Excellent thread indeed,
I shall bring a short testimony about the weight of titles ; my site (a "personal home page" site, I feel a bit alone in this webmaster category here) is structured hierarchically :
Main page -> A -> B -> C -> D
where level C contains small pictures, and level D enlarged pictures. Main page is PR >= 4, probably 4 (I have no toolbar to check, but I know it can appear as the result of a link: query). Since the structure is hiearchical, C pages are likely to be PR nearing 2 and D pages PR nearing 1.
On the previous version of my site, D pages only contained the picture, nearly no text, no title. I modified this : now they contain some short text (a date, a place name and upmost the title of the picture -under h1). This title is also the title of the page, and the content of an "alt" tag.
I check what it changed on the name of a friend who appears on four of my D pages (10000 "competitors" on this query). On the previous version of my site, one of the C pages containing this name appeared #18, now two of the D pages appear #3 and #4.
So for SERPs title is of the strongest importance, I find it stronger than could appear through what I read here on webmasterworld (but probably PR is less important in my "hobby" niche, where competition is not so strong as in e-commerce).
The sites that have linked to the actual PR 10 main site will nearly all mention one of the 2 keywords because of what the site is about. I didn't notice this until after I posted that the PR6 page is forwarded to the actual page and is not the page stated in the Google SERPS. These are the things that differ between the two sites. The PR4 has both keywords in the link text e.g. http://www.foo.bar/widget1/widget2.html whilst the PR6 has http://www.widget1.bar/../../something.widget2.php (forwarded from http://www.widget1.bar/widget2). The PR4 is the only one that links to the page internally (that I know of) using something widget1 widget2, whilst the other sites generally just have the link pointed at, but a significantly larger amount.
In addition to the main points mentioned in the previous messages, I have noticed 2 changes:
1. Google is undertaking some kind of I.P. based seperation on it's core Google.com site; seperating International users from US based users. I cannot fully confirm it as I am still investigating this. The lead factor pointing in this direction is that I lost significant traffic since yesterday from US based users, while users from partner sites and international I.P.s remained as it was.
2. The update is not as comprehensive as it usually is. Many of my deep pages are not properly updated...yes I know you might say it is because of my late update to them; but that is not the case, the pages were done well in time to meet the deep crawls.
I had around 67 pages chached. Today it became 217.
I was ranking 4th on my main keyword and after the update like every body I am on same positions for this and most of the other keywords.
Also last month I had more than 10 new sites linking back to me and around 4 with 6 points. The sites linking my sites are cached but no backlinks to my site?
So What I assume is that this time it was just a superficial crawl. May be a bit extension of Everflux. But it considered the back links of pages within site very much more.
Regarding google giving more weightage to Keyword density and position, I dont think google has given any special consideration.
One other thing I observed, is google seems to bring in news site more in to the search results.
Now as my pages chached changed from 67 to 217 today, I am assuming that some thing might happen soon.
The Real Dance is about to begin Yet!
PR decrease across industry? Maybe that's the reason my PR remained same. Was expecting a rise (tripled the incoming links). Also google is considering links from different pages of same website to another website. For me i still believe the link text as the most important factor.
mil2k, check the PR for the new links to see if they remained the same as when you got them. My guess is that many were demoted.
Its just a theory of course, but as I said, it makes sense to me. If Google doesn't raise the bar sometime, then theoretically, anyone can become a PR 7,8 or 9 without being a real authority. All you have to do is go on a massive link campaign ... regardless of themes and relevance which is what many, many webmasters have done and are still doing.
If they make it so tough that the average webmaster will stop trying (and therefore stop manipulating the algo) ... then link collecting should settle down to a reasonable (and relevant) level, making the algo much more useful. Think about it. If a site used to need X number of links (from PR 4 or higher) to become a PR 6 or 7 site, then making PR7 somewhat unattainable to the "average" site by raising the bar to XXX number of links ... who's going to bother? Only a "real" authoritative site or hub would be able to collect so many links.
