McDar Experiment

Number of backlinks. I just did an accurate count of the pages in my InfoPool and it comes out to 48 pages. Three of these are new articles and while they have been cached by Google have not been assigned an PR yet. So they won't show up in a backlink search yet.

That means there are currently 45 pages with PR5 or PR6. So it will be interesting how long it takes Google to report them all.

 

Bob,

My gut says it was coincidental. I do find it hard to beleive, myself, that a single PR7 is worth that much more than 38 to 40 PR5 and 6's.

BUT, lets just say for the moment that it is. My next question would have to be HOW MUCH MORE?

Google had time to discover and chew on 38 to 40 GOOD links. YET, there really was not that much movement in the serps.

Then along comes one PR7 and the page jumps 60 positions. I still maintain that IF the PR7 link had that much effect within hours of it's discovery, then we can only assume that every other link thus far has been weighed in and accounted for as well. If this is the case, we should see virtually no more movement until Foxy's links come into play. (of course there will be small movement back and forth 4-5 positions -but no more real gains).

 

That is another commonly held claim -- that Google gives extra value to links from pages with relevant themes or content.

I think Google would like to do this, and I think that is what they tried in the Florida update, but I think they have abandoned it for now and that the only thing they are using for relevance is anchor text. That in my opinion is another test within this test. Except for McDar's internal links all other links are from pages without any possible relevance to sleeping bags.

 

Shawn,
I would have to say that I agree with you regarding PR playing a big part, HOWEVER,
you are looking at the advantage your tools have of being located within a Large site and that site has a high PR which adds weight to it's pages.

In the case of the experimental page, it is located on a small site (256 pages) and the site is a PR6. The PR7 is a single link from another website.

I have seen time and again that site size does add conciderable advantage in the serps.

I guess a way to prove this one way or another would be to replicate this by pointing one single PR7 anchor text specific link to another struggling page and see if the magic bullet works there.

It would be a quick experiment as it only took Google less than 2 days to find Alek link.

 

EXACTLY, Bob! This is a VERY important piece of information.

I agree with you 100% regarding these "theories" being tossed about regarding only getting links from sites within your category/theme and "too much of the same anchor text will hurt you".

We may have an opportunity to prove or disprove them here.

 

Hi Mick and WELCOME!

I hope you don't think I was ignoring you.

I just took a look at your site with the keyword tool using web site design uk as the keywords (not certain what your target is) and you look to be heading in the right direction.

There are some big players at the top but also a few smaller sites that you certainly could match numbers with in a relative small amount of time.

There definately is a similarity between this experiment page in that this is also a relatively small site fighting to get in amongst the "big guys".

I kind of think of it as a featherweight boxer gearing up for a bout with a heavyweight boxer. You have to overcome the size difference by becoming much stronger and much smarter.

One way, I think we are proving here is to MAXIMIZE your anchortext.

I had been building links for my main page, prior to this experiment, and was using sleeping bags and tents - Mcdar for the anchor text for sites to use as the link. (now, I now there are other factors at play, but) That page was at position 365 in the serps for sleeping bags when I started. I had managed to push it up to position 202 (as of today). Now that is a lot of progress BUT, if you do a search for sleeping bags and tents it has moved from nowhere to position #3!

The reason I chose to start a whole new page instead of persisting with this one is that this page had the backlinks to it diluted with many links from directories which just mentioned the mcdar, etc. It seemed it would take many more links with very specific anchortext to turn that around than to just start fresh with every link in the anchor.

Thus, today...
new page position#29 allinanchor score of #14 (brand new internal page)
old page position#202 allinanchor score of #26 (site's main index page)

Good luck and certainly feel free to join in the discussion here.

Caryl

 

FYI Guys: Please remember that what date/time you add your link(s) to your pages is irrelevent - what REALLY matters is when GoogleBot comes and finds it. For instance, I added the www.komar.org link at 0645 MDT on April 16th ... and Googlebot didn't spider it until 2225 MDT that evening as part of his daily visits. I'd suggest that it's an important data point to keep track of as I'm not sure that everyone is spider'ed daily.

