McDar Experiment

Oh, and Bob, Caryl and z+2 (how did you come up with that name, by the way? ), I can help out with some linking as your experiments progress.

It seems like there is some upward movements in the regular serp for the McDar experimental site again, but if it stabilizes I could for example give you 730 links from one of my sites. It is a low pr site (PR 1-4), and was deep crawled really rarely until I started using it for experimentation

I just need to clean it up first, i.e. get rid of the other outbound links (too see the effect on the internal links and the external sites). But if this is interesting, let me know and I'll remove the links and set yours up.

 

Rational,

You make some good points. I agree that there appears to be a correlation between allinanchor: and the position in the SERPs, but I have never believed that that was all that was necessary to drive a page into the top ten.

And stemming is a very real factor. I have a site www.yourfriendlypharmacy.com which never uses the term "online pharmacies" in any anchor text anywhere. And BTW this is a very competitive search term.

McDar's tool reports that the site is #15 in the SERPs for "online pharmacies", but the allinanchor:, as one would expect, is not in the top 100.

For the term "online pharmacy" which is used in many many of the backlinks to this page McDar's tool reports the page as #29 but the allinanchor at #13.

In fact I just checked and our experimental page is #31 in the SERPs for "sleeping bag" (singular) and of course not in the top 100 for allinanchor:.

So there maybe a correlation, but certainly not a one on one or absolute relationship between SERP placement and allinanchor:.

I have seem many occasions when a site ranks higher for a stemmed version of the search term that for the actual anchor text.

I don't like long posts, so I'll talk about online poker rooms in a separate message.

 

Now to online poker rooms. I purposely chose that domain name and keyword phrase for exactly the reasons you point out. It combines two legitmate two word phrases "online poker" and "poker rooms" and should stem to both "online poker room" and "poker room".

So in my mind and afer doing some keyword research, I have one keyword phrase that may get my page ranked on 5 different contractions or variations. Damn clever don't you think

Now then the question is what should I use -- or we use, if you guys want to get involved -- for anchor text? Should I stay with the one complete phrase? Should I diversify the anchor text and use some or all of the phrases. Should I only use the plural phrases and let Google apply stemming to rank for the other two (singulars)?

Now McDar has a theory that the size of the site has a significant impact on it's SERP placement. How would you suggest that I add size to this site? I could add a forum. I could add poker tournament news feeds. I could link to all the books on Amazon about gambling. I'm sure there are other options I can't think of at the moment.

We didn't do any reciprocal linking to the sleeping bag page. Gambling sites love to exchange reciprocal links. Should I set up a resouces (link) section and aggressively pursue link exchanges?

I think it is probably impossible to set up two absolutely parallel pages to experiment with as Caryl has suggested. But we could add one or more features to this page that the original page did not have or use, and them track the results and progress against the time line of the original page. In otherwords after one week the sleeping bag page did X and after one week the online poker page has done Y etc.

Does this experiment make sense or appeal to anyone?

 

Well, this is going to be one long post, so I'll follow Bob's example and break it into several:

Hehe... It most definitely is.

 

This is very promising. I think we here have the chance of monitoring many different factors/matches/aspects of the Google algorithm. Interesting. Very interesting. I'll pledge you a run of the site with links from the site i mentioned in my previous post (after I have taken down the other links).

I think you should go for "online poker rooms", as you initially had planned -- and not vary the anchor text at all. That way we could monitor and measure the stemming effect isolately.

Phrases to monitor could therefore be:
- online poker rooms
- online poker (adjacent words)
- poker rooms (adjacent words)
- online rooms (non-adjecant words)

In addition I'd like to examine stemming:
- online poker room
- poker room

And relative words (in accordance with Google's ontology):
- online card rooms
- online casino rooms

According to Google "card" and "casino" are related to "poker", for example. It would be nice if some of these words could be found in the text and some not. One of my own sites rank high for a related word (it is low in density -- like all my targetted words -- and non-existent in allinanchor), but it would be interesting to see the effects in a more controlled experiment.

