Fact - If you build a website based on yesterdays facts, you will rank well for yesterday, not tomorrow :)
I was actually specifically refering to major changes in traffic on Wednesday 25th in particular the travel booking theme sites were hit with a very big filter stick.
I researched possible reason because it wasnt across the board on all terms and compered pages which had been hit with those that wernt. My post on page three outlined my findings. That was my hypothesis which btw requires one amendment and that is that you can have too many internal links with the same link text. I noted one page had over 2000 internal links and also dropped. Those pages with around 80 internal backwar dlinks actually rose. This was fact not hypothesis. I also believe footer links are also not weighted anywhere near what they used to be so the tests I am doing are increasing the internal backward links to a page (destination) that dropped 20 places from 1 backward link to 50 and I am embedding the link in a paragraph rather than link ¦ link2 etc.
It is only through these kind of tests that can provide evidence of a hypothesis/ theroy (call it what you will). The big problem with testing is by the time the new links are weighted and a ranking change noted Google has moved on and tweaked their algo some more. However I believe the drop in serps on Wednesday is down to the fact some internal pages simply didnt have enough internal links with the right link text. I'm also adding an external backward link to the page I am using for testing as it is most likely not just an internal linkage problem but a pr prob as well.
I stress my concern is really only for those sites that dropped on Wednesday not before, as on Wednesday something changed that has affected many travel sites and from what I am hearing other competitive keywords as well.
I'll keep you informed of the tests.
I agree however that it is very easy to come up with a hypothesis. Too often hypothesis posts on this forum remain that because no one actually bothers to put them to the test and publish results. It all ends up in speculation which helps no one.
Because urls, specific examples aren't allowed here you cannot really show/prove your hypothesis anyway.
Interesting (if scary) potential black helicopter tidbit...
New fresh tags just appeared on one of my sites, although only about 40% as many as usual.
Those pages with fresh tags seem to be universally ranking a few spots higher, close to where these pages generally rank. Those pages without fresh tags either stayed the same or dropped a spot or two. Of course it is varying on different datacenters, but I got a suspicion that bad crawling and/or trying to better deal with new/fresh pages is involved in several of the bits of weirdness going on these days.
willbfriendly, nice points overall [but look up theory here [wwWebmasterWorld.com]] and you'll see the definition is slightly different than yours.
Problem with anonymous forums where you can't post either relevant urls or search terms, nor can you know what type of site most posters are referring to in their comments, it's not really possible to ever empirically verify anything any poster says, with a few exceptions where they list their sites in their profiles, which most seo types don't do for obvious reasons. This means you have to try to figure out what each poster actually does, what type of sites they actually run, whether they are serious one site hobbiests or pro seo types etc, what category of keywords they are testing, whether those are the keywords being run more or less straight, or through the various new filters that have been installed.
All in all not the most ideal environment for applying the scientific method, sort of like if anyone could submit anything to Nature and get it published, leaving it up to the readers to determine what was empirical fact and what was conjecture. This is what makes this forum especially entertaining though.
Here is a week by week stat display for one of my PR6 sites. These are unique visitors from google by week. It is very easy to see where algo changes happened. My rankings have returned to April levels.
|Fact - If you build a website based on yesterdays facts, you will rank well for yesterday, not tomorrow |
I understand your point that the algo is always subject to change, but the fact is that my site was optimized and I rank #1 for my targeted keywords last week, and I still do today, have been for months in fact. The algo doesn't dramatically change everyday, so using historical data to optimize is and has been very successful both to webmasters and SEO companies.
When tweeks occur, we adjust and get back on top until the next tweek.
Why bother lurking here if you think it's futile to try to adjust to the changing algo?
<<<<Hmm, well, we were doing very well for a keyphrase for the last month, now its back on the 3rd page from a #2 result.
You have to search for the good results after doing a search.>>>>
And last month when you were #2 the search was perfect?
Something strange is afoot. My home page was nicely up on SERPS with a cache date of 8/25. Today I note that it now just has the generic "cached" tag, marked as cached on 8/17 (!), and has dropped a few places.
Observation...yes. Meaning...who knows?
I offer this regarding theories - [teacher.nsrl.rochester.edu...]
"A scientific theory or law represents an hypothesis, or a group of related hypotheses, which has been confirmed through repeated experimental tests."
