Back to previous post: Don’t Be Confused

Go to Making Light's front page.

Forward to next post: Go Bags

Subscribe (via RSS) to this post's comment thread. (What does this mean? Here's a quick introduction.)

October 28, 2007

Exasperated with Technorati (again)
Posted by Teresa at 11:56 AM *

Over at The Whatever, John Scalzi says:

…despite record visitorship, Whatever’s Technorati ranking’s been tanking, relatively speaking: It’s at about 2,500 at the moment, which is down from a high of about 800 eighteen months ago or so (that’s right, before BaconCat). I believe Technorati’s changed its ranking formula since then, and anyway, the Whatever being more or less out of commission in August and September will work to reduce links in. I suspect switching blog software might fiddle with things as well.
I don’t think it’s Scalzi’s blog software. A month and a half ago, on September 12th, Making Light finally broke into Technorati’s “Top 1K.” It wasn’t a fluke; we got to that rank via small steady increments over a long stretch of time. A few weeks later, I checked our rank and we’d dropped to 1,750. During those same few weeks, we’d had unusually steep growth in our readership.

That’s just flaky. How is it that in late September and early October of this year, there was a swerve in weblog traffic and linking patterns sufficient to produce those results on Technorati, but the change wasn’t otherwise observed in the blogosphere?

Moreover, what kind of change in weblog reading and linking patterns would drop The Whatever from a position 200 places ahead of to one 750 places behind Making Light? The Whatever is a thoroughly established and much-linked-to weblog. More to the point, it gets a lot more traffic than we do.

However valid or invalid the previous Technorati rankings were, the new versions make no sense. Given that the same people were responsible for both sets, I have to ask: if the old stats they produced were so bad that they had to be replaced with this new system, why should I believe that the people who were responsible for maintaining the old stats know any more about what they’re doing now?

Comments on Exasperated with Technorati (again):
#1 ::: Leva Cygnet ::: (view all by) ::: October 28, 2007, 05:42 PM:

It may not be Technorati, exactly, but rather that people are getting better and better at generating links by illicit means. And there is now enough money in blogging to make it worthwhile to game it and to try to increase the rank of a blog with artificial links.

So if you're a scammy blogger who's just in it for the money, you go to blogger and blogspot and umpty zillion other blogs that allow free blogs and start creating free dummy blogs that just link to your main blog that you want to promote. Since technorati doesn't have any real filtering that I've seen, it's going to count each blog as a link.

I would be completely unsurprised to find that there are people who've made a business of creating those links, too. Blogger's free, right? And Blogspot? And bunches of others? So you create a script that logs in to ten thousand clone blogs, generates a few lines of semi-random text, says, "I like!" and generates a link. Then when you've got a blogger who wants to promote a blog, the blogger pays $1000 or $2000 or $50 -- whatever the market will bear -- to the guy with all the clone blogs. He pushes a button, his script starts logging into the clone blogs and generating links, and voila, instant technorati top 10K rank. At least, this is my theory ...

Mind, this may not be intended to game Technorati directly, but may be a side effect of people who are trying to increase their google rank. Google's cracked down on "paid" links pretty hard, where people simply sell a link for between about $5-40 a month (depending on the rank of the site) so they need to find other methods of generating those all-important links.

Firefox News, for what it's worth, has dropped 2,400 technorati places practically overnight, from 7200 to 9600.

(And don't even get me started on the sites that simply fill their pages with rss feeds, slap a few ads on it, and call it a web site, without generating any original content of their own.)

-- Leva

#2 ::: Casey ::: (view all by) ::: October 28, 2007, 05:45 PM:

It looks like their system took a nose dive a couple months ago. There's another post on their blog that talks about relocating various bits of their network.

On the other hand, Leva@1 has a good point

#3 ::: Martin Wisse ::: (view all by) ::: October 28, 2007, 06:35 PM:

Does Technorati actually take traffic to the blog into account in its rankings, or is it purely linking that they count? For I thought it was the latter.

More links to your blog almost always means more traffic to it as well, but just because Scalzi has a huge and loyal readership does not necessarily mean more people link to his posts than they do to ML posts.

Also, the growth of the web also comes into play: having 1,000 links coming in from a total of 10,000 blogs means something else than having 1,000 links coming in from 10 million blogs...

#4 ::: David Harmon ::: (view all by) ::: October 28, 2007, 06:45 PM:

Hey, IMHO it's all just handwaving... does anyone actually know enough their algorithms to discuss their stability?

#5 ::: Bruce Cohen (SpeakerToManagers) ::: (view all by) ::: October 28, 2007, 07:37 PM:

Technorati's list is high,
many seek to scale it.
But do we know what gets some by,
and what makes others fail it?

