| 4:44 am on Dec 8, 2007 (gmt 0)|
My site is -950'd for pages that only have 1 item, so I just don't think it's keyword repetition. For example one -950'd second-tier page looks like this:
Title: "My Site Name - Small Widget"
Main Body Text:
Browse ¦ S ¦ Small Widget
Small Widget is made up of blah, blah, blah, blah
Small Widget with brown whozawhatzits from ThisCompany."
And that's it. I mean, there's a navbar at the bottom and some text info boxes, but nothing that in any way relates to the -950'd phrase "Small Widget" -- they're just site information. So is repeating a two-word combo 4 times on the page and 1 time in the page title really strong enough to trigger a -950 penalty? I just can't imagine keyword repetition is the cause... at least in this case.
[edited by: ALbino at 4:45 am (utc) on Dec. 8, 2007]
| 4:46 am on Dec 8, 2007 (gmt 0)|
This -950 critter just loves URLs with a lot of links. It eats the links and punishes the URLs, so I think less links per page are a better solution, it also means less text per page which might reduce co-occurrance.
If it does, then its appetite is pretty random. The #1 site in one of my niches (which has site links) and gets links through questionable means (link counter spam, mainly) has 1800 links on its homepage at least, it may be upwards of 2000 if I included everything.
Also, the links all what I'd call spammy. They look like this:
"Product - The best www.example.com site on the planet"
Nearly every link has the phrase "The best www.example.com site on the planet" on the homepage. It's repeated in at least 60% of the links. You'd have to see it to believe it how miserable that looks but google loves it. Ranks #1-#3 for every keyword.
| 5:09 am on Dec 8, 2007 (gmt 0)|
How about excessive number of occurrences of keywords in navigation links?
I noticed some sites seriously hit who had just that, and even got it down to a percentage limitation at the point when I'd seen enough of them. Sure, there are other factors in the "illness", but that's the most compelling symptom I saw on a good number of sites.
| 9:17 am on Dec 8, 2007 (gmt 0)|
Much to everyone's help in this forum, my main site has recovered! Just had 60+ phrases return to the top 2 pages a few days ago, and I am positive the penalty was due to over-optimizing the internal navigation.
I removed a good chunk of our footer about 10 days ago, and noticed a handful of slightly competitive and semantically related terms show up in the top 10. Upon seeing this I took out the shotgun and completely removed the footer, and lo and behold everything is back!
What this tells me, which hopefully may apply to some of your situations as well:
1) Removing secondary navigation can help - this confirms what others have theorized regarding "link weight".
2) Changes to your rankings occur right after the next crawl of your site.
3) The filter is phrase-based - if you make a change and notice some pages coming back, those were likely on the edge of the filter and you just did something right.
Thanks so much to everyone on this forum for sharing your thoughts and ideas!
| 3:36 pm on Dec 8, 2007 (gmt 0)|
|I completely removed the footer. |
Did it have only links to internal pages? That would seem a bit drastic if so. I thinned mine, removing several internal links, including ones that had psynonym keywords, or repeated keywords from header navigation links, but I'm not removing all of them. Users should be able to find a link to the home page, if nothing else, in a footer, like this site and many/most others. I removed the keyword in the footer home link and replaced it with the neutral "home" text to avoid potential hassles from suspected "overoptimization."
P.S. I've suspected for a while the 950 lifts after regularly scheduled crawls, not reinclusion requests, etc.
| 8:18 pm on Dec 10, 2007 (gmt 0)|
Yes it only contained links to internal pages - though this was a supplementary navigation and many of the links already existed on the page through a primary left menu.
| 8:36 pm on Dec 10, 2007 (gmt 0)|
It seems like I'm getting a little recovery as well. Not a full recovery, but seems like pages are creeping back up to around the top 20 or so. Still down, but not forgotten.
Over the weekend, Google gave me the option to increase my crawl rate to "faster" - does this help at all with getting pages back? Is this a good thing? I chose it, of course, because Yahoo hosts my site and its not my bandwidth google is sucking :)
| 12:41 am on Dec 11, 2007 (gmt 0)|
Has anyone who suffered this 'penalty' have a site that targeted two separate (and maybe even unrelated) keyword phrases and only have 1 set negatively affected?
| 1:01 am on Dec 11, 2007 (gmt 0)|
I think many of us are targeting a very wide variety of two-word phrases that are only vaguely related of which some are affected and some aren't.
| 1:55 am on Dec 11, 2007 (gmt 0)|
"I think many of us are targeting a very wide variety of two-word phrases that are only vaguely related of which some are affected and some aren't. "
So if this is the case for others, then this may be a clue. I have a site that has two separate and unrelated sets of keyword phrases, and this Dec Google whack only affected 1 set of keywords, the other set still ranks great.
Anyone else see this?
| 3:07 am on Dec 11, 2007 (gmt 0)|
Multi-topical site...affected...need I say more.
| 11:45 am on Dec 11, 2007 (gmt 0)|
|I think many of us are targeting a very wide variety of two-word phrases that are only vaguely related of which some are affected and some aren't. |
Matt Cutts says the 950 is an "overoptimization" penalty. So you could assume that's the primary issue. Keywords and keyword phrases obviously are the traditional targets of optimization.