But, we drift far off topic I think.
that's right if you're a scientist, not for normal english though, so it's whichever one prefers, standard or scientific def, no argument.
Your main point is right though, but solid conclusions cannot in general be reached in the absence of actual data, whether keywords or urls, many questions could be resolved very quickly given the quality of some of the posters here if actual data could be used.
There's also the problem that some posters simply don't want to use empirical methods, for whatever reasons, I've seen this a few times where one says something that is easy to test, and empirically verifiable, that offends a belief or fondness for a company/site, but which is still objectively true, and people jump all over it until you conclusively demonstrate the truth of the claim, rather than testing it themselves first then responding. But I guess that's part of the game here.
Something has happened, but no one has a clue what.
Some people are affected and some aren't and no one has a clue why.
Some people are upset by the change, and others can't even see a change.
Oh, and there has been a debate about what it is to have a theory vs. a hypothesis.
there is a disturbance in the force.
Its not clear from the thread but this disturbance seems to happen about this time every month.
It's almost the full moon, in case anyone missed what time of the month was being referred to.
There are always periods in a month where forces get disturbed.
Sorry - could not resist it;)
Personally, I believe the recent "update" ... for lack of a better word, was another "keyword density" algo tweek. The difference between this one and Florida is that this one was very sensitive.
I found four or five pages in my site (which haven't been changed since Florida) which were suddenly penalized. The only possible reason was keyword density.
Do a site:yourdomain.com search and see what you find. Good luck!
Liane, which kw density value are you talking about?
Just to add an international flavour. I am running about 15 websites related to Thailand. My two main sites, both 'hand-made' extensive web directories, have been seriously hit by the latest Google rearrangements. It happened on 25 August and I had seen it coming, because for the last few weeks Google never visited at all (before there were daily visits). As a result, traffic dropped by about 70% overnight. My smaller websites are doing as before, and actually a few of them are picking up traffic. One site, somewhat of a mirror site of my two main sites, and a site I hardly ever look at any more, now starts getting more traffic. Interestingly, and likely coincidental, the two main sites affected are hosted in U.S. with a well known 'asp' sites webhost, that also has does a lot of free webhosting. My other sites are on other servers. It was mentioned that the latest update was travel related, and possibly Thailand-related websites fall in that category. I have Google advertisements running on the site, and certainly Google will have a few hundreds dollars less income now from my sites, due to the much decreased traffic.
Well, here are my experiences...
I have two websites relating to a sport. One site focusses on news relating to players in the sport, and the other focusses on statistics relating to the very same players.
Both sites used to rank well for searches of player names (especially the less famous players) and regularly ranked in the top 5. This is really almost entirely off the strength of internal anchor text as these pages don't get inbounds from outside the site generally.
In addition, each stats page for a player links to the news page, and vice versa.
Both sites are served on the same server and same IP.
Now, the stats site doesn't feature at all when searching for player pages, but the news site is as good if not better than ever.
Both have similar on-page optimisation, e.g. h1, a mention in bold etc. and both have very similar internal linking strucutre. The stats site homepage is PR6 with a dmoz link, and the news site has no dmoz link and the home page is PR5.
I can understand the loss of position on the stats pages if something has changed (not that I'm happy of course) but it's strange that it didn't equally affect the news site given the similarites in structure.
Keep in mind the content on the pages is totally different and there is no question of a duplicate content filter.
These pages are absolutely nowhere near travel of course.
I personally lean towards some sort of change to the value of internal anchor text rather than keyword density, but it's clearly nothing too simple.
When looking at the posts in this forum I get the feeling that google simply expands filters/algos to other areas as well.
Some things I read have been a no-no since beginning of this year or even Florida.
I've been following threads like this one since last week and suggested some patterns I've noticed. I think I just had a new insight that maybe some others can check on their results and concur or not.
We had a few primary pages drop from top 10-20 down to 100+ like others on here. I suddenly realized that the only pages we had affected were ones which were significantly 301 redirected TO! Is it possible G could suddenly be penalizing for 301 redirects? We 301 redirect most pages we remove, which are still backlinked by outside domains, to our home page, and it was majorly effected Wednesday dropping from #14 to ~#100. We also autorefresh our 404 page to our home page after n seconds. We 301 redirect another smaller group of deleted pages to a different page and it recently went from #2 to #18. Most others remained unchanged. Coincidence? I seem to recall there was some talk recently of blackhatters exploiting a 301 redirect bug to gain rank, could this be G fighting back and in the cross-fire catching innocent sites using it honestly?