Scalzi's rank now falls below
Making Light's own rating.
The problem is we cannot know
Just how they do their weighting.

#6 ::: Leslie in CA ::: (view all by) ::: October 28, 2007, 07:40 PM:

Casey @ 2, your links don't work.

#7 ::: Chad Orzel ::: (view all by) ::: October 28, 2007, 08:53 PM:

Moreover, what kind of change in weblog reading and linking patterns would drop The Whatever from a position 200 places ahead of to one 750 places behind Making Light? The Whatever is a thoroughly established and much-linked-to weblog. More to the point, it gets a lot more traffic than we do.

To be fair to Technorati, John's blog was basically off-line for a couple of weeks there. That's long enough without new content for a lot of links to move off the front pages of blogs, and that's what Technorati really counts (as far as I can tell).

I've also seen a drop from about 9,000 to about 12,000, without a corresponding drop in the "Authority" (which seems to be simply related to the number of blogs linking to a given site in some period of time), so there has been a screwy change to the system. But it's not inconceivable that the Whatever might've fallen farther, faster than Making Light, given the lengthy site outage.

#8 ::: Teresa Nielsen Hayden ::: (view all by) ::: October 28, 2007, 10:10 PM:

Leva (1), I know the fake blogs you're talking about, and they've been around for a long time. I wrote about them back in 2005, on a previous occasion of my losing patience with Technorati. For Google, Technorati, and other pageranking sites, they've got to be part of the background noise of the internet -- a problem to deal with, no doubt, but not a reason to make drastic changes in the system.

Casey (2), can you try those links again, or just post the URLs? I'd like to see those pages.

#9 ::: Casey ::: (view all by) ::: October 28, 2007, 10:39 PM:

Sorry about those links. Don't know what happened.

Two entries from Technorati's official blog.

#10 ::: Kevin Marks ::: (view all by) ::: October 29, 2007, 04:16 AM:

As Technorati's first employee and the one who wrote the spiders, I should respond here. Technorati's ranking was changed to integrate over 180 days a while back. Thus Scalzi's blogging hiatus is rippling through now, but will be repaired soon, as long as he continues to attract linkers.

#11 ::: Charlie Stross ::: (view all by) ::: October 29, 2007, 04:35 AM:

I've noted a drop in my blog from around 3500 to 4500 some time in the past few weeks. (I don't exactly check Technorati daily.)

#12 ::: midori ::: (view all by) ::: October 29, 2007, 08:58 AM:

The lagging 180 days thing is interesting.

Charlie, if I was going to guess, I'd say that the bursty nature of blog posting is influencing your (and Scalzi's) results. You had several quiescent periods, prior to, and during Worldcon. (And I skipped reading your excerpts while you posted them - I enjoy your pithy commentary, but I couldn't afford to buy another book just yet ;)

Anyway, if burstiness leads directly to ranking wobble, then I would expect multiple contributor blogs that post often, like Making Light (or Boingoing or Techdirt) to rank pretty consistently* - compared to other, non-bursty blogs. Do you suppose that's so? I wonder if there's a way to express how deeply ML posts go - I've never seen anything like the depth of thread here outside of Slashdot or DailyKos, and it seems like people actually read the whole thing.**

*consistently compared to one another. I.e. the difference between ranks could vary quite a bit, but the order probably wouldn't: Boingboing, Techdirt, Making Light
**TNH, is this why you prefer non-threaded discussions?

The search box on technorati is stupid has poor UI.

It would seem obvious that typing in the name of a blog, or it's url would display the summary for that blog. "techdirt" "" and "" should all produce similar results. The first returns nothing, the second only mentions of the blog, and the third nets you blog reactions and a capsule summary on the right. WHY? Are they trying to make your service hard to use for naive users?*

*bonus round:
If you are sent to an individual post on their own blog, there is no way to get to the main blog without selecting the url in the address bar and removing everything after /weblog/ . "Home" and the banner at the top take you to the front page. I guess those guys don't want people to read their blog?

#13 ::: Charlie Stross ::: (view all by) ::: October 29, 2007, 09:38 AM:

Midori: Making Light is itself somewhat bursty by nature -- no criticism of our hosts intended, but I've noticed that sometimes it falls quiet for up to a week when everyone's away on some business or other. Really, there's no way to maintain a non-bursty blog, unless you're (a) prone to logorrhoea and save up postings for a rainy day, or (b) able to pay your staff to post on a daily basis. (This latter type of blog we have a technical term for: "magazine".)

Nasty little thought: what is the chance that there's a political angle to this, such as a metric shitload of election-focussed political blogs coming on stream and cross-linking to each other in a coordinated manner? With genuine content being fed in via campaign managers, and a connection into the right- or left- blogospheres, that'd tend to have a knock on effect on everyone who isn't in that particular loop.

