Hoping to pick the brains of some old pros!

niccol:

We will let the readers of the thread decide if :

Option 1
I can't understand your point because it is far above my level of comprehension due to your superior intellect, knowledge and experience.

Option 2
I can understand and think it is wrong.

You didn't leave much in the way of options here did you? I've decided Option#1 is the most probable, even though you do give me far more credit than I deserve and yourself less credit than you deserve.

The problem isn't you, the student. The problem is me, the teacher. I have failed in my task :(

This isn't fair to you, so I feel duty bound to try again with a slightly different approach.

I have a theory. 4 cylinder cars are faster than 8 cylinder cars (hey, I never said it was a good theory).

How can I prove my theory?

1. I can read the manuals/specs of a range of 4 cylinder cars and compare them with the specs of a range of 8 cylinder cars, and I can ignore all the other specs and just take note of the top speeds specified for each, and I can come to a conclusion based on what has been written.

This isn't 'proof'... it is just an acknowledgement that I accept what the manuals have stated.
OR,
2) I can devise a test so I can see the results for myself.

I'll do the latter. A race!

For the race to show any useful results, I'm going to need need X number of 4 cylinder cars, and an equal number of 8 cylinder cars.
It just happens that I've got a got a friend with a Ford Dealership, He supplies me with 10x 4cylinder Fiesta's (identical in every way), and 10x8 cylinder Falcons (also identical in every way).

At the end of the race, I'll take the 1st 10 cars across the finish line and see how many of them have 8cylinders and how many have 4.

If my theory is correct (4 cylinders are faster than 8) I can reasonably expect most, if not all, of the cars to cross the finish line will be the 4cylinder Fiesta's. If the theory is incorrect I can reasonably expect most, if not all, of the cars to cross the finish line will be the 8cylinder Falcons.

Regardless of the outcome, I will be quite confident to state that my theory will been either proven or dis-proven. Agreed?

Wrong!. All this is doing is comparing one type of Ford Vehicle with another Ford Vehicle. It says nothing about the GM Holden’s or any other make of car. This is the problem we can expect if we attempt to 'make everything equal except the thing being tested'

The results (whatever they may be) will be hotly debated among the Ford and Holden supporters.

In order to prove/disprove my theory, I need to introduce other variables, specifically, a range of 4 cylinder cars from a range of makers, and a range of 8 cylinder cars from the same range of makers. For simplicity, lets say the new race consists of
5x 4 cylinder Holdens
5x 4 cylinder Fords
5x 8 cylinder Holdens
5x 8 cylinder Fords

The theory/test is still to see if the 4 cylinder cars are faster than the 8 cylinder cars.
At this point, some people will scream as loud as they can that this test/race will prove nothing because of the introduced/other variables, specifically the different makes of car, but the cars may also be different colours, some will have radios, others will have a CD player, etc,etc, (these are all irrelevant as far as the 4/8 theory is concerned).

The 'argument' being that all these other variables means we can't come to any valid conclusions in regards to the 4/8 cylinder question.
They may or may not have a point. We have no way of knowing unless we perform the test. (Strangely, those currently putting forth this argument aren't even interested in doing any testing because of these variables). I digress.

What are the some of the possible outcomes of this race?

a) The cars will cross the finish line in a somewhat random order.
b) The 4 cylinder Holdens will take the lead positions followed by The 4 cylinder Fords, etc
c) The 4 cylinder Fords will take the lead positions followed by The 4 cylinder Holdens, etc
d) The 8 cylinder Holdens will take the lead positions followed by The 8 cylinder Fords, etc
e) The 8 cylinder Fords will take the lead positions followed by The 8 cylinder Holdens, etc
f) The 8 cylinder Holdens will take the lead positions followed by The 4 cylinder Holdens, etc
g) The 8 cylinder Fords will take the lead positions followed by The 4 cylinder Fords, etc
h) The 4 cylinder cars take the lead positions (in no particular order) followed by the 8 cylinder cars (no particular order)
i) The 8 cylinder cars take the lead positions (in no particular order) followed by the 4 cylinder cars (no particular order)
..... and so on.

Can we come to any valid conclusions from any of these results? Some will say no. I say Yes.

a) The cars will cross the finish line in a somewhat random order.
Conclusion: Neither the make of car, or the number of cylinders have any bearing in which is faster. (an unlikely result in this scenario. Theory neither proven or unproven)

b) The 4 cylinder Holdens will take the lead positions followed by The 4 cylinder Fords, etc
Conclusion: The 4 cylinder cars are faster than 8 cylinder cars, and the 4 cylinder Holdens are faster than the 4 cylinder Fords

c) The 4 cylinder Fords will take the lead positions followed by The 4 cylinder Holdens, etc
*Conclusion: The 4 cylinder cars are faster than 8 cylinder cars, and the 4 cylinder Fords are faster than 4 cylinder Holdens
*
d) The 8 cylinder Holdens will take the lead positions followed by The 8 cylinder Fords, etc
Conclusion: The 8 cylinder cars are faster than 4 cylinder cars, and The 8 cylinder Holdens are faster than the 8 cylinder Fords