I would also be interested in monitoring a four words search term... a search term containing one word more than what the page is being optimized for. I don't know what this term should be though... something relevant, though.

There will be a lot of measuring/monitoring here... I'm not sure if it is necessary to watch all the datacenters, or if we could settle for one. There is also a question regarding which operators we should use for the monitoring all the search terms. Examples:
- regular
- allinanchor
- inanchor
- sandboxed

I'm a bit uncertain about the last one... after the big shift happened. I still believe in the sandboxing of links, since this makes to a lot of sense, but am very uncertain about what the -dsa search measures.

 

I don't think the size of a site has any direct impact on a site's placement in the serps (but I can't say this for certain, and an experiment is therefore always welcome). Indirectly, however, I think McDar is on to something. A large site with a massive amout of pages, means two things:
1) linkage (in itself)
2) internal anchor text

Both these are factors in the ranking, so when conducting the experiment to check if size matters, it will be important to rule out these (by only linking one-way through the hierarchy of pages).

 

Adding a forum is probably the best way to get unique content, but is also probably the most work. It would take some time to get a large number of pages. Taking a feed from Amazon seems to be the best way to add a massive amount of pages quickly. There is always the danger of Google spotting the content as not being unique, but this is probably a minor risk. ...if this is a concern, maybe you could use Barnes&Nobles since presumably fewer use their data..?

To make this an integral part of the experiment, maybe after having added some links, when the placement in the serps have stabilized, it is possible to add a huge number of pages (from for example Amazon) -- and watch the effect this has on the serps... By being careful in how the data is distributed throughout the site and how the pages are made accessible (re: links and anchor text), it is possible to do this in three phases:

1) Add a huge number of pages at once with only outbound links from the main page and throughout the hierarchy. Once the pages get spidered, it is possible to measure whether size matters.

2) Add a link back to the main page. Here we can measure how much effect there is in internal links without any anchor text -- once a deep crawl has been conducted.

3) Add anchor text to the links. Now we can also measure the effect of internal anchor text.

During this phase of the experiment, I assume all other factors are kept in stand by, i.e. no new inbound links.

 

Maybe it is possible to do the linking in phases? First focus on only inbound linking (while preparing a list of potential reciprocal link partners) and when this method has been exhausted, go for a large number of reciprocal link exchanges in as short time as possible (for then again, afterwards, to monitor the results over time until the serps stabilize)?

 

I think this experiment sounds very interesting! We have the chance of monitoring/testing a range of different elements:

• adjacent words in anchor text
• non-adjacent words in anchor text
• stemming of anchor text
• related words to anchor text
• size of the site
• inbound only links
• reciprocal links
• internal linkage
• internal anchor text
• etc.

After this experiment we should all - hopefully - understand the market we operate in better

 

I think Google has largely fixed the -das search as part of the last upgrade. As far as I'm concerned it hasn't worked since. It still gives marginally different results but I don't think it is voiding the sandbox effect.

 

Rational,

I'm pleased to see that you are interested and think we have something we can learn from. Hopefully Caryl and Foxy will check in tomorrow and make their comments.

I think it would be an interesting and worthwhile experiment, and I wouldn't mind the boost it would give the site either.

BTW do any of you Guys play poker. It is an interesting online activity.

 

Hi Rational and Bob

Golly you go off to have a glass of wine [australian semillon chardonnay] and the world has moved on.

Rational even though my user name is z+2 - I'm still called foxy. [Geez I'll have to alter my signature!]

With my name change [it is actually a personal thing] I put down my new site zplus2.com and it immediately proved, ie one day, the importance of allinurl going to no1 in that day for "zplus2" [if we really needed to prove that fact!]

I chose the name because of the difficulty of "using" [read getting ripped off by others] the site with the + in the name, and because there was no examples of it on the web as z+2 and only one example of zplus2 [its a transformer I think!!]. In this way we had a truely "virgin" site. The only links to it were using Search engine optimisation from two sites as I thought it could be fun to push it to the top [I don't do other peoples seo] and hand over enquiries to the main people here.