If this is what they are doing, its brilliant and I think it will be very effective.
Ergo, webmasters will go back to collecting relevant links which will drive traffic for them and stop trying to manipulate Google's algo so they can sell PR.
Honestly I agree - there was NO major changes! Funny thing is, even a new site that was launched 3 weeks ago, we put some really STRONG (PR6 and PR&) links out there giving the new site an immediate 5 - cant be found anywhere for our keywords! The site is very SEO too! Just like another thread read - all this waiting/hassle for a little fart.. All of our other sites didnt move an inch - backwards link went from 600 to 1250 (again strong) and no change!
|The Real Dance is about to begin Yet! |
yet another user that feels this update was dodgy
jady... I'm finding just the opposite. On one of our primary PR6s, we dropped to a PR5 (fortunately, though, still maintaining our high search results); and a NEW site went from PR3 to PR5, and every article is still on page 1 of the SERPS for the primary keywords.
Somebody mentioned it earlier, but it almost seems like these are FreshBot results. We keep all our news items current virtually every day, and FreshBot has been keeping them at the top - but I was fully expecting the 'dance' to do a lot of repositioning, at least in the short term.
I had backlinks showing on google drop from 134 to 60, site has same PR but had by far my best update ever. Page 7 to 4 for most competative keyword and another, page 8 to page 2?
What can we gather from that if anything? I added a few good content pages and increased SEO stuff on home and keyword pages. Does this indicate links where less heavily weighted this month?
Yes - on closer analysis I think size of site affects the results more now. A bummer for me as most of my sites have huge sections which are unindexed (due to Flash etc). Will have to rethink.
Has is occurred to anyone that occasionally there may be no change to the algo?
>>Has is occurred to anyone that occasionally there may be no change to the algo?
It has occurred to me that using anecdotal evidence to base decisions regarding algo changes is fruitless. Unless you can isolate all the variables involved that contributed to changes in the SERPS you are left with conjecture.
The amount of data that needs to be gathered to create a data set that is large enough to make informed decisions is prohibitive. Using two or ten sites as a control to compare and contrast simply won't provide you with anything but an opinion. ;)
To me there doesn't seem to be many changes this update; SERPs for keywords staying pretty much the same.
Maybe there is more emphasis on inbound link text.
Our positions are virtually unchanged, as many have said, but I see many more total results per term or phrase across the board.
Our PR is unchanged, also. I was expecting/hoping for a point bump up as we'd landed a few choice high PR links. The link query is not showing them, however, and we actually went down again (3rd update in a row) on the links returned. I know that lesser PR links are not returned, but Fast has us with almost double the links returned even when only considering those of PR3 or higher... Guess I should not complain as the serps remain strong for us, but Google has never come close to returning all of our PR3+ links.
Algo tweaking has been done to the old index in end of March so in this update, only the index changed (no algo changed).
My observation -
- PR calculation algo is more strict and thus many people are getting lower PR. This happened the same in the last update. So this is the 2nd PR weight reduce.
- Importance of PR DOESN"T change, it remain strong and important factor. What change is the PR calculation algo.
- This is what I am wondering. I lost a lot of backlinks showing in the index but the links are still exists. As I forsee few months ago, Google will start ignoring links page in the PR voting calculation even if that page has PR. Or at least, I didn't see a lot of my PR5 back links appearing. All these links page has lots of links in it.
Netguy - our links do show, but they didnt pull as much weight as previous updates have. We launch many new sites each month, and this month we launched a new site that we put forth the most aggressive marketing we have ever done! However, even such the site was assigned a PR of 5 with 120 inbound links, it is still on page 6 for the top keyword we want to place well for. Now if you add a regional keyword to the search term, the site is number 1. But looking at the competition, there is no reason this site shouldnt be on page 2 or 3 at the worst.
I am thinking that Google is NOT giving new sites, no matter how high the PR is or how well it is optimized as much weight as it used to get. In all, what I am saying is, I THINK SITE AGE MATTERS ALOT!