Since Mcdar says my link was "Reported by Google 4/18/2004 (all datacenters)", we can set an upper bound for the propogation from GoogleBot to the Data Centers. I've mentioned before that if we wanted to micro-analyze, it would be interesting to know what TIME (MDT is GMT-6) on 4/18 that you found my link active?

alek

P.S. I got a chuckle out of Shawn's comments on the links from www.digitalpoint.com - I'll bet that page is a PR7.99 ... and would be real surprised if it doesn't go PR8 in the next update, which means it may allready be so internally.

 

Alek,
I am not certain of the time but I generally check around 7:30 am EST. This is not to say that it was not showing up on Google had I looked at 10m the evening before.
I think the best we can hope for here is correlational data. When you focus too closely one one thing, you tend to lose the focus of the overall picture.
As I have said before, to prove the value of your link, you could do one of two things. You could link out to a similar page in a similar manner and see if it has as dramatic effect. Or, you could simply remove your link to the experiment page and we will watch to see what effect it has.

 

I have a couple of questions regarding this experiment. First, to the person with the PR 7 site. What effects has it had linking it to other sites of your own? Does it really carry that much weight? And second, Bob's links are mainly from the same domain he stated. Do links from the same domain carry that much weight? The reason I ask is because I have a few sites, and I typically don't link the internal pages out to the other index pages on different sites that I own. I just recently started doing it a little more the other day. I'm scared to do it too much in case them being on the same ip will hurt anything. However, my girlfriend's sites are on a different deal with the same host. I could just link over a hundred pages to her easily this way. So again, do links on the same domain carry weight still?

Oh, and I've emailed a few people with PR 7 sites for a link exchange, if they end up giving me a link, if I see results quick enough, I'll try to post them if I'm still reading this thread. Might not be as controlled of an experiment, but maybe it'll still help.

 

Thanks Caryl

I think you're right in that it demonstrates an effective method for the little guys to compete. I set about my experiment to see how long it would take (and what it would take) to reach first page for a fairly competitive term with a new site.

Interestingly I've just rechecked my figures and have jumped 20 places immediately after G picked up a link from a new internal page. I'm wary of jumping to conclusions but the results had been reasonably constant until I added the page to the site a couple of days ago. Maybe enough encouragement for bpal to try more internal links?

I had just added another page today before the change in the allinanchor results so it will be interesting if I get the same kick in two days time after G picks up the page tomorrow. I will let you know.

Mick

 

Thanks for that insight. Hopefully that will work for you. If things look good, I've got around 200 pages or so that I can kick links towards my girlfriend's sites and see what happens. They are only around PR 2 and 3, but I don't think they really have any outbound links, only inbound back up the tree on the same site. I'd link them to my own, and I have a little, but I don't want to get hit by the search engines for cross linking too much. So my girlfriend's sites will be what I try it on even if they are off topic.

 

If you are worried about links from the same IP block then you may not want to link to your girlfriend's site. You said earlier that it is on the same server as your page. In this case it is probably on the same IP block. Remeber there are 256 IPs in a full block. The first 3 sets of numbers are the same for every IP in the block.

However I don't think Google is checking this. It would mean they were using either Local Rank or Hilltop technology and there is no good eveidence that they are.

Remember though, what is different about the McDar experiment is that while all 48 of my links come from the same web site, and therefore the same IP, they are on a completely different IP block than the target site.

 

Alek,

I don't really see the relevance or importance to this timing discussion to the experiment. The results will only be meaningful over time. Regardless of when you posted and when Google crawled your site we can't be sure that the big jump is directly related.

The data over the next few days will help us reach our conclusions, but the timing of the bot either on GMT of local time is probably not really germane to this experiment.

 

That's a good point about the ip blocks. Thanks for sheading some light on the subject (I don't know much about them). I'll need to look more into that.