e) The 8 cylinder Fords will take the lead positions followed by The 8 cylinder Holdens, etc
Conclusion: The 8 cylinder cars are faster than 4 cylinder cars, and The 8 cylinder Fords are faster than the 8 cylinder Holdens

f) The 8 cylinder Holdens will take the lead positions followed by The 4 cylinder Holdens, etc
*Conclusion: Holdens are faster than Fords, regardless of engine size *

g) The 8 cylinder Fords will take the lead positions followed by The 4 cylinder Fords, etc
*Conclusion: Fords are faster than Holdens, regardless of engine size
*
h) The 4 cylinder cars take the lead positions (in no particular order) followed by the 8 cylinder cars (no particular order)
*Conclusion: The 4 cylinder cars are faster than 8 cylinder cars, regardless of make. (This unlikely outcome will prove my theory that 4 cylinders are faster.
*
i) The 8 cylinder cars take the lead positions (in no particular order) followed by the 4 cylinder cars (no particular order)
Conclusion: The 8 cylinder cars are faster than 4 cylinder cars, regardless of make. (This is the most likely and expected outcome, which will disprove my theory that 4 cylinders are faster)

**In all cases (except 'a') the results clearly show, first and foremost, whether 4 cylinder cars are faster than 8 cylinder cars (or not), with a possible *secondary observation *as to whether Fords are Faster than Holdens (or not). **

In other words, this one test/race, with added variables, tells us more about whether the 4/8 cylinder theory is correct (or not) than if there were no other variables at all.

If the theory was to test if Holdens' are faster than Fords, we can use the exact same results to prove/disprove this theory to, with a ** possible *secondary observation ***as to whether 4 cylinder cars are faster than 8 cylinder cars (or not).

Summary: We do NOT need to 'consider every possible variable' in order for any given test to be considered 'valid'.
The conclusions derived from any given test is dependent in what the test is designed to prove/disprove and 'other' variables will be of only secondary importance.

End of lesson.

If I fail you this time, I will need to refer you to another teacher... I think 'bn' may be a good candidate.

Rod

DivaVocals:

Uh huh.. Okay..
Why?? are you gonna get the bouncers and force me to leave???

As usual, you missed the point.

Some/many of us here are trying to advance our knowledge and skills, by asking questions and performing various tests (which may or may not demonstrate anything conclusive).

It is obvious that you feel you have nothing to learn from being here, and by all means, hang around for the entertainment value if that's all you are interested in.

The bulk of your (and Nicks) input has been little more than unfounded negative criticism designed to do nothing more than confuse.

As I've been told many times, If you can't say something positive, don't say it at all.

Rod

SPH:

This little one-product niche market is, I think, a very good way to test and refine SEO skills and theories. Either independently or participating with others who come along.

Spot on Steve! This is the very reason I saw the need to change the original "Wurlitzer Debobulator" product to "Wurldbitzer Debibulator".

Although both 'products' are unique in their own right, the individual words are not. If we'd used the original product name, rather than the current 180 or so listings, there'd be many thousands of them. We'd not be able to see the results our our tests due to all the 'external noise'.

Cheers
Rod

Gosh, I have progressed to 'student' now? Wow, that's great. My aim is not to confuse, by the way. My aim is to point out that you are talking a lot of laughable rubbish.

What you are not understanding in my discussion is it is not the 'factors' that you have included in your test that I am doubting. It is the ones you have not included in your test that I am doubting.

So let me try again:

**What have you done in your experiment to eliminate the effect of domain names ( may or may not exist )? How do you know that the effects that you are viewing are not because of domain names?

What have you done to eliminate the effect of possible domain name rank ( may or may not exist ) ? How do you know that the effects you are viewing are not due to domain name rank?

What have you done to eliminate the effect of domain name age ( may or may not exist )? How do you know that the effects you are viewing are not due to domain name age?

What have you done to eliminate the effect of differences in content? How do you know the effects you are viewing are not due to differences in content?**

I really can't be bothered to go on after the differences in content. That is such a big glaring omission.

To use your example of the cars. None of the results have any meaning if some of the cars are being driven by Lewis hamilton and some of them are being driven by me. If that was the case there is a big factor that is being overlooked.

The first big factor that is being overlooked in your experiment is that the content is not the same. Duh! Earlier in the thread you said that you agreed that content is king. So, that surely is quite a big factor to overlook.

It is hard to understand how any one who believes that content is one of the most important factors in ranking would not understand that testing pages with different content is laughable. More than that it is hard to understand how anyone who believes that content is one of the most important factors in ranking might intentionally alter the content of some of the pages in the test ( to write a smug little 'victory' message which he later had to eat along with this hat #post443 ) cannot understand why some people think he is a complete jackass.

If you believe that your test is valid in any way then you will be able to calmly answer the questions in bold above. If you can't do that then I think I may have to look for another 'teacher' .....