On a high PR6 site:
Normally (two years now) we get 50,000 pages indexed out of 100,000. All of links are picked up, but the half of the site that is unindexed show up only as one-line anchors. Those don't get returned in searches unless you know how to do it, because you have to allow for the filenameing conventions we use and make them part of your keyword. So the half of the site that is unindexed remains invisible to all Google users except us.
This time our PR went up slightly, as we expected from a substantial increase in external links. The site itself is the same as it was during the last update.
And this time we took a major hit -- instead of 50,000 indexed pages, we have just 35,000. The other 65,000 pages are just showing up as useless one-line anchors. All those anchors prove is that Google knew it should crawl and index that page, but didn't get around to it. So we're down 30 percent for the number of pages indexed.
I was hoping that freshbot would be assigned to pick up some of those unindexed pages. After investigating what freshbot was doing the last couple of days (some 8,000 freshbot hits on one day), I have to say that yes, freshbot may get some of them, but no, it's not deliberate. It appears random -- freshbot has its own game plan.
Then we checked the old index at a point where it was still showing, to see if those useless one-line anchors were still present at the end of the previous cycle. The answer is yes, they were still there. In other words, we have no expectation that freshbot will make a serious dent in our 65,000 unindexed pages. More likely, it will be spinning its wheels on stuff already indexed.
By the way, the entire site has been spitting out 304 Not Modified when asked, for the last two update cycles. When GoogleGuy says to get your 304 act together, I'm wondering what they're smoking at the Googleplex.
This Google update was "crawl lite." By the way, Inktomi dropped thousands of our pages about a week ago, although their crawling is still very active. Don't know what's happening there either; hope it's just a phase they're going through.
Our goal is to someday stick up a disallow for all bots. If we could do that, it would reduce the stress considerably.
Since the recent update, something I've wondered about before seems more likely.
Many of us have suspected that new inbound links have more weight than older inbound links. This squares with my experience.
However, it also looks like it takes a couple of updates before new inbound links have their full effect.
When I get a high PR link to one of my lower level pages, I notice that the page moves up some in the SERPs after the first update. Then it moves up more in the next update. Then it tends to move down a little or stay the same in the following update.
Maybe just coincidence, but food for thought.
I have been thinking about this "crawl lite" and also what Google Guy said in another thread about "good site structure"...I am starting to believe the two are connected. It appears that it is necesary to give "deep links" from level 2 & level 3 pages in order to get the deep pages crawlled. This "structure" properly done will allow crawling of upto 1 million pages.
Let me restructure my site this way and see what happens next update
A bit further analysis of the changes to my SERPS. I have an old earthlink account that my site was/is set up on. Back before I knew what I was doing I had many links to my site, but some links were to the earthlink URL and others to my domain name which simply redirected to the earthlink account.
Now, I have never wanted to "waste" all those old links to the earthlink account, so as an experiment, a few months ago I moved my domain name to a new, dedicated account and threw a about 5 pages of content. The new site links to the old from every page, but the old only links to one interior page of the new.
The new site finally showed in the index last update. This update it shows a PR of 3. The old site has moved from a PR of 4 to a PR of 5.
In short, it appears that my effort to gain benefit from the old redirected domain name's links has been successful, which helps to explain why I have risen as much as 20 places in the SERPS on various search terms.
Still though, as I look at other sites that I have been involved with I am left with the impression that this update is giving more weight to on page factors than has been the case recently.
|It almost seems as if older sites are being favored a little more this update. |
My 1996 root level homepage lost over 100 backlinks with the update a month ago and this time it seems to have gained them all back plus about 40 more. I am really how the algo changed both of these updates. Did they change something a month ago to lessen the advantage that older sites may have had and somehow it didn't work out? Any theories? Neither update seems to have affected the serps on my top key words. My new (couple of years old) site did not show these dramatic changes.
|Has is occurred to anyone that occasionally there may be no change to the algo? |
Then why the massive backlink changes on my old site?
