Welcome to WebmasterWorld Guest from 18.104.22.168
Forum Moderators: open
Many webmasters in the thread above are reporting significant drops in traffic, up to 80% in some cases within the last 24 hours. In contrast others are reporting significant rises in the SERPs which indicates a big change (which imo is not sandbox related).
Rather than having a "I've dropped/risen too" thread, perhaps it is a good idea to attempt to work out what has been happening. I personally have seen big changes on numerous travel booking related sites. One of the theories appears to be too high keyword density. This may have some merit, as on some sites ranking drops are not across the board on all themes/phrases. So it could be that templates focusing on one term have a much higher density than others. I'll certainly be looking into this. I'm also wondering if this is somehow theme related as as I've mentioned it defintely has affected the travel theme. or perhaps it is based on just highly competitive terms as a whole.
It would also be interesting to see if there is a concensus from those sites that have lost traffic on..
1. Are they affiliate sites with many links to the affilate host.
2. Do they have satelite domains which they cross link/one way link with sites particularly on the same ip c-block.
3. Are the drop in serps keyword specific or site wide.
4. Are some inbound links on mass from single domains. For example, a great number links from a footer link on a major site.
5. Have changes been made to keyword density recently.
6. Is there a distinction in ranking between major and minor terms.
7. Has there been any PR changes or backward link changes.
Google isnt likely to give us the direct answer, but maybe they wont need to if we work together in cracking this change.
If your site has been affected from wednesdays update, it would be helpful for all, if you could perhaps answer the 7 points above relating to your own site.
The first pages of the SERPs have been invaded by multiple listings for affiliate sites. Half my stable competitors' sites have dropped out of sight. (Travel category)
The last time the pages that I watch were like this, it turned into Florida.
Among other less than stellar results are:
1) An amazon book
2) A Yahoo directory page
3) A blog spamming/pay for links brochure site
4) An inner of a well established site (not an indented entry)
#4 deserves a closer examination. Yahoo shows 11 IBLs, and G show none.
Now, the homepage of that site is ranking at #6 purely on IBLs. It has no visible copy! It is entirely image based and is absoultely bloated with JS.
I think that I could just about replicate 90% of these serps by combining allinachor from the plural and singular keyword terms and overall IBLs. PR is not a major factor in what I am looking at - rather, shear numbers of links is what is doing it, and that counts on-site links.
It is the remaining 10% that just seem to drop into the SERPs from nowhere. ANd then there are the mysteriously vanishing sites that were ranking well and have disappeared for no discernable reason.
Overall, except for those odd ones, the SERPs are not bad, just different. Of course, if my sites had been the ones to vanish I might have a different opinion about this :)
I doubt that you and I will ever agree on this point, as you make it every time the subject comes up.
If the SERPs were for the query "sites about widgets" I wouldn't argue that it was relevant.
If the query is simply "widgets" then the most relevant would be pages about widgets, not lists of "sites about widgets", nor pages with "widgets for sale".
As long as G insists that directory pages are relevant, there will be many made for adsense directory pages showing up in the SERPs, a common problem if what we read on these boards is accurate.
Heck, I've got some of them myself. Perhaps you do to?
Finally, if you had read my entire post, and not quoted completely out of context, you would have seen that I stated that overall the SERPs were not bad, just different.
I see the most anchor text trash since pre-florida -- completely worthless sites based on thousands of anchor text links from blogs. (As well as many sub-domain redirects.)
These two factors do demonstrate a difference between now and the poor serps of last summer... getting tons of blog links works now, but making thousands of internal pages works much less well. In terms of which garbage is worse, I'd rather see internal crap links valued more than blog spamming. That is the dregs.
Of course it would be better if internal linking of quality domains would be valued, since building deep, content-rich domains has been the mantra of Google for some time now. The current "scraper algorithm" is diametrically opposed to that though.
The brave new world of seo... the hunt for PR2 links is on!
[edited by: steveb at 6:36 am (utc) on Aug. 27, 2004]
yep thats 2 of my sites, I've lost about 80% of the traffic, the first hit was 25th then another site on the 26th both are primarily affiliate based sites, right now I'm trying to work out what best to do.
I don't know if it's a direct attack on affiliate sites as I have others that are still ranking well, but one thing I did notice is some of the doorways on site A that were pointing to site B (rather than the affiliate) have still held there positions where as other doorways on site A that were linking straight to the affiliate have lost top places - you figure it out
I don't see any real changes in serps, I've lost a little ground but nothing to speak of. What did happen is a lot of my newer pages got dumped, became url only, or Supplemental. They were mostly from a subdomain that is all affiliate links as opposed to the root domain which is a company site. However I'm guessing the loss of indexed pages is just due to drop in PR. I went through my link 'partners' today and was astonished to see how many of them had ratted and especially how many had just disappeared.
Back to link building, urgh.
One page in front of me in the SERPs is a shoping site with a few hundred pages each one saying "cheap product" but with a different product on each page. Internal links only, PR 4 page. Content consists of mostly "buy you cheap product here", pictures and an affiliate link. Complete garbage (and my standards are pretty low).
I have a handful of posible theories on the changes that I will offer up for critical review here:
1. Google dropped the ball. Things will come good in a few weeks.
2. Links from the footer area devalued.
3. Links from sites deemed offtopic devalued.
1. The majority of those sites that lost a great deal of traffic have over 100 pages.
2. Not all terms have lost ground. In my case the major terms are steady or in some cases have improved. For minor terms (read destinations)it is the other way around.
3. Pagerank, backward links have not changed. It appears to be an algo filter which imo is not based on authority, keyword density, affiliation etc. In fact from what I can see the cause is not related to any of the 7 points I mentioned.