Even the only other vocal supporter of this 'test' of yours has said:

Perhaps the main conclusion so far, RodG, is that we're not that great at SEOing a site? ;)

So, it seems that one set of conclusions would be that you aren't great at it. You sure as hell do not know how to design an experiment. So, what attributes exactly do you bring to the table as a 'teacher'?

Geez, even SPH says that it is not an experiment. It is more about some competition about who can get the highest ranking as some kind of way of 'honing your SEO skills' .

So, basically it is a pissing competition.

I tend to avoid pissing competitions.

Have fun.

niccol:

Gosh, I have progressed to 'student' now?.

If You wish to play the part of teacher then you need to do a lot more that ask (silly?) questions. You need to demonstrate the points you are trying to make. You have made no attempt to do so.

niccol:

My aim is to point out that you are talking a lot of laughable rubbish.

I have now provided no less that two different scenarios that demonstrates that you are talking a "lot of laughable rubbish."

niccol:

What you are not understanding in my discussion is it is not the 'factors' that you have included in your test that I am doubting. It is the ones you have not included in your test that I am doubting.

What you are not understanding is that my examples are demonstrating that we don't need to include all the 'factors' in order to make valid and logical conclusion for any given test.

niccol:

So let me try again:

What have you done in your experiment to eliminate the effect of domain names ( may or may not exist )? How do you know that the effects that you are viewing are not because of domain names?

If the domain names were relevant/significant this would reveal itself in the observation(s), and we wouldn't see the pattern of 8 of the top 10 listings without the KeyWords. The domain differences would randomise these results and we'd observe no discernible pattern/trend in regards to the URL's..

Furthermore if the domain names were relevant/significant we'd be able to observe that from the current listings... I dunno, maybe they would appear in alphabetical order or something, with the type of URL being a secondary or, tertiary observation, or perhaps we'd see no pattern to the URL's at all. In which case we could conclude the domain name is more significant than the URL's.
Your next two 'points' would be equally observable, even though they aren't being specifically tested either.

niccol:

**
What have you done to eliminate the effect of differences in content? **
How do you know the effects you are viewing are not due to differences in content?

The test (as I specified/designed) was for all sites to have identical content for the test being performed.

I didn't need to do this to prove/disprove the URL/KeyWord theory, because just like the domain names, if this were a significant factor for the test being performed, we would see the 'content' pattern in the test results, and the URL formats would be somewhat randomised as a result.

In fact this is what I was expecting to see happen if I didn't put this particular control in place (the assumption being Content trumps URL's), which is why I tried to get as many sites as possible to all use the same content for the initial tests.

As you know, I somehow lost this control factor, and as such, I was honestly expecting the URL 'test' to be inconclusive as a result, instead of the test showing anything useful about the URL it would/should have shown the sites with the 'best content' taking the top spots.

My next tests were to have been to keep the URL format constant, and change only the page content, and again, observe the results.

It isn't the individual tests that are important though. The important thing is what can we observe as a result of any given change/test, and what can we deduce or conclude from those results. As has been shown, this can often lead to surprises, with unexpected patterns emerging that may not even have anything to do with the test(s) being performed.

niccol:

To use your example of the cars. None of the results have any meaning if some of the cars are being driven by Lewis hamilton and some of them are being driven by me. If that was the case there is a big factor that is being overlooked.

Now this is the kind of thing I'm looking for. It isn't a problem though, we have twenty cars and only one Lewis hamilton. If lewis happens to be driving the 4 cylinder ford and makes one of the top 10 positions this would be an anomaly to the pattern observed that most of the top spots were taken up by the V8's.
We can either ignore the anomaly as being insignificant in regards to the test results, or, we can investigate this anomaly by having Lewis drive one of the other cars to see if it was just a fluke, or whether he is such a good driver that he can drive a 4 cylinder car faster than the average person can drive an 8 cylinder car.
This doesn't invalidate the observations, it just means we may need to do additional tests to isolate the significant from the less significant.

niccol:

The first big factor that is being overlooked in your experiment is that the content is not the same. Duh! Earlier in the thread you said that you agreed that content is king. So, that surely is quite a big factor to overlook.

See above. I didn't overlook this at all. I was very peeved that I lost this particular control before we were able to perform a few other tests that could have been difficult to observe otherwise.

niccol:

It is hard to understand how any one who believes that content is one of the most important factors in ranking would not understand that testing pages with different content is laughable.

I agree, but you really can't blame me for that.

niccol:

More than that it is hard to understand how anyone who believes that content is one of the most important factors in ranking might intentionally alter the content of some of the pages in the test

I did NOT 'alter the content'. That would have been pointless. I added a new page, identical to my 1st one in (almost?) every way except for a few relatively minor changes to the content. This is an independent test unrelated to the URL tests currently under discussion.

niccol:

( to write a smug little 'victory' message which he later had to eat along with this hat #post443 ) cannot understand why some people think he is a complete jackass.