The conclusion of many here is that overoptimization of keywords/phrases in internal linking, i.e., navigation keyword stuffing/navigation spamming causes the 950 penalty.
If you target many related phrases, and thereby use repetition of the keywords or phrases in your internal links, you can certainly expect to get hit with the 950.
Although it must be noted some sites doing that still haven't been squarely hit. However, that could change at any time with the turn of a dial, or some other questionable aspect of your site Google finds.
I was breaking the unwritten rule for months and then finally got the 950.
P.S. The only two- and three-word phrases that survived the 950, with few exceptions, are the ones for which I was #1.
| 3:52 pm on Dec 11, 2007 (gmt 0)|
I'd also add that different sets of semantically related phrases seem to have different thresholds - which may be why one set of phrases is 950'd and one set isn't.
The threshold seems to be based on the competitiveness of the phrase (in my experience at least).
| 4:58 pm on Dec 11, 2007 (gmt 0)|
|I'd also add that different sets of semantically related phrases seem to have different thresholds - which may be why one set of phrases is 950'd and one set isn't. The threshold seems to be based on the competitiveness of the phrase (in my experience at least). |
Good post. That would make sense b/c the most competitive phrases are most likely to get overoptimized.
I wish we had better insight into the phrase-based SPAM-detection patent.
Despite everything I've read here, I still have no idea which phrases Google will see as positive and which as negative; which are predicted to be found, and which will be seen as spam.
In the past, there was a keyword density issue we had to focus on. Now do we have to worry about keyword-phrase density?
Google, in recent months, seems to have divided its algo into at least two--one for single keywords and one for the rest.
Perhaps it will divide (or already is dividing) again, with one algo for keywords, one for keyword phrases, and one for all the rest.
| 8:32 am on Dec 12, 2007 (gmt 0)|
"The conclusion of many here is that overoptimization of keywords/phrases in internal linking, i.e., navigation keyword stuffing/navigation spamming causes the 950 penalty."
It's always puzzling and interesting to read a lot of experiences here. I almost never see any SITES get 950ed anymore, so obviously keyword repetition in navigation has nothing to do with penalties to single pages, or several pages on a domain.
| 8:15 pm on Dec 12, 2007 (gmt 0)|
|How about excessive number of occurrences of keywords in navigation links? |
Especially when the target pages are very thin. What's the justification to an SE for many internal links full of keywords to virtually non-existent pages?
If nobody else backs the page (inbound links), it's unjustifiable and primed for a 950 hit.
| 4:12 pm on Dec 14, 2007 (gmt 0)|
I came back today - not a 100% recovery but around 90%. Some pages are still below where they were. But still. Best Christmas present ever.
From my count, I dropped for exactly 60 days. Hard to tell cause and effect, of course but I felt mine was a part duplicate content (I had content stolen from me), maybe over-optimization of the navigation (though I doubt it, it was never that bad).
Yesterday I was looking at the source code and my base href tag (or whatever its called) was gone since our redesign. I put it back in yesterday and today I'm back - makes me wonder if I had noticed it earlier would I not have been gone 60 days :/
On the plus side, it gave me a chance to get my PPC campaigns profitable so if it happens in the future, it's not a wipe out.
| 5:00 pm on Dec 14, 2007 (gmt 0)|
Duplicate keywords in anchor text links is the new form of duplicate content Google hates.
|I put it back in yesterday and today I'm back |
Do your logs show Google crawled your site in the last 24 hours? I'm not sure that issue/fix would lift the 950. Don't know that anyone else had a similar experience reported here. But fixes of all kinds are good for the Web.
| 5:19 pm on Dec 14, 2007 (gmt 0)|
Do your logs show Google crawled your site in the last 24 hours?
I'm a yahoo store. I don't have any real log files. WMT shows the last crawl of the homepage on the 12th; but that isn't updated in real time as far as I know.
What's interesting is that, now, my WMT shows:
Pages from your site are included in Google's index. See Index stats.
Pages in your Sitemap(s) are included in Google's index. See Sitemaps overview.
Before, the second thing (pages from your sitemap....) wasn't there. It just showed up this morning.
| 9:54 pm on Dec 14, 2007 (gmt 0)|
We think we may have fixed some over optimization on our site. We had been hit for many months before we were able to indentify the problem.
Assuming we fixed the problems, will google pick up the changes quickly and put us back to our previous postitions? Do we need to initiate any reinclusion request (we were always indexed, just got killed on ranking)
| 10:49 pm on Dec 14, 2007 (gmt 0)|
-- Duplicate keywords in anchor text links is the new form of duplicate content Google hates. --
You mean Like this?
<a...>Unleash the fury Mitch, Unleashed fury </a >?
| 10:54 pm on Dec 14, 2007 (gmt 0)|
I think problems come more like this - on every page and targeting a competitive keyword (1.e. not 'fury')
<a href=...>winter's fury</a>
<a href=...>tiger's fury</a>
<a href=...>hell's fury</a>
<a href=...>fury of vengeance</a>
<a href=...>sound and fury</a>
<a href=...>full of fury</a>
| 11:25 pm on Dec 17, 2007 (gmt 0)|
One site is back in top 5 (near its reg SERP of #2) today after it had slipped to about #81.
