Help forum Feed

The canonical  tag on www/masspublisher.com has a fully qualified link to "www.masspublisher.com".  this should in effect, according to google, act like a 301.

http://googlewebmastercentral.blogspot.com/2009/02/specify-your-canonical.html
"Can this link tag be used to suggest a canonical URL on a completely different domain?
No. To migrate to a completely different domain, permanent (301) redirects are more appropriate. Google currently will take canonicalization suggestions into account across subdomains (or within a domain), but not across domains. So site owners can suggest www.example.com vs. example.com vs. help.example.com, but not example.com vs. example-widgets.com."

I believe the site, a few years ago, was named SWIMBO.  when it was renamed to MassPublisher, the 301 was put in place to direct past users to the new site.

would this cause a problem?  we can remove the 301 now.  would it cause a problem?

Phil - Thanks for your answers.  We're just trying to figure this out, so if anything we've said here seems wrong / off base, let us know.  Thanks again.

I just noticed that we seem to no longer have any links to the site according to google "link:" command (yesterday, we had 50, today we have none)

link:www.masspublisher.com
link:masspublisher.com

what happened?  why did google flush all links to our site?  will they be re-built over time?  This is probably directly related to why the site vanished from almost all searches

It's odd, when you look at "site:masspublisher.com", our site's content is still indexed "about 7,870 from masspublisher.com" but our pages show up near the bottom of any search they should be near the top of.

Example, "Robby's Code Snippets", yesterday without quotation marks, masspublisher.com profile was number 1, now it's nowhere... yet if you do the same search with the quotation marks, it CAN find the page.

my new question... is it possible to experience the google "sandboxed" effect on an existing site that has been running fine for a year?  the effect allegedly artificially pushes your site down in the rankings it would usually appear high in.  Allegedly it does this on new sites until they have gained enough "trust" to get there true organic ranking.  but would it do this on an existing site that has not changed much in the in the past few months?

http://www.masspublisher.com/content/profile.php?ItemID=35103&PublicMode=true

I think you guys snarked your massive site update.  How extensive are the differences between the site today and the way it was structured on 10/18/2009?

Michael - Not exactly sure what you mean, but that site update was made this evening, AFTER the problem occured last night / this morning.  The change itself impacted almost nothing that Google would see.  Thanks, though...

If the general question is whether we did a great big site update before all this happened, no.  We did have a problem with our XML sitemap (it was not found by Google for some reason) but we reposted and it was fine.  What was IN the sitemap hadn't changed too much in several weeks

"The change itself impacted almost nothing that Google would see."

Sorry.  Hard to see what happens in retrospect.  The loss of your link profile might explain what's going on, then.  Even Yahoo! doesn't report many links for your site.  How many did you used to have?

We only recently launched.  There were 50 or so links if you checked "link:masspublisher.com" on Sunday.  Now, 0.  I get the ranking change if we were caught in Google update or something, but how can all your links vanish? It's just freakin' odd.

Autocrat -- Thanks for the note.  I've taken a look at Yahoo Explorer several times.  We launched in late July and Yahoo appears to the be last to index us, and their link counts appear to be lower than Google's were but rising.

To answer your question, the site is about 3 months old, thought it's been up with robots.txt blocking crawlers for more than a year.

To answer the 2nd question, no.  There are still several links out there from sites that you can find if you search for "masspublisher.com" on Google, but none of them appear on "link:masspublisher.com" which seems to say that Google is simply wrong.  It's very strange.

Thanks for the help.  Any other thoughts you have are more than welcome.

Google's "link:" operator only shows a random selection of sites that Google knows about.  They apparently update the data that the "link:" operator reports from on a periodic/occasional basis.

It could be that your site is just too young, has too few stable links pointing to it, and hasn't earned enough trust to be stable.  It could also be you've tripped a filter.

A colleague posted this response separately to our problem, and I thought it might be useful:

It's hard to know exactly what happened unless you had someone do a thorough SEO analysis of your site a month or two ago, and then redid a thorough analysis today and then you could compare the two to see the differences.