#14 ::: Marie Brennan ::: (view all by) ::: October 29, 2007, 10:18 AM:

Related question, that I've been wondering about for a while: does anyone have an idea why my Technorati watchlist has become utterly useless? When I go to that page, it invariably, for months now, says there are no items in my searches -- even when I know new things with those terms have been posted. For a while I was just retyping in whichever ones I felt interested in, and getting my results that way, but now even that fails. The most recent posts it pulls up are a month or two old, when I myself could point to half a dozen within the last week.

I've essentially given up on using it, but I do miss the convenience, so if there's a way to fix that, please clue me in.

#15 ::: Fred ::: (view all by) ::: October 29, 2007, 12:02 PM:

I've also noticed problems with my searches, and the rss feeds attached to them. Sometimes I don't get posts; more often than not, I get the same posts repeated over and over again. I was getting one result a couple dozen times a day in my news aggregator for a couple of weeks.

Seriously, though, if Technorati's page ranking is incorrect and its web/blog search is this flawed, is it of any use?

#16 ::: Kevin Marks ::: (view all by) ::: October 29, 2007, 12:32 PM:

You can look at historic charts of links to Making Light and Whatever at Technorati, and you can see that you both have a big spike just moving out of the 180 day window (Scalzi's was bigger, and in March).

#17 ::: Bill Higgins-- Beam Jockey ::: (view all by) ::: October 29, 2007, 02:52 PM:

How exactly does one obtain the Technorati rank of a blog?

#18 ::: Teresa Nielsen Hayden ::: (view all by) ::: October 29, 2007, 04:15 PM:

Midori (12), that is indeed why I prefer non-threaded comment sections. As a secondary reason, I dislike threading's assumption that every comment is the descendant of another single comment. It artificially segments and compartmentalizes the conversation.

#19 ::: midori ::: (view all by) ::: October 29, 2007, 04:35 PM:

17, Bill Higgins-- Beam Jockey:
How exactly does one obtain the Technorati rank of a blog?
Funny you should ask. I don't know the best method, but what I do is type the blog url (e.g. but not the full url (see above) into the search box on the Technorati site. This gets me a summary blob on the right about the blog, which has the number.

I found a better way is to do it just now while trying to answer your question. Take this:
and replace the example with the url of your choice.

I have no idea how you scrape the whole list. I understand there is some kind of firefox plugin you can use to compulsively check your status. Does technorati even work with livejournal? I wonder if it is even possible to monitor the relative positions of a subset of blogs?

10, Kevin Marks,
I realize I was (indirectly) busting your chops about the UI, but do you and your crew realize how obtuse and non-obvious your service is? Both in terms of what it is, and how to use it? I'm not trying to be insulting, I'm just wondering what is going on inside your heads, and I don't really have many other tools than to ask.

18, Teresa Nielsen Hayden,
(yay! I got something right!)
You also wrote:
As a secondary reason, I dislike threading's assumption that every comment is the descendant of another single comment. It artificially segments and compartmentalizes the conversation.
Ah, that makes sense. Sort of the folksonomies are more efficent than arbitrary categories thing: people are bad at deciding where a topic splits. (And do a poor job of re-integrating insights if a conversation has been flagged as being 'about something else.')

#20 ::: Avedon ::: (view all by) ::: October 29, 2007, 09:53 PM:

I've noticed the same thing.

On the new ranking system I've been in the top 10k, somewhere between 7k and 7.7k, ever since they started showing it on the top right corner with the little thumbnail.

Then I suddenly started sinking and I'm outside the top 10k now.

This despite the fact that I've been posting more at Eschaton lately and link to myself in my signature (in honor of Brendan Nyhan blaming Atrios for something I once posted). This also, of course, generates more links from blogs that have not previously linked to me.

I'd think links from a high-authority blog like Atrios' joint would have pushed me up rather than down.

#21 ::: Jan Vaněk jr. ::: (view all by) ::: October 30, 2007, 06:45 AM:

#9 / #2 on why some links are broken: Apparently ML's installation of MT swallows element A's HREF (and others' other?) attribute if it is not enclosed in quotation marks - cf my blog that could use some PageRank pimping [oh dammit, preview shows rel="nofollow"! :-] (in Czech only, alas - on LJ, my favourite nick is squatted by somebody who seems to have gone offline years ago). Yes, I consider it rather a bug than a (good) feature and would welcome any change, if not an outright solution then at least a warning in the "HTML tags" section.