The fact that at least in my own case, the drop in rankings was not site wide but only affected minor terms "hotels in unpopulardestination" it gave me a good chance to compare pages that were ranking well "hotels in populardestination" with those that dropped significantly.
I thought it might be a kw density thing but they use the same templates so that couldnt have been the issue.
I looked at pr. PR was the same on the major and minor phrases. I looked at the internal linking structure and.. bingo!
I spent 3 hours last night doing extensive internal link structure research and found that practically all those internal pages that were receiving fewer than 7 internal backward links were the ones that lost ground in the serps. Those that had 30+ internal links had gained ground. I also checked this on some competitor sites that I'd tracked and knew had the same problem and found the same. Basically the internal link structure was simply not thorough enough. One destination which dropped 20 places when checked had only the one internal link pointing to it. Those that actually rose had +30. I tested quite a large sample and the common ground was the number of internal links. On competitors sites I also noticed that where there was no change or steady ranking the destination page had IBLs from external websites.
From the above I wasn't sure if it was a pagerank distribution issue or a simple number of internal link threshold kind of thing. One page on a competitors site had only 4 links from 4 different mid pr sites giving it a pr 5. This was sufficient for the page to retain its ranking. Therefore I would say that it is probably a linkpop/pr thing.
+20 internal links or 4-5 reasonably good PR IBLs seem to do the trick. With the pagerank update so out of date it is difficult to assess if it is a pr issue but it probably is. With better internal link structure you can gain a pr point for sub pages.
One of the sites I optimize was much more an authority that the other. This big authority site was not as badly affected traffic wise. The site that only recently got to being an authority dropped like a stone in traffic and had on average one pr less on the destination pages.
Conclusions so far:
The problem seems to be an internal linking/pr problem on sub pages. Better link structure without going ott I believe will pull the minor destinations back up. On a side note I wouldnt mind betting that those that have seen a significant rise in the serps have either fewer pages hence a better distribution of PR or have many pages that are very well interlinked with each sub page having at least 4 backward links (does depend on the competitiveness of the term.)
I am now testing with a page that dropped over 20 places by adding 20 more internal links plus one link from a different pr 5 site on a different ip c-block (thematic of course :-).
I should stress I may still be wrong here but the research so far is definitely pointing to pr distribution/internal link structure problem. Google seems to have tweaked the value of sub pages based on internal link structure/pr a bit. Many sites havent changed or have seen an increase. These sites probably have their pr spread well and perhaps have less pages to deal with than many large sites (in my case +60,000 pages).
I'll keep you informed how the test goes. I'm not doing sitewide changes yet as we all know Google throws a wobbly now and then and the results may go back to how they were within a few days.
PS. Well done all those who spotted the deliberate mistake in that Wednesday was the 25 and not the 26th :-) I cant seem to change it though (mod?)
It's never a bad idea to have a site map with inbound links to all your pages so that the spiders don't have to search for them, but I wouldn't go about creating link pools everywhere in your site based on this information, even if it is true right now.
i cant see any pattern, apart from the serps being worst i've seen.
typical that the 26th update was the worst for me but it seems to be sticking unlike all the others.
i'm from the UK and I am seeing a german language amazon clone site top of the results for my area.
nice to see Google still care about the quality of their serps.
It may well be the case that some terms dont trigger this change which may explain why many sites arent affected and can get away with one reasonable mid pr backward link on an internal page.
Or they could quickly find the exact category they wanted in the Google SERPs.
>I doubt that you and I will ever agree on this point, as you make it every time the subject comes up.
I do agree with you partially. I would think that in a case where for a specific SERP 6 of the 10 pages listed were directories, the a lot of people would find that unsatisfactory. However, if it is say just one of them is a link to a directory, a material number of people wouldn't find that objectionable. They still have 9 other choices. I've yet to encounter anyone in real life complaining about Y! or ODP directory cats showing up in the SERPs.
1. Dmoz listed sites of ours are suffering the most with a 35% drop, out of 21 sites only the dmoz listed ones are experiencing the 35%, call it a theory but when you have 21 sites all large and all established for years and the only ones to drop are the ones that all have a tie to dmoz for a backwards, this cant be coincidence. My 2 cents is that dmoz is finally being devalued from there algo.
2. Our largest and oldest "authority sites" with tens of thousands of backlinks are getting hit hardest.
We rely on hundreds of thousands of serps per day so we do not really track every one but I do have numerous ones I keep tabs on and we see a tremendous amount of scraper spam with a nice peppering of amazon/dmoz clone trash.
Another thing, this is definatly not effecting everyone, over half of our sites have only seen a 5% or less degrade which is not really anything in my book.
The main factor pushing me there was obviously footer links at 2 200+ page directories. Thus abundant links from a single domain must have been devalued dramatically. However as my site is now nowhere in top 200 this looks more like a penalty case - in my case I'm afraid this is duplicate content suspicion as my internal linkage forwards to the index page called other than index, default or home. What's even stranger is that I reappeared for all other 3 keyword combos (disappeaing from their SERPs for a day) but not this one. The scenario was the same - nowhere in the first 100 and then whooof - back to where I was before. Looks like google is selective where and how to filter for what he believes to be more popular searches.
Having spent much money and time to create pages with new content for what was an affitiate site .. I have found both these that have been worked on and the old ones i have are doing equally well / badly. Eg all have changed but afiliate backlinks are not the cause.
Are the drop in serps keyword specific or site wide.
Site wide .. 2 word key prazes back to Aug 2 serps 3 word ones as good as ever but still same as Aug 3.
Have changes been made to keyword density recently.
Is there a distinction in ranking between major and minor terms. See above 2 word and 3 word
Has there been any PR changes or backward link changes.
Backlinks .. all internal backlinks have been removed .. google showing PR0 backlinks all of a sudden .. lots of changes there.