But here are 4 possibilities:

1. Google made a big change to their algorithm. I haven't seen any big changes to any other sites' rankings that I'm aware of, but it's possible that Google tweaked something specifically related to your industry.

-- Possible.  Anyone have any idea here?   Did Google update their algorythm this month?

2. You changed something on your site that Google didn't like. Removed content, changed the internal linking structure, renamed pages, modified the home page or sitemap, changed the software you are using to make the site, moved your site to a new server, something like that.

-- No idea.  We're always changing the site, but NOTHING major changed during this time period.  We did change to a faster server with our host, but would that really do this?

3. You lost links from sites that had previously linked to you. Like maybe you lost links from www.sbiccabistro.com for example. Or the sites that link to you did something on their website that affected how Google views those links to your site.

-- We did not lose any of our main (few links).  We did seem to be found by some spammy link sites, but we haven't bought or intentionally listed on any of these sites.

4. You got links to your site that Google didn't like. For example, links from a spammy porn/casino/viagra website. Or you bought some links, which Google doesn't like, and Google caught you.

-- Nope.  Not our kind of stuff, and we don't buy links.

I would watch your rankings in Yahoo and Bing over the next few weeks and see if you go down in the rankings in those search engines as well. If you do, it's probably problem 2, 3, or 4. If it's only Google that you lost rankings in, then it might be problem 1.

-- So far Yahoo! and Bing are find and continue to improve...

So...help!

i think your site is not update time by time.Google marks the site which is updated by time to time. so you should update your site by time to time as to retain the position in google search ranking.

All - Some interesting articles that related to all of this:

- http://www.articlesbase.com/computers-articles/over-optimization-and-the-google-sandbox-1088424.html
- http://www.seobook.com/archives/001230.shtml
- http://www.earnersblog.com/aging-filter-tips/

At this point, to be honest, it seems less like there is a problem with us than there is with Google itself.  It's one thing to be secretive and careful with your algorythm, and to protect yourself from spammers and other bad actors.  It's another to offer a product to the market, to dominate that market, and to foist responsibility for customer service off on your customers.  Is it seriously that hard to offer suggestions to customers like "your site may have a few problems like.." without revealing anything proprietary (beyond the crawler / html notes in Webmaster).  Perhaps this is the behavior one should expect from an abusive monopoly rather than a good market competitor?  Just a thought.  I wonder if there's a sandbox penalty for that too.

That aside, thank you to everyone so far for their comments.  We'll keep updating this as we get new information and when, if ever, there's a resolution.

Autocrat - The canonical changes were made about 3 weeks ago.  The duplicate content stuff is gradually decreasing, but thanks for pointing these out.  Do you agree that (a) we have a canonical problem (and what specifically?) and (b) do you think it could cause the symptoms we're seeing?

I just checked source.  We put the canonical tag in place on august 31, so 8 weeks ago.  and google crawls the site a minimum of 3000 times a day, so it has known about it for a while.

The duplicated search is interesting.  If you follow both links, one is correct www.masspublisher.com, the other link has the canonical pointing to the www.masspublisher.com version.  the canonical tag is always fully qualified to point at www.masspublisher.com.  My guess is that google has not de-duplicated the links in that specific case yet.  To my knowledge, this is exactly what the canonical was designed fix.

Thanks for your help!

Robby - 3 weeks, 8 weeks, whatever ;-)  Thanks for correcting that

Autocrat - Thanks for the response.  It's good to see the positive information as well as the negative.  I appreciate it.  I always understood that the "honeymoon" was a few days after new content was posted.  For us, if that's the case, it was months.

John -- First, I'd like to thank you for commenting on our site and taking a look at what might be wrong.  Thank you.  That's really all I was after in the first place -- some way of knowing what was wrong so that we could address it, and some way to not spend hours/days/weeks tilting and SEO windmills only to find they have no impact or are even negative.

That said, I'm not sure I understand why this would result in de-ranking of pages that DO have content on them, but I think what you're saying is that if Google's system "chokes" (my word; there's probably a better one) on sites with a lot of pages that (perhaps) should be ranked poorly, and therefore de-ranks the whole site (including pages that should rank highly).  That seems like an odd practice given that most people's sites are not uniformly perfect or flawed, but at least it makes some sort of technical sense -- sort of the hammer response to spammy / empty site problems.