I see a lot of mention of the influence of authoritive sites. Exactly what is an 'authoritive site'? Yahoo and DMOZ? Are there others? Does the fact the site is .edu or.org make any difference?
I have been working hard to get my most active site (one that is about 3 years old) up on a prize one word key word. I got a huge jump on the serps a couple of months ago. I think the big factor was shortening my title in the title tags to just 3 words including the keyword. I did get a lot of new incoming links then as well.
Right after the last update I went through the rest of the site and made sure that the links to the homepage had the exact 3 word title in the anchor text. That was all deep crawled. Now I've gone from 19 to 11 in the serps on this keyword, a good % jump. I'm wondering if the new anchor text made the difference. It's hard to tell because I got a number of new related backlinks as well. Of course I want to get into that top 10!
French Tourist you aren't alone; there are several of us with non commercial or minimally commercial sites here. Content only sites are important too and even though I don't make my living with my sites I put in many hours of research and writing to put up information that can be found no where else on the net. I want it to be seen and doing well with Google is the key.
I saw a PR drop of up to 1 point across all sites, even though inbounds were UP or about the same.
Here's a wierdie.. I have link to my main site from a PR7 page.. it's the only outbound like from that page, and yet MY page comes out as a PR5.
However, SERPs are better than the last update. Outbound links from the site are still strong.
It does look a little like there's some tweaking going on here.
Sorry, juggling threads, misposted.
For the results I'm watching, nothing has changed in the algo (pr has changed, but I'm not sure thats algo) on either google, www3, or most of the datacenters, save -sj, where there seems to ba an excellent treatment (ie reduction) of spam sites. Personally I'd love to see -sj serps make it to www.google!
Anyone have any theories on what is going on with the datacenters this time? Seems weird to me, but maybe things just haven't settled down yet.
Looks to me like 2 datacenters haven't updated yet. I notice my main site seems to be fluctuating between PR5 and PR6, and as it was PR6 last month I have a bad feeling about this. :( My guess was I was low PR6 last month, and this would be consistent with an absolute PR drop across sites overall.
I made changes to titles, <h> headings and anchor text for the internal links and added 50% more back links. I saw dramatic improvement from #17 to #7, #7 to #1, and from #2 to #1 where the previous #1 held the #1 position for over 18 months.
After suffering horrible results in January, I reached the conclusion that Google was shaking things up. To further test this theory, I took an internal page with PR1 and no content and put in <H1> and 1 line sentence, 'Coming soon - new "keyword" section". The link from the main page was changed to 'keyword' 'keyword". The page is now PR4.
Did the same test with a new one page site. Only backlink was from my primary site link page (PR4) that has 90 outgoing links.
This new site is now #4 and it bumped several established sites with far more content.
My conclusion is that anchor text and on page optimization are becoming far more important, hopefully diluting the reliance on backlinks and once again increasing the value of hubs and authorities.
My website is ranked #4 for my main category keyword and this did not change. However this is first time we have not moved up over the past six months.
Here is waht I find curious. Over the past five months, on the day of the refresh, when I searched link:www.mysite.com my Google links went up or down. This is the first refresh that there is no change in the number of links.
Has this been the experience of others?
Is the dance over? My site didn't move up or fall down. In fact, none of the results for my primary key world change a bit. Hmmm...
This is just an opinion of mines, perhaps someone will be able to confirm or deney this.
Is it the case that one way linking is now a lot greater than recipricol linking?
With all the talk of backlinks being lost does this have anything to do with recips being dropped?
Perhaps google is getting wise to recips being for PR gain and one way links being a vote. I have not been able to find anything to confirm this it just sounds possible.
makes sense to me, a reciprical link shouldn't carry much wheight at all, but a one way should.
edit for typo
[edited by: NotePad at 2:25 am (utc) on April 13, 2003]
| This 113 message thread spans 4 pages: < < 113 ( 1  3 4 ) > > |