleagues in this industry about protecting the quality of interactive content by preventing access by those who would harm it with their contributions. But deletions can cause grave damage to online content as well. (See two past OJR articles on this topic, here and here.)

Deterrent against abuse

Here’s a scenario: You require readers to register in order to contact other registrants through the site. Someone registers, spams selected readers with who-know-what abuse, then immediately deletes his or her membership. It’s the online equivalent of a drive-by shooting.

If the publisher does not retain some of the offender’s information after his or she deletes the account, there might be little hope of ever catching them. Readers can figure that out, and sites that allow verbal drive-bys become far more attractive targets for this sort of behavior.

That’s why a policy of retaining user contact information, under certain circumstances, can help encourage more civil behavior on the site. Such a policy can also help a publisher resolve claims of impersonation and identity theft, since the publisher would have a record of who was behind an account that posted certain information, and when.

Publication is, well, public

All that said, Facebook’s argument in favor its retaining member’s information, that it makes reinstating an account easier, makes a much weaker argument. Members should learn that actions on a website have consequences. Quit, and you lose your profile, your lists, your blog. If you decide to rejoin later, you’ll have to do the grunt work of recreating all that you’d built before.

Yet keeping some of that content on the site, and public, can promote that same lesson, as well. Publication is just that… public. Therefore, people ought to be encouraged to think before they post. Maybe they take the time to adjust their account’s privacy settings, as Facebook allows one to do, to limit the information that people outside their approved social circle may see. Or, maybe, they decide that certain personal information ought not be on the Internet at all.

Of course, this is a lesson that ought to be taught before a person stats posting on a website, and not after that individual decides, “Um… maybe that online rant against my advisor wasn’t such a good idea, after all.”

Online publishers need to do a better job of promoting media literacy in the Web 2.0 world. As Newspapers in Education programs introduced kids to the content available in their local papers, perhaps we need a new program that introduces beginning Internet users to online publishing, to what happens to information that they post online, and what they can and cannot do to control that.

A publisher’s decision

A publisher could decide that he or she will allow readers to have complete control over the information that they publish to a site. Some websites explicitly cede ownership of and copyright over UGC to the users who created it. In those cases, publishers, to be consistent, should stay away from the Facebook model, and instead enable easy, user-controlled deletion of their content.

Whatever approach publishers choose, they can best protect themselves from Facebook-style criticism by taking every opportunity to communicate their policy to their readers, in plain language.

Some readers want an anonymous community that easy to join, and easy to drop. Others desire an online community with thoughtful comment from identified correspondents. There are as many options available online as there are publishers. Let’s just not lead readers to believe that their community lies in a different type of neighborhood than its publisher envisions.

Got something to say? Then say it!

The skill set for managing an online community lies somewhere between carnival barker and drill sergeant. You’ve get the crowd’s attention, draw ’em in… then train them and keep them in line once they’ve enlisted.

The job becomes even tougher for journalists, who want to draw traffic and elicit discussion while maintaining journalism fundamentals. It’s easier to open the doors for an anonymous shouting match than it is to craft a well-sourced and enlightening conversation.

Although we at journalism schools teach our students to write in an engaging and conversational manner, journalism is not casual conversation. The work we do to report and source our information tends to lend our words a formality beyond that offered by someone pulling their words from “thin air.” Ideally, we minimize that sense of formality in an effort to earn credibility for our work without intimidating the reader.

In addition, a reporter’s job, ideally, is to answer questions. If you’ve worked in a newsroom, think back to your first editor, or your basic reporting professor. When he or she told you to check out a lead, what would have been the reaction if you’d responded, “Uh, I don’t know”?

1) “Oh, gee, that’s okay.”
2) “Well, find the heck out!”

Journalists are trained from their first day on the job to find answers. That makes it hard for reporters to turn to their readers publicly and declare, “I don’t know. Help me out here.”

All these factors stand in the way of journalists running vibrant online discussion communities, even as our reporting skills and community know-how make us ideal candidates for those gigs.

We’ve offered dozens of articles on OJR over the years with advice on managing online discussion communities. And, as editor, I’ve tried to ensure that we’ve practiced much of what we’ve preached. Which is why I’m here to explain today a change we are making in the way that we are handling comments on the website.

Since I rewrote OJR’s content management system in the fall of 2004, OJR has required that readers register with the website in order to post a comment on the site. Our registration process is two-step, and requires registrants to retrieve a password from their e-mail accounts in order to log into the site.

In my experience, this system offers the best protection against spam bots and flame war trolls. The registration requirement keeps automated agents from exploiting input forms and the e-mail requirement deters anonymous hacks who want to cause trouble without consequence.

It’s not a perfect system; some spammers employ sweatshop labor to manually labor and submit comments to highly-linked websites. And even those who proudly attach their name to their comments can be jerks sometime. (Do I get some Fifth Amendment opportunities here?)

But, on the whole, I’ve found that this system, employed on other websites, helps keep the signal-to-noise ratio quite high, with a minimum of effort from site editors and moderators.

Yet a high signal-to-noise ratio doesn’t help readers that much when that signal remains weak. And in the relatively small world of the news publishing industry, sometimes people do not want their names attached to comments about their company’s vision and practices (or lack thereof.)

So, today we’re implementing a change at OJR: Readers may now submit comments to the site without registering.

That doesn’t mean we’re opening the gates to anything. Comments submitted by unregistered readers will be held for review before being posted to the site. And those comments will be identified by the poster’s IP address, rather than a log-in or reader’s name. (Unregistered readers will be able to include their name within their posts to OJR if they choose, of course.)

I hope that this alternative provides a way to readers to get introduced to commenting on OJR without having to go through the extra steps of creating an account and retrieving an account password. And that it provides a way for newsroom employees to add to the conversation in situations where they fear reprisals if their names were attached to their comments.

Of course, as journalists those of us reading the site will have to decide how much credibility to give to posts that come from unregistered readers versus those submitted by readers who have registered and supplied OJR with a working e-mail address. (You will know the difference because posts from unregistered readers will include an unlinked IP address, rather than a linked author’s name. Hey, at least we’re not slapping on the label “anonymous coward“.)

Ideally, from my perspective as editor, folks will try commenting using the anonymous system, decide that they like it, then register and becoming frequently contributing registrants on the site.

Spend more than a few days on the Internet, and you’ll see the whole range of conditions that sites impose on posting: from wide-open input forms without captchas to locked-down systems that require credit-card-verified user accounts.

Ultimately, we want more conversation, and less lecturing, on the site, and I hope that this change will move us toward that goal. And, as with everything on OJR, we reserve the right to change our minds — to make commenting either more or less restrictive than we will have it now.

Wanna share your experiences/frustration/success in running an online discussion. Hit the button and talk to us. Even if you haven’t registered yet.