Now we have this site to use here as a control for what we do with Bobs site - we can do whatever we want with it - it does not have to stay as what I set it out to be but I do think we should use it as a control because of its virginal quality [unlike me hehe!].

Other than that I think that you have covered most of the factors and I look forward to carrying out the experiment.

 

I'm with you guys on this! Plus, I would like to see Foxy's site used as a control.

I'll have to comment in greater detail tomorrow as I am off to go fishing (after I post update)

Rational, it is really great to see your involvement here. You bring fresh eyes and some very interesting ideas!

Caryl

 

NEW PAGE UPDATE:
*Google PR/Backlink update

The "New Page" sleeping-bags.htm
04/07/2004 "New Page" went live

Search for "Sleeping Bags"
_______________
NOTE: Due to size constraints, the entire report can no longer appear in a post.
You can view the entire report here... Complete Report
________________
Last 10 days

Postion #23 allinanchor: #4 - 05/28/2004 [3 DCs #22, 8 DCs #24] Note: Allinanchor #5 on 4 DCs **29 DCs are NOT responding**
Postion #25 allinanchor: #4 - 05/29/2004 [All DCs not recorded] Note: Allinanchor #5 on 4 DCs **29 DCs are NOT responding**
Postion #25 allinanchor: #4 - 05/30/2004 [All DCs not recorded] Note: Allinanchor #5 on 4 DCs **29 DCs are NOT responding**
Postion #25 allinanchor: #4 - 05/31/2004 [All DCs not recorded] Note: Allinanchor #5 on 4 DCs **29 DCs are NOT responding**
Postion #29 allinanchor: #5 - 06/01/2004 [1 DC #27, 7 DCs #28] **29 DCs are NOT responding**
Postion #29 allinanchor: #5 - 06/02/2004 [ 9 DCs #28] **29 DCs are NOT responding**
Postion #28 allinanchor: #5 - 06/03/2004 [ 4 DCs #26, 4 DCs #27] **29 DCs are NOT responding**
Postion #27 allinanchor: #5 - 06/04/2004 [ 2 DCs #19, 6 DCs #26] **29 DCs are NOT responding**
Postion #25 allinanchor: #5 - 06/05/2004 [ 5 DCs #26] **29 DCs are NOT responding**
Postion #24 allinanchor: #5 - 06/06/2004 [ 10 DCs #25, 2 DCs #26] **29 DCs are NOT responding**

_____________________________________________________
note: these pages are found using ( site:www.compar.com +sleeping Bags )
number of Bob's PR6 and PR5 links found:
56 - 5/18/2004 (All datacenters)
Unchanged from previous date
57 - 6/2/2004 (All datacenters)
58 - 6/3/2004 (All datacenters)
Unchanged from previous date
__________________
note: this page is found using ( site:www.komar.org +sleeping Bags )
Alek's PR 7 Link:
Not reported by Google 4/17/2004
Reported by Google 4/18/2004 (all datacenters)
Unchanged from previous date
Link removed / no longer pointing to page 6/2/2004
__________________
note: this page is found using ( site:www.ski-france-ok.com +sleeping Bags )
Foxy's PR5 and two PR4s
0 - 4/20/2004
3 - 5/04/2004
Unchanged from previous date

__________________
Bob and Foxy added link Sleeping Bags to their sig files on this forum 5/06/2004
Results for - ( site:forums.digitalpoint.com +sleeping bags )

3620 - 5/21/2004
3650 - 5/22/2004
3640 - 5/24/2004
3500 - 5/25/2004
3470 - 5/26/2004
Unchanged from previous date
3530 - 6/1/2004
Unchanged from previous date
3590 - 6/3/2004
Unchanged from previous date
3740 - 6/5/2004
3750 - 6/6/2004
__________________

Caryl and Foxy added link Sleeping Bags to their sig files on the SEO Chat forum 5/11/2004
Results for - ( site:forums.seochat.com +sleeping bags )

