Zemanta's Related Articles: Very Mixed Feelings [en]

For years now, I’ve been seeing my articles pop up on other blogs under Zemanta‘s related links. And for years, it has bothered me.

Actually, it’s a bit more subtle than that: it does not bother me at all that people list my articles under related links on their posts. It’s flattering. It’s linkage. I like it.

What bothers me is two things:

1. the trackback “spam” that often accompanies such links
2. the fact search engines do not differentiate between “real” in-content links and related links.

Somewhere in the middle of the first sentence of this article, I decided that if I was going to complain about Zemanta (though it’s not only “complain”, you’ll see — I said mixed feelings), I should at least install it and try it out. Being the old-school blogger I am, I hadn’t gone down that road yet, believe it or not.

To be honest, Zemanta is a blogger’s wet dream. No more hunting for that Facebook logo or an image to illustrate your post. No wonder most Big Blogs nowadays always have a picture handy! Zemanta crawls all over the place, comes back with a bag full of images, and lets you select the one you want. Ditto for links to services or brand names you mention. And ditto for other articles with related content that you might want to send your readers to check out.

Using Zemanta is not automatic, as they told me a few months ago when I complained to them about what I think of as the trackback spam issue. And indeed, as a blogger, I have to click on the Zemanta elements I want to include in my post.

So where is the problem?

The problem, for me, is that it does not mean the same thing if somebody actually takes the trouble to read my article, writes it into his post in context, or simply dumps it with a bunch of links at the end of his post because it had a nice title or sounded vaguely related to what he was writing about.

It cheapens the value of the link.

That semi-automatic link involves little or no effort, little or no research, no real endorsement. I’m ready to bet that most of the Zemanta related links bloggers put at the end of their posts are there because they look good, rather than because the blogger once read the article, remembered it, hunted it down again (or bookmarked it), and decided to link to it while writing.

With that in mind, sending trackbacks to these related articles is exactly the practice I frown upon in my recent post about Technorati, Tags, and Trackbacks. And in all honesty, I wouldn’t mind if they were systematically nofollow, or at least if blog search engines like IceRocket or Google Blog search learned to make the difference between in-content links and end-of-post semi-automatically-generated link dumps. (See my IceRocket search for CTTS and the Google Blog search one — and check out how many of those links come from Zemanta rather than human beings.)

Why am I so brutal about these related links?

I have no problem with the idea of listing related links to a blog post. I have no problem with automatic lists of related posts — I even use them here on Climb to the Stars. But c’mon, if we’re putting nofollow on comment author site links, we should also be putting nofollow on related links. Specially as I see, now I have installed Zemanta, how easy and noncommittal it is to include these related links.

Look at this screenshot: I see a list of blog post titles related to what I’m writing. I can hover on one and see a text snippet, click on “read more” to quickly check it out (am I going to do anything more than “quickly check it out” if I’m writing a blog post and impatient to hit Publish?), and then simply click on the post title to add it to the end of my content under “related links”. Easy. Too easy! These links are not content-driven — unless you consider their presence in the Zemanta “related articles” is content-driven by Zemanta’s algorithm. But their choice is not driven by the fact the blogger values their content.

One thing I was told by Zemanta (IIRC) was that bloggers could choose to add nofollow to their related links. Actually, they can choose to add nofollow to all their Zemanta links. All-or-nothing. And honestly, the way it’s phrased, who would want to select that option?

(No way I’m going to tell Zemanta to mark all the links it creates for me as “objectionable”. No way.)

So, what are my thoughts now?

• I like the idea of Zemanta as a content-enhancement support tool, I don’t want to trash it
• it seems specially useful for images as far as I’m concerned (though I’m disappointed it didn’t pick up the two screenshots above from my stream when I uploaded them — had to add them manually)
• I like my blog showing up in related links elsewhere, though I don’t give that much value to it, and I really don’t see it as a valuable source of traffic (my stats tell me that)
• search engines and blog tools should make a distinction between “manual” links and automatic/semi-automatic links, particularly of the “related” kind
• I don’t want to get trackbacks when somebody includes my blog in their related links: maybe Zemanta could provide a way for blog owners to record that preference? would there be a way for Zemanta to tell blog tools like WordPress “don’t send trackbacks or pingbacks for this or that link?”
• the nofollow setting in Zemanta needs to be a little more subtle than all-or-nothing, and do away with the scary wording (“objectionable”, c’mon on)
• and while we’re at it, is there a WordPress plugin which would allow me to “un-nofollow” links left on my blog by certain commenters? the honest-to-good human beings who do not spend their time trying to link-drop?

Note: in this post, I used Zemanta to link to… itself (in the first paragraph), add the logo top right, and that’s it. I’m going to keep it active for the few next posts though to see if I actually use it, other than just liking the idea.