My guess is that you have a (1) ingest all (2) index/rank all (3) review what's indexed (4) go deeper into indexed pages and re-rank (5) if overall site seems to have a lot of bad pages, de-rank entire site (6) slower page-by-page evaluation process to restore rank to individual pages (7) ongoing crawling / indexing / ranking (8) at some point the site becomes "trusted" and overall ranks more highly.  Totally guessing here, but it makes logical sense.

If that's the case, then I would interpret what you're suggesting to say that we've hit point "5" b/c we have a lot of low-text pages, and that if we work on our content on these pages we'll start to rank higher, slowly in "6" regain individual page rank and eventually "8" we'll stop being penalized overall for having emptyish pages.  If that's the case, we'll start working on this immediately and, again thank you.

But, and it's a big but, this doesn't make a great deal of sense.  Any reasonable business person who wants to start a consumer portal for business reviews, information, etc., is probably going to go out and BUY a great big list of (say) restaurants and put them online so people can review them.  This means that ANY startup review sites (kind of like ours) will by definition have mostly empty pages until we can generate traffic (via your search) to get users to make comments.

Given that this represents an entire class of very large and very small internet players, it seems like your system should recognize these kinds of sites and rank them RELATIVELY with the ASSUMPTION that they do and always will have a lot of empty pages.  I mean, look at UrbanSpoon.com and others - half these pages are empty of any real content.  As they should be until someone contributes reviews, etc.  Penalizing us for having empty pages (that are business profiles waiting for reviews) is basically just punishing any small business or startup that dares to challenge the dominance of large community sites or provide truly local content.

At the very least, it seems like Google should be careful not to penalize content-heavy pages that happen to be on low-density sites.  That way, at least, there's a way for us to crawl (pun intended) up to a trusted level in a methodical, linear manner without bouncing all over the rankings.  There's just too much noise in all of this to make good business decisions w/o having to hire a team of SEO people to sculpt our site -- which seems like the exact thing Google would not want.

All that aside, my focus is really just on running our business.  I appreciate the input you've provided and will work to improve our site in light of you suggestion.  I hope you'll take my comments as simple input from someone who just wants to run a business in interesting economic times w/o having to earn a PhD in SEO...

Thanks again.

Autocrat - Well, I'm no expert so I'll submit to your greater experience on search, and agree on the user experience side.  Thanks.

John -

Thanks again.  I really appreciate the comment, and I completely agree.  So much easier to move forward now that we understand the issue. Here's what we're doing:

1. 99% of the link types that you showed in your list are auto-generated on profiles, whether there is content to follow or not.  Just bad design for users, really.  We are either making these disappear when there is no content, or if we can't do that, making them No Follow until there is content.  For instance, the Community and Maps pages will all become No Follow until there's something there.

2. We found a canonical problem or two, and are fixing those

When this is all done, I'd say by tomorrow, our site should look much, much more clean to Google.  I'll post an update here when it's done and we've tested the new confifugation.  I don't expect any special quicky "reevaluation" by Google then, but then again it couldn't hurt ;-).

BTW, great link to Matt's blog entry.  Helps me better understand the exact issue from both sides.

Thanks again,
Shawn

Michael --

Thanks.  Interesting, though not 100% sure I understand.  We'll be using No Follow on dynamic links to sub-pages of the (non-empty but not very content rich) stub pages that are truly empty until there is content on them, not on the stub pages themselves.  Once there is content on the sub-stub pages, the no follow tags would go away.

Can you tell me a bit more about how the nofollow tag feels like a penalty.  I did a bit of searching online, but want to make sure we fully understand.

Thanks,
Shawn

Just read this and relax about the whole thing:

http://pythongoldmine.vndv.com/Convert/We_are_here_to_change_the_world_01.html

This one talks about the situation of this kind:

http://jsgoldmine.uuuq.com/Convert/zzGoogle_index_drops_like_a_rock_23_Ed.html