Blog du Tristank

So terrific that 3 of 4 readers rated it "soporific"

Why Changing Your Blog URL Sucks

Why Changing Your Blog URL Sucks

  • Comments 9
  • Likes

Jay posted about his languishing site traffic since a URL migration for his blog, and I thought I'd get in on the action as well.

Now, I don't work in Search, and I'm essentially speaking from the logical inference portion of my bum here, so please feel free to correct any erroneous assertions I've made.

The Way Of The Blog

One of the first things I read about corporate blogging (and I recommend anyone that's thinking about it goes to read it) was the Corporate Weblog Manifesto, by Robert Scoble.

Lurking near the middle section is The Terrible Twelve:

12) Never change the URL of your weblog. I've done it once and I lost much of my readership and it took several months to build up the same reader patterns and trust.

This is a really, really, really important point. I can't understate how important this is from a traffic perspective, especially with the added measures employed these days in post-honeymoon-period blog security.

Speaking for myself, I have a "core" group of people that read this blog regularly, and are probably subscribed to it directly. That number took an instant nosedive when the URL changed, and the traffic is still well below where used to be (which also tells me I'm not nearly as interesting as I was last year. I blame phenylalanine).

A reasonably large portion of traffic used to come from search engines, particularly Google. Since the migration, search traffic has all but dried up.

Why?

Well, this is where my crazy random uninformed theories come into play, but I think it's something like this:

I publish a story called "Tristan's Story". The URL is blogs.example.com/tristank/story .

Someone else thinks the story is cool, and links to blogs.example.com/tristank/story .

Other people (let's say bloggers) link to the same URL, or to the URL of a post that references the original URL, forming a tree of links (nb: I prefer "Pyramid Schemes" to "Tree Schemes", so I'm going to call it a pyramid).

Quick note on assumed search mechanics here: My guess is that most crawly search engines give each link in the pyramid a weighting based on the number of children (and their weight, based on their children and so on), and the overall weight counts towards the top link in the pyramid.

So when linking occurs using a regular static-HTML-page type of setup (a not-blog), weight is quickly added to that story, and it bubbles up the search engine results.

Trackbacks and Comments

For blogs in particular, there's another important and powerful mechanism here: Trackbacks - and in a similar manner, comments. I reckon bloggers have typically been even more powerful than regular news sites, largely due to these.

Reason being: if another blogger links to the story URL, a trackback is often sent to the site hosting the original story. This is a link to the new story by the new blogger that's commonly published on the original site. News sites tend not to generate trackbacks (but have a good search engine rank anyway, as many sites link to news articles). 

Let's say that the trackback link from the new blogger's post is published directly below the article, in the comments section. Next time a search engine bot crawls the page, the trackback has a reciprocal effect, also improving the rank of the linker, which improves the rank of the original article as well, and so on.

Over time, more people link to one of the links that link to my link, so the story improves in search engine ranking, and becomes more discoverable in that search engine (eg, a hit nearer the top of search results for "Tristan's Story").

Historically, for one of my more popular posts, this tends to mean that there's an initial viewing by The Regular Gang, some amount of trackbacks and linking, and then the search engine hits start rolling in a couple of days later, followed by some more links, and so on.

Comments typically allow the commenter to leave a URL with their comment, and Back In The Day, this would be crawled with as much vigour as any other link.

Throwing A Spanner In The Works

Here's where I'm unclear on methods employed by bots - if anyone wants to correct me on specific search engines, The More I'll Know.

Now, after all this "weight" has been built up under the old URL, I change my blog URL to http://blogs.technet.com/tristank/ , and the story URL is now http://blogs.technet.com/tristank/story .

What happens to all that weight?

Well, the weight essentially points at what's more likely to be a dead link now.

In our case, there's a redirect from a given blogs.msdn.com URL to the equivalent blogs.technet.com URL: The connection to the old URL results in an HTTP 302 (I think), and the content is usually still discoverable via the redirect.

But it seems (to me; just a feeling) that bots might not count a redirection to some content's new home as being "as good as" the old content. After all, someone could have taken over the domain and redirected the bot to any old page full of advertisements. I have no idea how clever the bots are with content comparison and checking.

Old Grey Trackbacks and Comments, She Ain't What She Used To Be

To add to the "your URL ain't what it used to be" problem, the landscape has also changed for trackbacks and comment URLs.

Because the system was so open before, trackbacks and comment URLs (where you leave your website address along with a comment you make on a story) became an easy target for link spammers, out to improve their search engine ratings using the same technique as arguably-more-legitimate users. In some cases, bloggers had to spend a significant amount of time deleting *hundreds* of comment spams to a single blog, all made over the period of a few hours. The payoff for the spammers: Better search engine ranking.

To address this (a guess, again), Community Server (and I'm assuming other blog engines) implemented redirection URLs for comments, so instead of leaving an actual URL on a page with a comment or trackback, you leave an address that's looked up internally, and redirected to, rather than a "raw" URL on the page.

And the nail in the coffin: the "NOFOLLOW" tag has been implemented for some of these URLs in many popular blog engines, telling search engines to back off and disregard them.

So whereas with .Text 095, you might have ended up with a link like <A href="blogs.technet.com/tristank">My Link</A> left directly in the comments section - a link with real value to search engines - with Community Server, we now have something closer to <a rel="nofollow" href="www.example.com/blogredirect.aspx?target=http://blogs.technet.com/tristank">, which is fundamentally worthless to search engines: even if they handle the 302, the "nofollow" tells the bot that the link is not to be crawled.