298 - 5/21/2004
318 - 5/22/2004
435 - 5/24/2004
490 - 5/25/2004
467 - 5/26/2004
461 - 5/27/2004
Unchanged from previous date
429 - 6/1/2004
427 - 6/2/2004
435 - 6/3/2004
Unchanged from previous date
457 - 6/5/2004
452 - 6/6/2004
__________________

"Sandbox" info - this search theoretically removes "sandbox" link filter
Results for - ( sleeping bags -dsa -dsa -dsa -dsa -dsa -das -das -dsa )
#3 - 05/15/2004
Unchanged from previous date
#2 - 05/21/2004
Unchanged from previous date
#3 - on 17 DCs and #2 on 14 DCs 05/24/2004
#2 - on 16 DCs and #3 on 15 DCs 05/25/2004
Unchanged from previous date
#2 - 05/28/2004 (later in day #17)
#19 - 06/1/2004
#20 - 06/2/2004
#18 - 06/3/2004
Unchanged from previous date
#17 - 06/6/2004
__________________
PR/Backlink info

~ May 31 Update ~
sleeping-bags.htm has PR6
170 Backlinks
(backlinks reported in the 170)
37 of Caryl's links
46 of Bob's Links
1 of Foxy's Links
85 for Digitalpoint forums
1 for Alek's Link
___________
~ April 23 Update ~
sleeping-bags.htm has PR5
49 Backlinks
(backlinks reported in the 49)
5 of Caryl's links
42 of Bob's Links
1 of Foxy's Links
2 for Digitalpoint forums
Nothing for Alek's PR7 link(the link on Alek's page is 53rd link - may validate the 50 link/page theory)
__________________

 

And what are we fishing for?

 

haha, I go fishing for Large Mouth Bass

I camp every weekend from May to October and there is a great little fishing spot there.

BTW, I do have a Road Runner connection at the camp and I bring my laptop so I am never really out of touch. But, I spend as little time on the computer as possible.

Caryl

 

NEW PAGE UPDATE:
*Google PR/Backlink update

The "New Page" sleeping-bags.htm
04/07/2004 "New Page" went live

Search for "Sleeping Bags"
_______________
NOTE: Due to size constraints, the entire report can no longer appear in a post.
You can view the entire report here... Complete Report
________________
Last 10 days

Postion #25 allinanchor: #4 - 05/29/2004 [All DCs not recorded] Note: Allinanchor #5 on 4 DCs **29 DCs are NOT responding**
Postion #25 allinanchor: #4 - 05/30/2004 [All DCs not recorded] Note: Allinanchor #5 on 4 DCs **29 DCs are NOT responding**
Postion #25 allinanchor: #4 - 05/31/2004 [All DCs not recorded] Note: Allinanchor #5 on 4 DCs **29 DCs are NOT responding**
Postion #29 allinanchor: #5 - 06/01/2004 [1 DC #27, 7 DCs #28] **29 DCs are NOT responding**
Postion #29 allinanchor: #5 - 06/02/2004 [ 9 DCs #28] **29 DCs are NOT responding**
Postion #28 allinanchor: #5 - 06/03/2004 [ 4 DCs #26, 4 DCs #27] **29 DCs are NOT responding**
Postion #27 allinanchor: #5 - 06/04/2004 [ 2 DCs #19, 6 DCs #26] **29 DCs are NOT responding**
Postion #25 allinanchor: #5 - 06/05/2004 [ 5 DCs #26] **29 DCs are NOT responding**
Postion #24 allinanchor: #5 - 06/06/2004 [ 10 DCs #25, 2 DCs #26] **29 DCs are NOT responding**
Postion #27 allinanchor: #5 - 06/07/2004 [ 1 DC #26] **29 DCs are NOT responding**