Don't ask why. The site does have issues (it was built years ago, pre-SEO self-education) and does bounce occasionally for weeks/months on end.
Its PR5 may help push it back higher. Who knows. I'd noticed about a week ago when the home page was #81, one page had been 950d.
The PR juice is mostly from inbound links, which have been known to lift 950 penalties, so that could help explain the comeback.
I was getting ready to fix its spammy internal links, etc., but now may not bother if it holds.
In other Google Search News, I just saw an article site has been 950d! Now let's hope this is the start of the long-overdue Google Article Link Farm Spam Elimination Campaign. Instant experts be gone! 8->
| 6:02 am on Dec 18, 2007 (gmt 0)|
Well, I've been looking at a site that seems to be ranking because of the repetition of a keyword in the anchor text, and it's doing so well that it's hard to get the client (or me) to consider changing it. Removing the keyword repetition on several test pages did, in fact, cause those pages to drop significantly for searches that included the keyword.
I'm wondering, therefore, whether, like many Google filters, the level of competition is a factor... and that phrases with relatively low competition levels are below the threshold of the filters. A very sloppy measure of competition level on this particular set of phrases puts pages returned containing exact phrase matches at between 50K to 150K matches, not super competitive.
Anyone else relating this apparent anchor text filter to the competition levels? I'm guessing it might kick in at higher page counts. Google counts have been so all over the place, though, that it's hard to trust the numbers.
| 6:10 am on Dec 18, 2007 (gmt 0)|
I agree that something like the level of competition is a factor. In fact, the way the phrase-based indexing patents are written, the thresholds for spam-detection penalties would naturally vary in different keyword sectors, according to common practices in that area. I've posted this summary before, but I think these ideas bear repeating:
|1. The re-ranking is triggered by crossing a threshold. |
2. The threshold can be different for different search terms.
3. The threshold can be different for different markets or website taxonomies.
4. The threshold is set by measuring and combining many different types of
mark-up and grammatical factors, and not by absolutely measuring any
5. The threshold is NOT set absolutely across all web documents. So phrases
in the travel space can be held to a different measure than, say, phrases
in jewelry e-commerce.
| 7:39 am on Dec 18, 2007 (gmt 0)|
Expanding on point 1 above... as the web grows and niches expand, it's likely that a phrase that was below the radar eventually will become competitive enough that it will rise above a threshold and get hit by filtering.
As has been discussed, it's becoming increasingly important to anticipate these levels or triggers, whatever they might be.
[edited by: Robert_Charlton at 7:43 am (utc) on Dec. 18, 2007]
| 3:33 pm on Dec 18, 2007 (gmt 0)|
Back to oblivion for my site again this morning. This is fun. I had a good 4 days or whatever, but now I'm back to where it was before. This is driving me nuts.
| 8:41 am on Dec 22, 2007 (gmt 0)|
Yeah, the unpredictable bouncing into and out of the 950 is making me lean towards a brand new domain and new site.
You don't know:
1. If/when you'll be 950d.
2. If/when you'll be un-950d.
3. If/when you'll remain un-950d.
All the while you have to guess what you've done wrong, what corrections will work. It gets really silly. How do you know that your old site will be back to normal by the time you've got a new domain and site up and running?
I'll keep the old one up, but if the Google Spam Gestapo refuses to explain what it's doing, how much more time does it expect me to waste p*ssing around, fiddling, revising, whatever.
What great plans does Google have for 2008? Ten new 950s? Who's the knucklehead who came up with the 950 in the first place? He or she will have to hope karma doesn't bite.
| 9:20 am on Dec 22, 2007 (gmt 0)|
The 950 penalty is knuckleheaded not because it is a penalty -- obviously plenty of pages online should be penalized.
It is knuckleheaded because its implementation is incompetent.
"Let's penalized pages that deserve it" is not bad.
"Let's make a penalty that we totally suck at implementing" is bad.
| 11:48 pm on Dec 22, 2007 (gmt 0)|
Yeah, the unpredictable bouncing into and out of the 950 is making me lean towards a brand new domain and new site.
I'm not convinced this is a solution. A new domain may not fix the problem; then you have to work a long time to "age" the domain (if domain age is relevant at all - not sure), work all your inbounds again when, whatever is going on, could still be some sorry mad scientist experiment which Google (I believe) does not quite understand.
If it is, indeed, something onsite, then the problem is still there.
To answer your previous question about "whether the logs show google hit my site" in the recovery on the 14th. WMT eventually showed that Google hit my site on the 14th (recovery date) and again on the 18th (the date of most recent demise).
| 12:42 am on Dec 23, 2007 (gmt 0)|
Even if your domain ages, one day it will get kicked by the -950 nonsense.
So it's time to look for short term success.
The Google spam team did a poor job implementing this penalty.
[edited by: SEOPTI at 12:46 am (utc) on Dec. 23, 2007]
| This 199 message thread spans 7 pages: < < 199 ( 1 2 3 4  6 7 ) > > |