Au Contraire. A Jackass (or its closest cousin, a donkey) is a very stubborn creature. What that post has shown is 1) I acknowledge I'm not infallible, and 2) I am willing to learn and adapt as new information becomes available. I'm not in any way ashamed of that. I'm more concerned about the type of person that denys the evidence of something because it doesn't fit their beliefs.

niccol:

If you believe that your test is valid in any way then you will be able to calmly answer the questions in bold above. If you can't do that then I think I may have to look for another 'teacher' .....

I have done as you have requested. Perhaps you could have the same courtesy and tell me why and where the car tests and/or logic is flawed.

Do I really need to test every variable to come to the conclusions I did?

Would it have changed the results if the cars were all different colours?
Would I need to test each colour car to come to the same conclusions?
Would I need to race during the day and the night to eliminate this variable?
Would we need to run the race with all air-conditioners on? off?

This list is endless and we can't hope to test every variable, nor do we need to.

Yes, you did raise a very valid point that the driver could make a difference. You should also have noted in my 1st set of car tests that I eliminated this variable as a factor by specifying the cars were to be driven 'pedal to the metal', therefore we'd be testing the cars speed only (ok, so now you'll argue that Lewis is better at steering and braking, so the pedal to the metal still doesn't cover all bases). This is just nitpicking though.

My final? input (for this aspect of this thread) is that (correct me if I'm wrong) there are so many variables involved that no test we perform is going to be of any use whatsoever?

Do my car examples really prove nothing at all because I haven't considered all variables?

If this is not correct, then isn't it about time** you** proposed a test or series of tests that will demonstrate how we **can **determine what methods do, and what methods don't work for SEO purposes?

It all fine and dandy to sit there and proclaim I've got it all wrong, but unless you can come up with something better you aren't really doing much to help.

Do you want to help our endeavours to see what is/isn't important for SEO, or is your only purpose here to criticize?

Cheers
Rod

No, sorry, you are still wearing the 'Jackass running a Pissing Competition' t-shirt in my book.

All you have to do to persuade me is answer a very simple question:

What have you done to eliminate the effect of differences in content? How do you know the effects you are viewing are not due to differences in content?

For a man of your intellect that should be easy, right?

(Or alternatively you can explain why, despite the fact that we all seem to agree that content is a major ranking factor, ignoring the effect content has on ranking is the best way to proceed.)

See previous posts about my positive suggestions as to how to run a meaningful series of tests. No need to repeat them.

The test (as I specified/designed) was for all sites to have identical content for the test being performed.

but they don't, do they?

which is pretty much the same as saying

'we thought we would have all equal drivers and they would all go pedal to the metal but in fact they all just did what they felt like.'

niccol:

Even the only other vocal supporter of this 'test' of yours has said:

"Perhaps the main conclusion so far, RodG, is that we're not that great at SEOing a site? ;) "

So, it seems that one set of conclusions would be that you aren't great at it..

I've never said I was. Fact is, other than content, I've always considered most of the 'SEO' methods are of such little significance they aren't worth a second thought.

If as a result of these tests and observations I'm proven wrong I'll most certainly adopt the ones that show significantly positive results.

niccol:

You sure as hell do not know how to design an experiment..

I'm waiting with bated breath for you to show me how it should be done.

niccol:

So, what attributes exactly do you bring to the table as a 'teacher'?.

Apparently nothing. This is why I'm handing the reigns over to you. The 1st thing I'd like to find out is if KeyWord loaded URL's are more (or less) advantageous than the non-keyword loaded URL's.

Please propose a test (or series of tests) that will demonstrate this. Or do you think it is an impossible task?

niccol:

Geez, even SPH says that it is not an experiment. It is more about some competition about who can get the highest ranking as some kind of way of 'honing your SEO skills' .

That's not quite how I read Steve's comments. Not that it matters. I'm still keen on finding out the significance (or lack of) in regards to the URL formats (which is essentially where this thread started).

Please propose a test (or series of tests) that will demonstrate the whether adding KeyWords to a URL will improve Google rankings (or not)

I've done the best I can, and you say its isn't good enough, which implies you either think it is an impossible task, or you have a better set of tests that will give us valid results.

Put up or shut up.

Rod

So if you are really trying to investigate the effect of keywords in URLs then you need to improve your testing environment. As I have said before my way of doing this which is kind of industry standard is to use matching sub-domains. In the past I have named them something like aaaaaaaaah.domain.com and aaaaaaaaai.domain.com . That is to try and reduce the influence of the actual domain name. Obviously I have tried to create them at similar times. And I have created them on domains that have no history in google. All of that is just to try and eliminate the effect of domain names. It is not perfect but it is a lot closer to perfect than what is going on now. I have taken similar measures to isolate the factor that I am looking at with regards to content, links and any other factor I can think of. One has to do that otherwise the results and observations are not valid.

This was there to demonstrate how one might attempt to eliminate the factors involved in domain name.

Again, purely in as an example, in the past I have run multiple installs of Zen Cart from the same database as a way of ensuring that the content is equal. Although this ultimately causes problems when investigating modules. Problems that can be overcome but still they add complications.

So, which bit of this are you struggling to understand?

