Good observation hitprof and I think some comments to that "doubling-effect" have been burried in an earlier post(s) in one of the huge sj-dominic threads.
However many are also observing backlinks not showing that were placed for example in early april (before the deepcrawl). As googleguy commented - these might still show up.
Maybe in some cases but not all.
The SJ server is showing a total number of links that is not even equal to the whole nubmer of backlinks from my own internal pages, let alone the links from other sites.
To be drawing any conclusions from the SJ data in my opinion is premature at this point.
The dance has not happened yet and until it does, we really won't know what data is going to show and how it is interperted.
HitProf, your hypothesis is not congruent with the results I'm seeing.
My site shows (actually displays) over 70 IBLs on WWW right now (out of an estimated 146). -SJ shows 32 out of 39.
-SJ is definately using an older data set for my site. I think the differences between -SJ & WWW are more dramatic for newer sites (or sites that have had significant link building within the last two months or so).
I agree that we have to wait for the real update to say anything definite about sj-results. Lots of webmasters are still missing backlinks they now exist and should show in the index. This is another issue: those missing links are most likely not yet counted and are not yet included in the reported number. That might explain the fall from 381 to 276 on one of the sites I checked. This may be fixed when the missing links and last crawler data will be include in the index.
What I say is that the *estimated* numbers are now far more accurate and this accounts for the *huge* loss of links, compared to the relative small numbers of missing links in the new index.
Some of your links are still missing. How old are your missing links?
The -sj index seems to be a fairly old index with some fresh inclusions so your missing links may show up as soon as they have fixed that. I hope your backlinks will go up to 70 soon, but I still think 32 out of 39 is a better guess then 70 out of 146. You never had those 146 links indexed anyway.
If my theory is correct, then by that time Google will show something like 70 links out of 'about 70'.
HitProf, I really like your insights. I think our newer systems do a much better job of estimating link counts.
The reduction in inbound links showing isn't the concern of most webmasters. Rather, the complete lack of anyu recent deepcrawl data.
What we have noticed is a lot more internal links, up by about 40.
I have seen that a lot of our other links, especially those from countries other than the US (yahoo japan, korea, etc) are missing.
I always assumed that the link estimates included links from sites that were lower than PR3, and that accounted for the discrepancy.
For instance, I have a site that has been included (without my permission) in a good number of these so-called "directories" because it is a top 3, 5, 10, and 20 result for several, search terms related to "our" industry.
Anyhow, since I chose to include my name in my meta description tag, if I search towards the bottom of the "heap" in the SERPs for my name, I see a lot of very (feeble) attempts to create "hubs" using links to my site and others with good rankings.
I wish I could set a link to one of these - they have unintentionally created a piece of art - a keyword collage, if you will. Cool effect, varying the size of the header tags... Well, apparently the site took their subdomain page down, but the cache still displays it in all of its "glory".
Anyhow, google sees this link to my site, but it isn't listed in my backlinks. Therefore it shows fewer backlinks than I have, and fewer than Googlebot itself has discovered.
Thanks for the confirmation. I'm glad it wasn't just me :)
But it's still not working properly.
Yahoo's backlinks have gone up to 617K. (I'm not going to count those!) so I've rechecked the others. They are now 21 out of 25, 278 out of 294 and 71 out of 75. That still seems OK.
But I've also found another one that shows only 18 out of 31 reported backlinks on www2 (40 out of 80 on www). So this site is still missing links and has a big gap between reported and showing links.
The numbers of low PR incominging links do not relate to gap between reported and show links in the old index. That gap always seems to be 'about 50%'.
If you do a spam report on the 18/31 site, I'll ask someone to check it out. Thanks..
|I think the differences between -SJ & WWW are more dramatic for newer sites (or sites that have had significant link building within the last two months or so). |
Welcome to Webmaster World Bernard!
I've come to the same conclusion myself.
The problem with the back links on fi, sj, www2 and www3 is that they are old as far as my site is concerned. Probably 3 to 6 months old. I stopped advertising with a few sites but the backlinks are still there, even though they aren't on www and none of the newer sites that have linked to me have appeared.
The overall effect is that I've lost ground on all keywords.
I hope the dance continues and there is a full update of April backlinks.
Ditto to only old links showing. The newer ones aren't showing up, but they're still linking to me and are PR4 and up.
For me no new backlinks after march 1st are showing...
Come on deep crawl data from April! Where are ya hiding? An issue with 99.9% opinion and GG with the facts. I bookmarked GG's profile page so I know when a fact has been posted. thanks GG
very interesting ... I have a site, which I did a lot last month and for "some" reason it is #1 on very competitieve keyword (was #15).
And for this site I expect to have more links shown .... in fact it shows ONLY all old links, just exact number ;) So I suppose command link for now doesn't show any new links.
No doubt this will be solved after the next update. If not, I'll let you know. Thanks for your concern.
<added>Not unimportant: this whole thingy has no infuence on ranking whatsoever. 5 out of 3,9M on -sj, 5 out of 4.8 mln on www and 5 out of 4,7M on -fi :)</added>
Grin! - wakes up next morning
Orphan opages - redirects - Doesn't that sound like old fashioned spammy doorways?
We'll see what happens after Google brings in the spam filters. Let's hope refreshing the content with deep crawled data comes first.
Boy, threads get buried quick around here.
Hitprof, my site is old, but the missing backlinks are (relatively) new. I checked yesterday and the backlink counts on -sj & -fi have increased to 41. I'm confident that Google will eventually find all the links out there. I'm not worried about it.
I was just pointing out that the index on -sj & -fi is old and the differences in actual backlink "droppings" (I'm not referring to the estimated figure which you rightly point out as being more accurate now) are more dramatic for sites that added a lot of links over the last couple of months.
I agree with you that the 'new' index is old :)
One of my client's sites is sort of a monthly e-zine and the index is missing some complete issues. The most recent one is included however, must be freshbotted. It's still strange...
We'll wait and see.
Are you able to divulge when recent deepcrawl data will be factored into the new indexes?
Appreciate this does take time, but was wondering if we are days or weeks away from this happening?
Now that the new index is beginning to emerge on -fi I did some quick checking. (Congrats UK_Web_Guy with your find!)
The new backlink numbers for these sites are now almost exactly spot on (hmm.. 387 out of 396 isn't bad).
The 'problem' site 18/31 has turned into 18/18 but rankings are still unaffected.