 

What I was trying to say Bob is that I changed my link at 0645 on 4/16 and Googlebot crawled my main index.html 16 hours later. If Googlebot had NOT crawled my page yet, then there would have been ZERO input from my link so far, and there is no way that my PR7 link could have caused the big jump in ranking seen shortly after my link was not only crawled, but did show up in Google.

Just to be really clear, I am not saying it is conclusive that the PR7 DID cause the jump in rankings, but again, if Googlebot had NOT crawled me yet, I could say pretty much for sure that was NOT the case.

I don't know how frequently other web sites are crawled, but the "clock starts" when your page is crawled, not when you make the change - this may be significantly longer than a daily crawl for other people (can you and/or others share any data?) and I think has possible bearing on trying to determine cause/effect.

I agree over time that it all gets "absorbed" and shows up eventually, but I think you guys are trying to match certain changes in file-contents/back-links/etc. to changes in Google various rankings.


alek

 

Alek,

I think I know where the misunderstanding is here. The only reason I want to know when a link goes up is so I will know WHEN to start including a search for evidence of that link in Google.

This is a section of my daily update...

Every morning I perform this search on the datacenter tool
site:www.compar.com +sleeping bags

This search gives me ONLY the pages found in Googles index that are located on Bob's site that INCLUDE the words sleeping bags on the page. Since there is NOWHERE else on his site that sleeping bags is mentioned, the results are only the pages with the links.

You will note that on 4/17/2004 I found that Google had listed 37 pages of Bob's that had sleeping bags on the page.

It does not matter when these links where actually spidered all we are interested in is whether Google has them listed that day. If you look closely, you will see that numbers have been fluctuating from day to day. This is most likely due to the fact that some are still only showing up due to freshbot visits and really not indexed yet.

One advantage a PR7 page may have over a lower PR page, is that it probably receives freshbot visits more frequently.

Does this make sense to you?

Caryl

 

I guess I'm confused - I've been looking at this data you have been posting:

Isn't the above information gathered by doing a "regular" search for "sleeping bags" (and then one using allinanchor) and seeing where the www.mcdar.com sleeping bag page ranks. I.e. aren't we all linking to the www.mcdar.com sleeping bag page and what we are interested in is seeing what actions cause a raise in the rankings.

The "site:" search you mention above just shows which sites/pages have been spidered and "recognized" by Google. www.compar.com happens to have a bunch of 'em and it appears it took a little while for 'em to all show up in Google (assuming all added/spidered at the same time - you might check the later). www.komar.org only has one link and we saw mine showed up in two days - that's fairly easy to see.

What is the $64,000 question is: if you add link XYZ, how and when does it "ripple" through the Google ranking system and affect the rankings of the target page, with the allinchor position being a possible secondary indicator of relative importance? The data I quoted above shows something "significant" happened between 4/17 and 4/18 - can we isolate what it was?!? ;-)

alek

P.S. I agree with you that one definition of when the "clock starts" is when it shows up in Google using your "site: search" ... but I suspect that's probably not very long ONCE your site has been crawled - this may be a fixed time (my guess) or it may be a function of the importance of your site - I've never tested that, since once the page is spidered, Google seems to "have" it withen a day or two - if you still have the historical data for each of the www.compar.com pages, it might be illuminating to cross-reference against when the googlebot spider'ed 'em - my guess is they all showed up with 2-3 days AFTER the spider (not neccessary after you added 'em). Another easy way to test this is next time you add/update a page, do a Google Search a few days later for a complete sentance or two that you added (shouldn't have to say site if unique enough) and I bet it will show up.

I think I'm confusing/clouding the issue, so maybe I'll sit back for a bit and see how things develop - again, what is very interesting to me if the effect of a link on the SERP's (and allinanchor) with a secondary interest of how fast does this occur.

 

Your link from the PR5 is up and one from a PR4 on the same site but in a Number 1 position and one PR4 in a number 2 position same keyword phrase - "snowboard France"

Would you like more?

 

Thank you FOXY!

Please let me know the urls of the links so I can add them to my daily search.

Caryl