I started running those kind of tests myself and with others about four years ago. To be honest I do not any more. The reason for this is because many of those factors had been researched to death. and the results had pretty much been demonstrated ( I am not saying that occasionally revisiting those experiments is not worthwhile, by the way ) . The problem in involving others, as you have pointed out, is that others may not stick to the experiment the way you'd like. So I always felt that solo investigation tended to have less chance of being screwed up by someone clicking on the wrong link or adding content to a page. That kind of action for me could screw up the results of a six month series of observations.

The other reason that I stopped these kind of tests was that the factors that i think are interesting ( just personally ) at the moment are more to do with usability and conversions. So you really do need the involvement of the public to make any testing worthwhile. The kind of testing in isolation that I was doing previously doesn't fit too well in this scenario.

One interesting thing that always emerges from doing this kind of test is that it always evolves. So, you chat to someone about your tests and they say ' But hang on, couldn't that effect be due to factor A that you haven't considered'. At which point you usually say 'But I don't think it is' Which is fine but what you need to do is demonstrate that it isn't. So, for example, I started using the sub-domains a.domain.com and b.domain.com until a colleague pressured me by saying how do you know that 'a' and 'b' aren't effecting your results. We did a series of tests that showed that probably they weren't but nonetheless we changes to using aaaaaaaaaaaaaaah.domain.com and aaaaaaaaaaaaaaai.domain.com because it seemed to make sense and when we tested it those two domains appeared to be equal.

So, in my opinion, the correct action when someone says 'your experiment is flawed' is ( obviously after a bit of self-righteous huffing and puffing ) to see if they have a point and design any ( even the slightest ) chance of them being right out of the experiment.

So, as a for instance. I might say to you. Can you prove that the effects that you are seeing are not being created by the actual domain name. Now, the honest answer to that is 'no I can't' . So my advice ( and this is positive advice by the way) is to go and redesign the experiment so that any chance of the domain name influencing or skewing the result is removed. It is not hard to do. My way of doing it is to use matched sub-domains. There are probably other ways of doing it.

That 'evolution' of your testing process will make your tests more meaningful. The next time someone asks 'what are the effect of domain names on your results?' you can proudly tell them that you have isolated consideration of domain names. And move on to thinking about other problems your testing environment has.

This approach to developing a test environment is not something I have just made up on the spur of the moment. It is pretty much how most research in both academic and corporate environments works.

I honestly think you are smart enough to understand this if you just take a step back. It is not pleasant to have your test described as flawed. But if you take some notice of the people who are describing it as flawed then it can be improved until no-one can see flaws in it. That is just part of the process. Those of us who actually have done research will just accept it as part of the process.

The big flaws in your test environment at the moment are :

1. you are comparing pages on very different domains
2. you are comparing pages that have very different content
3. you have no idea who, and how many times, people are clicking on links both on the pages and in the SERPs

all of those are actually fairly easy to solve if you have the energy to do so.

Products/category removed from both sites 10am EST on 1/14. Curious to see how long before they fall off. I would love it if someone showed proof that gibberish descriptions are not detrimental, all you Google algorithm experts out there who are all piping in. They can detect keyword stuffing in descriptions, they have many secret components of their very sophisticated algorithm that no one but a select few inside of Google know, so show us the proof please. It's obvious that having a different description was creating a lot of contention and I am really not into this verbally abusive arrogance and snottiness going on so no longer want to visit this thread or be a part of it. But it has made me want to run an A/B on subdomains to do some additional research for myself in various aspects of SEO/content/etc so thanks for that. Adios.

<pause_lurking>
Sorry SPH, I'm done. I've been sick as a dog, no where near recovered (gonna see the doc again if not better Wed), now I have a sick kid, too. If/when I can spend time in front of the computer, I have boatloads of work to catch up on. I really don't have the time to fiddle with it.

I'm kind of hoping RodG will stop trying to go tit-for-tat about wether the 'test' is scientific. If he just lets that ride, then maybe there would be more time to come up with a chart that tries to sort out what everybody did… all in one orderly spot. At least then there would be a final cheat sheet for looking at what HAS been done.

I don't think our original plan was really even trying to be 'scientific'. But I did find it helpful in putting my fears to bed. I think it might be interesting to see a final accounting, sans the debating.

I notice that your keyword url prod (prom mart) is #3. Above it is my native site (bizesitehost). But the meaning of that is clouded by the fact that my site has an incoming link from my original top placed prod link (in the footer of the other site). I did that (before this thread started) because eventually bizesitehost will be my actual website (too busy doing real work to fiddle with it) so I've been putting the link in the footer of all work I do.

Also above you (at #1) is the sabianstore native prod page. But that is clouded by altered description and whatever else was altered.

So if you set those two clouding issues aside, you are still doing quite well. I wonder if you have altered anything also?