BTW, I just noticed in the last 15 minutes that the page I mentioned above which went from #2-18 since last Wednesday is now back up to #5. Maybe the rest will return shortly!
Can you be more specific DirkZ on WHICH things have been no-nos?
Yes, i was thinking the same thing as too i have 301 on major pages as the urls were chnaged due to new layout. But in my case pages that were not having 301 dropped too, i guess i will remove the redirects and see what happens.
I have seen a slight increase in traffic today too, i cant say for sure until tomm though. I lost around 80% of my traffic 25th...i really hope it comes back.
I think google have recently been turning down the 'weight' of variation words (stems/synonyms) and thus increased the weight of exact-matching words.
At least, this is how it appears in my field -- mine and my competitors sites do not change much over time (well, certainly haven't changed much recently), so all other factors remain about equal.
I posted a longer message on this subject in another thread (in which it is probably a bit off topic [shrug]).
How does the above idea fit in with the changes people are experiencing / have experienced in their ranking?
Maybe it's not the only factor at work, but I've definitely seen changes in stemming/variation/synonym weighting over the last weeks.
I MIGHT be inclined to agree somewhat with Frogg if you define what "exact-matching" means. In our case we dropped majorly for "X Y" whereas most of our backlink anchors are of the form "X Y Z" "W X Y" "X Y A". Insisting on an "exact TRUNCATED match" of this type would be terrible, since, maybe the searcher doesn't KNOW what kinds of "X Y" widgets there are out there, and only those sites with only cursory info about X Y and very little specific info will show up under a general search for "X Y". This would also have an unfair advantage for keyword only domains since a large proportion on anchors to them will simply be their domain name: XY.com
|Liane, which kw density value are you talking about? |
On page plus title.
I have been reading through this forum and have seen a lot of interesting posts. I find it weird how a lot of Travel related sites are getting hit. I also have a travel related website, though it is an informative website.
My stats show about a 30% decrease of my normal traffic on August 25th, the day everybody seems to have been getting hit. The thing that kind of gets on my nerves is that on August 16th I saw a 30% increase in traffic for about 3 days. Then my traffic went back to normal on the 19th before dropping by 30% on the 25th.
Those are very inconsistant stats. I haven't really seen anybody else talking about any significant increase on the 16th. I thought I would mention it. Google must be testing something. Two major shifts like that isn't normal.
Has anybody else seen this?
I had an increase 4-5 august till 25 august and then it went down :(
Our travel site has been hit hard. Index page has been at 3-7 for about 3 years now down to 13. Sub pages that had been at 1-5 for years now are not in the top 50. UGH!
To me it seems like more "authority sites" (based on incoming links not content) may be rising. These sites have lots of incoming links because they are huge corporate sites, the smaller sites sites/companies are getting squeezed out.
For the sites/keywords I watch, I have not seen that KW density has much to do with it yet, it seems mainly to be incoming links.
Welcome to WebMasterWorld Makaveli...
I just lost my homepage in the SERPs for several phrases, however, those same phrases are pulling up sub-pages better ranked then they were before.
One other thing I noticed (for one of my sites anyway) is that links to my site are ranking higher then my site. In fact, on my money phrase, a site with a link to mine is popping up #3 while I am #534 with a sub-page. When I examine the site linking to me, the only instance of the money phrase is the anchor in my link. So what's up with that?
The main page had a 301 redirect that I had to remove because for some reason googlebot was not following the redirect to the new page. I believe this will straighten out with time. The PR on the new home page is still the same as before, just not listed in the results. I have posted several times stating that 301's were not being crawled correctly and I have seen other posts that are starting to point in that direction. I have removed all 301's and now am starting over.
Is this making sense?
Not to me!
The PR display in the Google Directory just updated, again. New today, not the one that occured a couple weeks ago.
Some datacenters at least, like 220.127.116.11 as in [18.104.22.168...]
[edited by: steveb at 2:24 am (utc) on Aug. 31, 2004]
Local search is in beta and has been appearing/disappearing in some keyword spaces for several months. Did it just turn up in travel related searches?
| This 130 message thread spans 5 pages: < < 130 ( 1 2  4 5 ) > > |