_____________________________________________________
note: these pages are found using ( site:www.compar.com +sleeping Bags )
number of Bob's PR6 and PR5 links found:
56 - 5/18/2004 (All datacenters)
Unchanged from previous date
57 - 6/2/2004 (All datacenters)
58 - 6/3/2004 (All datacenters)
Unchanged from previous date
57 - 6/7/2004 (All datacenters)
__________________
note: this page is found using ( site:www.komar.org +sleeping Bags )
Alek's PR 7 Link:
Not reported by Google 4/17/2004
Reported by Google 4/18/2004 (all datacenters)
Unchanged from previous date
Link removed / no longer pointing to page 6/2/2004
__________________
note: this page is found using ( site:www.ski-france-ok.com +sleeping Bags )
Foxy's PR5 and two PR4s
0 - 4/20/2004
3 - 5/04/2004
Unchanged from previous date

__________________
Bob and Foxy added link Sleeping Bags to their sig files on this forum 5/06/2004
Results for - ( site:forums.digitalpoint.com +sleeping bags )

3620 - 5/21/2004
3650 - 5/22/2004
3640 - 5/24/2004
3500 - 5/25/2004
3470 - 5/26/2004
Unchanged from previous date
3530 - 6/1/2004
Unchanged from previous date
3590 - 6/3/2004
Unchanged from previous date
3740 - 6/5/2004
3750 - 6/6/2004
3790 - 6/7/2004
__________________

Caryl and Foxy added link Sleeping Bags to their sig files on the SEO Chat forum 5/11/2004
Results for - ( site:forums.seochat.com +sleeping bags )

298 - 5/21/2004
318 - 5/22/2004
435 - 5/24/2004
490 - 5/25/2004
467 - 5/26/2004
461 - 5/27/2004
Unchanged from previous date
429 - 6/1/2004
427 - 6/2/2004
435 - 6/3/2004
Unchanged from previous date
457 - 6/5/2004
452 - 6/6/2004
431 - 6/7/2004
__________________

"Sandbox" info - this search theoretically removes "sandbox" link filter
Results for - ( sleeping bags -dsa -dsa -dsa -dsa -dsa -das -das -dsa )
#3 - 05/15/2004
Unchanged from previous date
#2 - 05/21/2004
Unchanged from previous date
#3 - on 17 DCs and #2 on 14 DCs 05/24/2004
#2 - on 16 DCs and #3 on 15 DCs 05/25/2004
Unchanged from previous date
#2 - 05/28/2004 (later in day #17)
#19 - 06/1/2004
#20 - 06/2/2004
#18 - 06/3/2004
Unchanged from previous date
#17 - 06/6/2004
#19 - 06/7/2004
__________________
PR/Backlink info

~ May 31 Update ~
sleeping-bags.htm has PR6
170 Backlinks
(backlinks reported in the 170)
37 of Caryl's links
46 of Bob's Links
1 of Foxy's Links
85 for Digitalpoint forums
1 for Alek's Link
___________
~ April 23 Update ~
sleeping-bags.htm has PR5
49 Backlinks
(backlinks reported in the 49)
5 of Caryl's links
42 of Bob's Links
1 of Foxy's Links
2 for Digitalpoint forums
Nothing for Alek's PR7 link(the link on Alek's page is 53rd link - may validate the 50 link/page theory)
__________________

 

Right on - nice to hear

 

I think that all we would need to do is have Foxy set up a site of two or three pages and target the same keywords/phrases.

What would have to be controled for is the linking to Foxy's page would mirror those to Bob's page.

Rational has a good point about internal linking and we could control for that by having Bob's page have only one internal link to it.

Certainly, the on-page factors will be different, but even if Bob's site is in position #20 and Foxy's is in position #60, if there is a point that our linking strategy "trips" any "fliters" i.e. sanbox" it should effect both sites the same regardless of their individual ranking in the serps.

Caryl

 

I also caution about attempting to investigate too many aspects at the same time. As this may only serve to blur the cause and effect evidence we are trying to establish.

Having Foxy's site as a "control" would allow us to run "sub" experiments down the line especially if both sites show similar results to our linking changes/strategies. If the sites react similarly enough, we could get to a point where we could change links only to one of the sites and monitor changes that are reflected only to the site targeted by the links.

Again, the MOST important factor between the two sites would be maintaining the same exact links to both.

Caryl