Quick side note, you really need to quit sweating the 'gibberish' business. It is completely immaterial EXCEPT if it does not match the gibberish on the other sites. Gibberish for testing is fine, has zero effect on your real site, and simply gives you something to use for testing. As for extrapolating (maybe I should add gibberish to all my real products) on this unimportant aspect… well, that is just silly. :)

EDIT (Saw lankeeyankee's post just b4 I hit submit): Yea, all the negativity/hostility is what killed my interest, too. Talk about a downer. With regard to gibberish, I think if you use systemically (as SPH has considered, I think jokingly?), then you will have a problem. The fact is that using 'gibberish' or 'lorem ipsum' is a simple way of life for designers (professional designers, anyway). It is a very accepted tool that predates the internet and is used by pretty much every serious designer on the planet. If SE's penalized the overall ranking of otherwise good sites for having a few pages of test material, then my sites would have been in the toilet long ago. I very frequently use the method to get a feel for what real content will look like.

You might find this an interesting read:
http://en.wikipedia.org/wiki/Lorem_ipsum

[Back to SPH] I like what you say about just turning this product (wurldbitzer) into an SEO challenge. Bet that would be fun. See who could get to the top, hopefully they will share how. Other than that, I think the words wurldbitzer debibulator have run their course. Hopefully any new tests designed will choose a new set of words to play with. And hopefully you guys will all turn off the stuff that automatically generates all those ancillary links this time (unless you seek to test the use of those; if so, you will need to standardize the number of pages/products on all testing sites).

A while back RodG was wanting to do a different test. Might not be able to achieve rigorous standards in this environment, but testing still could be interesting/entertaining. I did find the original test to be helpful, despite weaknesses.

If you guys do move forward, to eliminate all the side chatter and noise, maybe another locked thread could be initiated? Discussion/debate here, but only test related information there. I don't know about you guys, but I don't think I could find RodG's chart post if I tried. If you divide the conversation from the recording of simple facts, there'd be less clutter and the efforts would be much easier to follow. Just a thought.

++++

I've kind of expected my original site to completely disappear by now. But the home page is still showing up (last page of results). I wonder what is causing that?

I have not removed the keyword URI that points to the home page. But because there are no links on my site to that URI, it should go away right?

A couple thoughts on that. If it persists beyond a month, is it because once google knows a URL/URI exists it will continuously check it until it receives a 404?

And if the home page (which is what that uri now points to) continues to show up in searches for the worldbitzer, is it because of that now defunct uri that contains the keywords? It seems like that is the only possible reason?

Here is the logic (might be a misunderstanding of Ceon here): If google penalizes duplicate content (they do, of course; I always thought it was completely ignored), then the second will be eliminated, right? So the home page is the primary content. But the URI won't show because it is duplicate content (just shows home page content). But if the URI containing worldbitzer remains, is still being followed by google due to legacy, and the only difference is the keywords in the uri, those keywords point at the home page, and google persists in showing it in the results… could that be caused by anything other than the keywords in the uri?

Definitely worth watching, IMO.

Thought you might find that interesting, RodG.
</pause_lurking>

RodG:

As usual, you missed the point.

Some/many of us here are trying to advance our knowledge and skills, by asking questions and performing various tests (which may or may not demonstrate anything conclusive).

It is obvious that you feel you have nothing to learn from being here, and by all means, hang around for the entertainment value if that's all you are interested in.

The bulk of your (and Nicks) input has been little more than unfounded negative criticism designed to do nothing more than confuse.Not missing the point at all, and yes I do find this thread quite entertaining..

However it seems that YOU are missing the point that both Niccol and I are trying to make.. Niccol laid it out (again)..

niccol:

The big flaws in your test environment at the moment are :

1. you are comparing pages on very different domains
2. you are comparing pages that have very different content
3. you have no idea who, and how many times, people are clicking on links both on the pages and in the SERPs

all of those are actually fairly easy to solve if you have the energy to do so.

You believe that it is possible to draw a valid conclusions from this "test". Awesome. I disagree for ALL of the reasons that have been pointed out. Following this thread one can see that this "test" is no longer a controlled experiment at all. There are no conclusions can be positively be extracted from this "test" due to this loss of control. I will continue to express this opinion because I am ALLOWED to participate in this thread so long as I am NOT violating any forum rules. If you do not agree.. GREAT.. but I can (and will) continue to express my opinion on this topic as I see fit.

Though you don't seem to think so, I am interested in advancing my knowledge (which is why I clicked on to this little thread to begin with). However, as others can attest to (and have) this doesn't seem to be possible without a barrage of insults and implying that folks are less intelligent than you are. "questioning" anything seems to leads to snarky responses and hand slapping. and so I am CLEAR, this kind of behavior has come exclusively from YOU RodG.. But like I said earlier I think this is just how you get down.. Perhaps you are not aware that you speak to folks using the pompous tones you do when you respond. (and if you do realize this, then that's worse..)

Questions are part of the process, and if that annoys you, well that's just too bad.. You do not have the right to dismiss anyone from participating in this discussion because you don't want anyone expressing an opposing POV.. (and your responses do give the APPEARANCE that you do not wish to be opposed or questioned) You can choose not to answer/respond if you think it's somehow beneath you (which based on your response to all but those who agree with you, you do), but the continued browbeating and snarkiness that you display is TRULY unwarranted.

