Best use of nofollow by a commercial site
Sure, Answers.com blows away every other reference site, with a clean display of dictionary, thesaurus, specialized dictionary, Wikipedia, and translation results, with cruft-free URLs for easy linking (anything is at http://www.answers.com/anything, nothing to remember) as well as copy-and-paste HTML for fancier links one click away from anything, an RSS feed of “Today’s Highlights” to keep you interested in them, along with a program you can install to access them from any other program, that I don’t use, and a Firefox search plugin I use constantly. But that’s not what I want to talk about…
Maybe it’s just because I don’t spend much time on the commercial web, but I’m surprised by how little use of rel=”nofollow” I’m seeing (in blinking lime) on non-blog, non-wiki sites. It seemed like such an obvious tool to use, to micro-manage how you transfer PageRank within your site and to other people. However, Answers.com is doing a great job of showing how to use it to their advantage. Every page of answers that includes dictionary and thesaurus results is apparently required to link to Houghton Mifflin Company, and link it does, with nofollow. Answers from Wikipedia link to Wikipedia, and also include a pro forma link to the GNU Free Documentation License, nofollowed. Because they are generous, goodhearted folks with the best interests of their users always foremost in the every action, they link to search results for your query, including image and news searches, Technorati, and Amazon searches. Because they aren’t foolish enough to push the competition’s results page higher than their own if they don’t have to, those links are all nofollowed. Every single page links to their privacy policy and terms of use, but because having a high-ranking page for “terms of use” just means people too cheap to pay for a lawyer will steal yours, they’re nofollowed. All in all, an outstanding example of how to use rel="nofollow" for the one thing it’s really good at, moving PageRank only where you want it moved.
Well, I guess I should have known this kind of abuse would happen! I just never expected anyone to use it for links to pages on their own site!
Whyever not? For some sites, like corporate brochureware, having policy pages is handy. You want to optimize for the main page, and a TOS or privacy page gives you another way to have every page link to something that links back to the front page. But when you want to optimize for every single page on the site, they become PageRank Motels, where it comes in from every page, but can’t usefully go back out.
Nofollow never stood a chance of doing anything to stop or slow comment spam. Unless you boil the ocean, and force it on every single site with comments (which also means forcing every abandoned blog off the web), it can only increase comment spam. Spammers don’t read blogs; they just write to them. If half the comments around use nofollow, then a spammer needs to post twice as much, to stay even. If 90% use nofollow, he needs to post ten times as much.
If you are a search engine, and you feel that not only will it be valuable to your results to not count comment spam links, but also to not count most actual links in comments and comment author links and TrackBack links as well, then it has some value, but to the rest of us, the only possible value in it is in micro-managing how search engines see our links.
Unfortunately, when you don’t know whether it’s a case of 1/m or 1/(m-n) it’s harder to really (ab)use it.
A big part of search engine optimization is to decide how pagerank should flow within your own site.
I haven’t used Answers.Com for reference; I usually check Information Please first.
Bleah. I’m willing to believe that InfoPlease has better information, somewhere on the site. I’ll never see it.
The first thing I see is a popup-blocked infobar, followed by an enormous banner (sometimes delayed, depending on whether the Flash ad is the top banner, or the sidescraper), followed by a mysterious stew of navigation links and advertising links. Query links aren’t memorable (I can type answers.com/germanium without having to think), so I have to load one set of ads, and then search, and when I’ve searched for a simple term, I still don’t get answers(.com ;)), I just get search results that I have to dig through. The ”cite this page” tool doesn’t give me a link, it gives me someone’s idea of how to cite in a dead-tree paper. But the real killer is returning search results, not answers. Answers.com does provide links to other related articles that may or may not be what you were after, but only down at the bottom of a page of actual content that, for the sort of things you should be searching a reference site for, was probably what you wanted.
That’s funny. I didn’t realize how uglied-up InfoPlease was until you pointed it out. I’ve been using it for so long I focused on the postage stamp of real content they place among the torrent of advertising.
However, the Answers.com RSS feed just sucks. It’s a single item (amusingly placed before the channel-level stuff, which I’ve never seen in all the thousands of feeds I’ve looked at), so you have to check it at least once every 24 hours to not miss an item. The single item is apparently rotated at 00:00 UTC, reasonably enough, but the HTML item isn’t available until some time later, so when the new item first appears in your aggregator, the link redirects to the main page, showing the content from the RSS item you saw the day before. The feed description is just a plain-text summary, so if out of the Topic, Word, Day in History, and Birthdays you are only interested in the Word of the Day, you’ll have to click through to the daily HTML page, and then to what you wanted. Bleah. Unsubscribed.