Re topic drift to handling comment threads: I'm quite sure I would read ML more if there were less than hundreds of comments to go through under virtually every post. (This one was apparently too uninterestingly specialist - and I may be coming too late to get noticed :-) / :-( - how does one type an ambivalent smiley?) Also, while I appreciate TNH's sentiment/argument @18, it is annoying having to pick out only those lines of discussion that I'm interested in between several overlapping offtopic/topic-drifted ones, with the average lag about 10-20 comments; some threading would be nice.
But I admit that introducing comment #s made it rather easier (that is, humanly possible at all); as I was saying recently, getting some way of converting those # references to direct links would solve most of the remaining problems.

#22 ::: Avedon ::: (view all by) ::: October 30, 2007, 07:01 AM:

on LJ, my favourite nick is squatted by somebody who seems to have gone offline years ago

When I opened my LJ account, I discovered that "Avedon" was taken by someone who had set up an account for an RSS feed to my blog. Only the facility that was creating my RSS feed died, and I can't contact the owner of the account to get him to pick up the feed I now generate myself, so it's just moribund.

getting some way of converting those # references to direct links would solve most of the remaining problems.

The date-stamps are the permalinks for the individual comments.

#23 ::: Jan Vaněk jr. ::: (view all by) ::: October 30, 2007, 08:55 AM:

The date-stamps are the permalinks for the individual comments.

I know this (Bob ;-) But even I am not that anal to manually write out HTML links when referring to another comment, and it certainly can't be expected from people who lack the 133t c0ding ski11z. What is needed - that is, what would be, or what I'd like to see, or what I'm talking about - is automagically taking strings "#<number>" in people's comments' raw text and turn them into permalinks to said comments' anchor. Which is somewhat complicated by the fact that the system (in its default setting only?) uses global unique IDs for comments and not "human" relative numbering starting anew with each post (my #21 above has 222609 in the permalink, Avedon's #22 is 222611 - inbetween there was 009528.html#222610), and would most likely and ideally require a change in its source code (as I was saying, I still think/hope there must/should be a MT plugin readily available for task so obvious as this) - once I saw this achieved on another blogging system (without the global numbering complication) purely by an external JavaScript, so I guess this would be principally possible here as well, but I certainly lack the coding skills necessary, as well as access to ML administration.

(BTW, "Don't me me type all this again" seems not to be working for me. Good think we have Autofill...)

#24 ::: Nancy C. Mittens ::: (view all by) ::: October 30, 2007, 10:53 AM:

Jan, I've always seen the pipe used for ambivalent:


#25 ::: Bill Higgins-- Beam Jockey ::: (view all by) ::: October 30, 2007, 02:06 PM:

Nancy C. Mittens writes in #24:

Jan, I've always seen the pipe used for ambivalent:


Greg Ruffa, who was in the satellite-launching business, used to say that this stood for

#26 ::: Teresa Nielsen Hayden ::: (view all by) ::: October 30, 2007, 03:40 PM:

Jan (21), just for you, I took the "rel=nofollow" out of your comment URLs. I also added quotation marks to the second one, which you'd automatically typed without them.

Midori (19): right again.

#27 ::: Michael Roberts ::: (view all by) ::: October 30, 2007, 05:57 PM:

Oh, so that's how Technorati works. I'm ecstatic to say that my blog is ranked 2.8E6, give or take a few insignificant digits. And the only links to it are from ... me.

Ah well. Posterity will value my writing, if not the cruel present. :-)

#28 ::: Mez ::: (view all by) ::: October 31, 2007, 03:31 AM:

Hmm. Looked up (as I extremely rarely do) some blogs & websites of mine or those known to me. Some are just not on the map 'No one has linked to this page yet.'; one at rank 8,911,336; *three* at rank 4,446,976. A very precise sort of rank for several things, especially ones that have a different number of links to them. I wonder how many other sites are at 4,446,976 or 8,911,336?

#29 ::: Jan Vaněk jr. ::: (view all by) ::: October 31, 2007, 09:40 AM:

#26: Thanks, but I was just joking - and I left out the quotation marks in the second link on purpose, to show what happens in such a case, which was the main point of my comment. (Anyway, do repeated links from the same page add more PR than just one?) Dammit, I should have gone with the original idea of a href="" vs. a href=; I must be having a bad week for expressing myself intelligibly... :-(( (At least the system now seems to remember my data.)

#30 ::: Casey ::: (view all by) ::: October 31, 2007, 11:40 AM:

#21 thanks for pointing out the error of my ways :)

#22 here's the LJ FAQ on changing a feed for a syndication account.

Smaller type (our default)
Larger type
Even larger type, with serifs

Dire legal notice
Making Light copyright 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012, 2013, 2014, 2015, 2016, 2017 by Patrick & Teresa Nielsen Hayden. All rights reserved.