RodG:

As I've been told many times, If you can't say something positive, don't say it at all.

RodWell that is certainly the pot calling the kettle black isn't it.. Too bad you've not demonstrated that in this thread.. So critiquing this "test" for it's OBVIOUS flaws is not saying anything positive eh?? Thanks for the information.. Let me make a note of that..

niccol:

The big flaws in your test environment at the moment are :

1. you are comparing pages on very different domains
2. you are comparing pages that have very different content
3. you have no idea who, and how many times, people are clicking on links both on the pages and in the SERPs

You have already emphsised enough times that my logic is flawed, and repeatdly told me why it is flawed (at least in regards to SEO). No need to keep repeating.
I noticed that you chose to not answer my yes/no questions in regards to the car scenarios (the principles of which can be used for SEO tests) What was the reason for ignoring these? Is Yes/No too hard for you? You don't see them as relevent?(why not?)

niccol:

all of those are actually fairly easy to solve if you have the energy to do so.

This is the problem Nick, I've done my best to devise and develop tests that should produce useful results. I've put a shotload of energy and thought into this 'fairly easy task', and apparently I still can't get it right.

I will ask once again, please demonstrate, not by telling me all the things I'm doing wrong, and not by telling me what I should be doing. I am incapable of solving the puzzle of how can I come to any valid conclusions without taking every conceiveable variable into account. (not just with SEO, but also the car scenarios, and indeed anything that has more than one variable to consider). I'm finding this incomprehensible and I'm 100% positively convinced that we don't need to consider every possible variable (even the unknown ones) to come to valid conclusions about any given test. You are clearly saying we do. Please, help me understand. Nay, not just me. I'm sure there are many lurkers here just as much in the dark as I am. Will you help us? please?

Thanks
Rod

feznizzle:

edit (saw lankeeyankee's post just b4 i hit submit): Yea, all the negativity/hostility is what killed my interest, too. Talk about a downer. With regard to gibberish, i think if you use systemically (as sph has considered, i think jokingly?), then you will have a problem. The fact is that using 'gibberish' or 'lorem ipsum' is a simple way of life for designers (professional designers, anyway). it is a very accepted tool that predates the internet and is used by pretty much every serious designer on the planet. If se's penalized the overall ranking of otherwise good sites for having a few pages of test material, then my sites would have been in the toilet long ago. I very frequently use the method to get a feel for what real content will look like.

You might find this an interesting read:
http://en.wikipedia.org/wiki/lorem_ipsum
---bingo!!!

I am not sure how I can demonstate without saying what I think is a good approach. In my opinion:

I would set up two sub-domains
I choose to name them as similarly as I can
I would password lock them so they can't be indexed
I would install zen on both sub-domains simultaneously ( yes I know that is impossible)
I would not add any templates or other stuff that alters the html that is sent to the browser ( as much as possible )
I would be particularly careful to keep the link structure identical so have the same number of products on each site with the same description etc.

at this point I have pretty nearly identical installs which logically should perform identically in google. ( This itself is a bit of an assumption but you can run a control pair of subdomains to test that assumption )

Then I would change one single thing on one of the installs. So, if you are interested in 'keywords' in the URL just change that. Nothing else.

I would not publish where I am doing this publicly because people clicking around can effect the results. Sorry, by the way, I clicked around your sites without thinking about it!

Then remove the password lock and let the games begin.

This way I have isolated the single change that I have made. Logically if there is a difference in performance then it must be due to that single change I have made. Of course, some idjit like me will come along and say that just doing it on a single pair of sites then the results are dubious. But can you really be bothered to set up a hundred pairs of subdomains? I might also say that any result you get in the first period is going to be noise rather than signal so that you need to wait a bit longer than you have.

That is how you might, if you chose to, set up a slightly more rigorous 'test'. I genuinely do not mean this negatively. It is a positive suggestion about how you can proceed and improve your results.

Having said that I kind of appreciate the pissing competition approach because it does make people think about stuff and does raise questions that may not have been considered before. I agree that it can be kind of fun and can highlight interesting areas.

Listen, Rod, I genuinely apologise if at any point I have over-stepped the bounds of politeness. Which I know I have. I am going to quietly leave this thread now. I really wish you the best of luck with it and hope you get useful and interesting information.

PS. I think that if we go back and look at the first post of this thread we are all going to have to admit that we are all very much 'off topic' :-)

Feznizzle:

<pause_lurking>

Welcome back. Hope you're feeling better (sincerely).

Feznizzle:

I'm kind of hoping RodG will stop trying to go tit-for-tat about wether the 'test' is scientific.

I've effectively given up going 'tit for tat'. It is now left for Nick to develop tests that will be useful.

Feznizzle:

If he just lets that ride, then maybe there would be more time to come up with a chart that tries to sort out what everybody did… all in one orderly spot. At least then there would be a final cheat sheet for looking at what HAS been done.