Starting Over

And so, the loyal cadre of regular blog readers still know where to find you, but your old articles probably won't (ever) have the same weight that they used to, so your posts aren't considered as relevant as they used to be, and the search engine traffic tends to dry up (or at least, it has for me so far).

Gradually, as you post stories that people link to, you'll start rebuilding an audience, and a better search engine ranking, and so on.

Comments/trackbacks aren't nearly as valuable these days, so the primary objective is to get linked in the body of the post by another blogger. Meaning, you have to say something interesting that's worth linking to! Meaning, I'm damned. Forever. Ah well, it's not like we're allowed to sell ad space on our corporate blogs anyway...

At the end of the day, the landscape is harsher than before, and if you're blogging for a living, you might want to work out how to keep that clunky old URL (and slap some banner ads on it, quick!).

Comments
  • It was a tough call for me to change domains. When I first started blogging almost 2 years ago, it was something I never expected to last this long (it actually started out w/ me building my own *very* simple blogging engine just to see if I could write one), I made a couple of simple posts, did some trackbacking to other posts...and was amazed at how quickly it took off (for sh*ts and giggles, read some of my first posts...they're TERRIBLE as I was just playing around).

    The domain I used initially was a psuedo domain using dyndns, again I never thought I'd stick with it this long. In hindsite, I should have made the switch a long time ago.

    Here's the kicker though: I wrote an HttpHandler utility to automatically route all requests for my old url's to my new one (the complete url, including the stuff that Telligent changed). Traffic stayed level for about 2 weeks after the switch, then all but disappeared after that, so maybe google didn't like the forwarding? Who knows, but you're right...it's back to square one :-(. Excellent post btw.

  • Quick question: would disabling rel nofollow in CS help out any? I haven't gone through the code to see how Telligent implemented it, but I would think it'd be pretty simple.

  • Hey, Jay!

    This is where some knowledge of search engine mechanics would be really useful, and it probably varies per engine, but here's my thinking:

    Ain't going to help much in the grand scheme of things.

    If I were a search engine vendor, I'd be low-weighting links that were self-referential, otherwise everyone trying to game a search engine would just make a site full of links to itself and wait for a crawl.

    So if you (alone) were to disable the nofollow links in CS, I'd guess that while your site would be more agg-friendly to others, it wouldn't help you personally a whole lot - the real gains come from URLs that point to a different domain from the page on which they're displayed.

    So for example, your disabling nofollow on your site would make my trackback for this article on your site worth more in SEO terms, but it doesn't provide a benefit for comments left on this site.

  • And thinking about it some more: perhaps the MS switchover to CS did have a larger impact than I'd considered on the overall "weight" of Microsoft blogs - perhaps self-referential posts between different directories *do* count.

    ... sigh ...

  • Thats why I use feedburner www.feedburner.com - its great cause you just specify the feed all your users just use this 'proxy' feed.

    - Angus

  • Hi Angus,

    Feedburner looks like it addresses the problem of a shifting RSS feed location and adds features to RSS feeds to boot.

    This might alleviate the initial drop in subscribers when a feed moves, but probably still runs into the "stale link" problem with search engines, within the content.

    The links that are most highly valued (if everything works The Way I Think It Does/Should) are usually the links that are held on sites other than your own, and are usually going to be Permalinks (or MostlyPermaLinks) that are tied to deep-rooted content.

    Unless you were to force all HTML links through a redirect service (like the FWLink system used by Microsoft, for example) - which makes life at least a little more difficult for the blogger - and keep them constantly pointing at the right spot, you've still got the/(my?) question mark over the value ascribed to a redirect by a search engine.

    It could be that for most search bots, redirects are handled gracefully, and it's the NoFollow that's caused the devaluation.

  • 1) I'm surprised you didn't consider the URL change a little more carefully before you leaped. This is, IMO, a well-known issue. And if this is *really* a concern you should be registering www.extrabits.com right this very second.

    2) But really, who cares? Are you writing to satisfy a search engine or an audience? Why not just write to satisfy yourself? Good content gets linked and that's why Google works. Keep writing stuff you like and the rest follows.*

    * Unless you suck. But you don't suck. Much.

  • Ah, Jeff Atwood, Motivational Speaker* - where would I be without you? :)

    1) Yeah, I knew it would be risky moving, but I didn't quite grok the full effects of the change (or set of changes) that took place at the time.

    Well-known, but perhaps not well-understood. And I like doing random things that have very little effect, like yelling "Karl!" at people I don't know.

    I've considered using the slow period to move this blog to another site, but in the end, it's just a work blog. Er, mostly.

    2) Way back when, I originally started blogging to get stuff into the search engines quickly!

    It worked pretty well at the beginning - which leads me to believe that most search engines are dumber than I give them credit for - but it's fun starting over.

    The real problem is that I'm creatively dried up. I'm a lonely water cracker among gifted succulent tomato authors. See? I can't even string together an eloquent metaphor any more.

    Oh, woe is me*.

  • Just getting back at the&amp;nbsp;trackback spammer that infested my classically-late-to-the-party-and-factually-questionable...

Your comment has been posted.   Close
Thank you, your comment requires moderation so it may take a while to appear.   Close
Leave a Comment