For what purpose? Unless we cover every conceivable variable, what wev'e done, and all observations made as a result will be totally invalid. (Niccol and DivaVocals have taught me that lesson).

Feznizzle:

I don't think our original plan was really even trying to be 'scientific'.

Perhaps that wasn't your plan, but it was mine. I never realised I was setting myself up for an impossible task though. I truely beleived that we could get some useful/valid results without taking every conceivable variable into account. Boy. did I end up with egg on my face.

Feznizzle:

Quick side note, you really need to quit sweating the 'gibberish' business. It is completely immaterial EXCEPT if it does not match the gibberish on the other sites. Gibberish for testing is fine, has zero effect on your real site, and simply gives you something to use for testing.

Up until this morning I would have agreed with this, but we can't come to this conclusion (that it is 'immaterial' and has zero affect on a site) because we haven't tested it with all domain names of all ages, not to mention all the other variables we haven't tested with.

Feznizzle:

EDIT (Saw lankeeyankee's post just b4 I hit submit): Yea, all the negativity/hostility is what killed my interest,

Pretty much killed mine too.... but, I've now got something to look forward to, Nick is going to devise a new set of tests for us, something that elimates all of the flaws in the tests I devised. I can't wait to see what the new test data and results reveal.

Feznizzle:

I think the words wurldbitzer debibulator have run their course.

Not only that, but any and all observations made from the results are totally worthless. Don't take my word for it, both Niccol and DivaVocals have made this very clear. At least to me. I've learned from them, and I think you should too.

Feznizzle:

Hopefully any new tests designed will choose a new set of words to play with. And hopefully you guys will all turn off the stuff that automatically generates all those ancillary links this time (unless you seek to test the use of those; if so, you will need to standardize the number of pages/products on all testing sites).

I'm sure that Nick will have all of this covered.

Feznizzle:

I've kind of expected my original site to completely disappear by now. But the home page is still showing up (last page of results). I wonder what is causing that?

This has caused me much headscratching (probably more than you'd imagine). It is an anomoly for sure... I couldn't explain it, until now :)

Feznizzle:

I have not removed the keyword URI that points to the home page.

That's the answer. :) Thanks. There is/was no way to determine this from Google's side, nor from visiting your site (unless we knew about the keyword URI/URL).

Feznizzle:

But because there are no links on my site to that URI, it should go away right?

Wrong. It is the URI/URL that is keeping it 'alive' . It won't go away until you remove this.

Feznizzle:

is it because once google knows a URL/URI exists it will continuously check it until it receives a 404?

Yup. That is the way I see it.
Another possiblity, along the same lines, is if the URL is in the google sitemap. That will also keep it alive, even if the page itself contains no relevent data.

Feznizzle:

Here is the logic (might be a misunderstanding of Ceon here): If google penalizes duplicate content (they do, of course; I always thought it was completely ignored), then the second will be eliminated, right? So the home page is the primary content. But the URI won't show because it is duplicate content (just shows home page content). But if the URI containing worldbitzer remains, is still being followed by google due to legacy, and the only difference is the keywords in the uri, those keywords point at the home page, and google persists in showing it in the results… could that be caused by anything other than the keywords in the uri?

Definitely worth watching, IMO.

Thought you might find that interesting, RodG.
</pause_lurking>

Well, it certainly has solved the mystery for me. Thanks. Your logic is flawless. (unlike mine, apparently) :)

Cheers
Rod

bn17311:

just wondering how many sites joined in test how many used standard zencart url

Here's my take, using ONLY ORIGINAL "GIBBERISH" DESCRIPTION:

DYNAMIC URL (NATIVE) SITES:

1. RodG: shop.vcsweb**.com**

2. BN: onyourbike**.co.uk**

3. BN: cscycles**.co.uk**

4. FEZ: bizsitehost**.com** (included because it doesn't look like he pulled this site?)

PRETTY URL SITES:

1. SPH: promdressesforless**.us** (Ultimate SEO URL)

2. SPH: fashion-mart**.biz** (Simple SEO URL and changed template the other day.)

3. BN: lessthan**.co.uk** (CEON, haven't yet noticed in listings)

4. SPH: prom-mart**.com** ( CEON, recent entry now making appearance with ancillary links. Not to be confused with prommart.com*)

I'm emphasising the domain extension, in case it is, or can become, a factor. So far, mine are holding well without .COM'S.

So far, looks equally matched re numbers of sites entered.

I recently entered three subdomain entries under prom-mart.com - demo-1, demo-2, demo-3 - two with Ultimate SEO URL and one Native Dynamic URL, but have not yet noticed an appearance.

*I also changed templates the other day on prommart.com. Wonder if changing templates can be a factor with google. Must watch fashion-mart.biz for that.

RodG:

If I fail you this time, I will need to refer you to another teacher... I think 'bn' may be a good candidate.

Rod

:laugh:

EDIT:

1. RodG: shop.vcsweb**.com**

Rod actually has two entries/links: product id #222 and product id #223. Last I looked, #223 had more added text.