[quote]Had Ann ONLY done the channels in her test, then it would be easier for me to discount the role all those AdLinks might be having on her bottom line and get on board that custom channels were indeed somehow responsible for the boost. [//quote]

I have said over and over again that I put the adlinks on following the introduction of the channels portion of the test and after the test was well started and showing improvements.

Why? Because after I was done with that part and it showed me such good improvements I thoughtr that MAYBE more page views would make my site look more enticing to the Googlebot and perhaps even higher quality. It took me a little time to hunt down (using my own webstats) and manually add 1 adlinks to the pages I found.

On the channel pages I had not changed the type of ad that was on it already, I just got a brand new version of the same ad to replace it with.

Got company, must run.

Ann

I don't think anybody believes that removing channels will cause smart pricing to go into a flurry and dish out more cash.

You don't? Reread the thread, a LOT of people believe exactly that and have developed some rather elaborate theories about it. I'm not trying to bicker with you, because you're one of the voices of reason in the thread, but I think you're speaking for yourself and not for others.

Well, yes that's true of course, but I thought we'd pretty well beat that idea out of them? ;) Quite honestly to believe it's that simple is nuts. I'm not afraid of saying that!

BTW your suggested mechanism is very similar to the one I posted back in msg #3028147 (LOL, I hate this new numbering system).

So great minds obviously think alike :). It's been niggling at me since this thread started. I don't disbelieve posters, and nor do I believe it's as simple as people might think. It clearly is fairly random who it works for / against, therefore it has to be something triggered by the channels removal and not the removal itself.

> I have said over and over again that I put the adlinks on following the introduction of the channels portion of the test
> and after the test was well started and showing improvements.

ooops, missed that (sorry about that Ann ...)

s'all right.:)

Ann

What about the theory that up until this experiment Ann was penalized by G for no apparent reason (no disrespect Ann, just speculating the possibilities), and by removing the ad channels, the mod date on the files was changed and ole man crawler saw that, hence re-crawling the site with a fresh point of view. AND in doing so, it recalibrated Ann's site to where it should be at now. Therefore, bringing her revenue to where it was suppose to be versus some kind of tactical promotion as we have been perceiving it... Just a thought.

Penalized? What are you talking about? You have me confused with someone else.

Google has always crawled my site on a regular basis and I have spent years in the top ten with my main site.
Ann

There is an Anne on this board as well as an Ann.

Penalized? What are you talking about? You have me confused with someone else.

Google has always crawled my site on a regular basis and I have spent years in the top ten with my main site.
Ann

Probably meant smartpriced. What you wrote has nothing to do with smartpricing "penalties".

It's plausible. You (none of us) can possibly know the effects of smartpricing on our sites.

So now it's become a grand conspiracy, even though Google already have a database that contains every impression and every click from every account.

I'll restate the hypothesis, since it's clearly not understood.

• We don't know what the algorithm for SmartPricing is.
  
• Given past behavior, it would be unsurprising for the SmartPricing algorithm to take advantage of data mining based on every available piece of data Google has.
  
• While we see data to infer that Google does track impressions and clicks, I know of no data to indicate that Google secretly maintains this on a per-URL basis. If you believe they do without some evidence, then I think that requires a conspiracy theory to explain why Google secretly can scale up to huge numbers of URLs per account, but still restricts us to 200 channels.
  
• Given that it's reasonable to assume that Google doesn't secretly maintain per-URL information outside of publisher-defined channels and arbitrarily refuse to give publishers access to more than 200 of those, it's also reasonable to assume that channel data is fed into the data mining effort that produces the SmartPricing algorithm. When you're data mining, the general philosophy is to throw everything into the pot, since the whole point of data mining is that the computer can uncover correlations that humans would never think to look for.
  
• And, of course, if channel data can affect the SmartPricing algorithm, then a single publisher might affect their own SmartPricing outcome for better or worse by tinkering with their channels. It's also reasonable that channel tinkering would have no measurable affect at all for many publishers, those for whom other variables in the algorithm happen to swamp the effect of the particular variable supplied by channel data.
  

Even though there's no earthly reason for them to want to downgrade sites that use channels.

This hypothesis does not assume that, on average, using channels will represent either a gain or a loss for publishers as a whole. It hypothesizes that an individual publisher could suffer a loss, a gain, or no change at all as a result of tinkering with their channels.

Even though a decision to base smart pricing on channel data, when a significant number of clicks aren't tracked that way, would have been monumentally stupid.

This sounds much like the viewpoint of folks who keep searching for the 5 key SEO tweaks that will keep them on top of the SERPs. Data mining is not a process of turning "knobs", as is so often visualized in these forums. It's a process of throwing as many different data sources into a pot as you have access to, and letting the computer figure out what factors actually correlate with the desired outcome (quality web pages, advertiser ROI, etc.). There's no process of "deciding to base" SmartPricing on anything -- you let the computer figure out what to base it on.

This hypothesis may be completely false; I don't care one way or the other, since neither outcome will affect what I'm doing. But I know of no existing publicly available evidence that strongly suggests it is false. And it certainly does not rely on any kind of conspiracy. It merely assumes that Google uses data mining here, as they have done elsewhere, and that they use all available information, as they have done elsewhere.

Kind of flattering that you'd pull my comments apart in such detail, 5 days and about 50 posts later.

I'm not going to respond in detail [sound of distant cheering], but I will point out that of course they have access to per-URL data, if they choose to use it, and there doesn't seem to be any "this hypothesis" yet anyway. Everyone seems to think it works a different way, some of the beliefs being consistent with what you wrote and some not.

Ronburk, thats an excellent summation of fact and hypothesis. Genuinely enlightening.

Of course they'd throw the whole lot in. I feel pretty daft not realising that in the first place.

Hey Ann!

If nothing else you started one heck of a thread, running over 3 weeks and 340 posts!

Which means you've provoked a bunch of thinking, which is usually a good thing. :)

I've been a skeptic on this, but as stated by ronburk, I can see how Channel Pricing Analysis, or CPA for short, would work....

Still not sure I want to join the experiment, as I might be one of the ones who would lose out.

Thanks, ronburk, and thanks, ann, for kicking this off.

Ronburk,

Thanks for the insightful post. Your points on data mining were eye-opening, for me.

If I stirred up some creative thinking then I think you all have made this a good thread. :)

I could not have sustained it alone but you are all welcome for the many thanks and thanks to all of you.

Update:

Income still holding up and rocking although not as rapidly as it did at the start. I am happy it is lasting this long, LOL Never know if or when it could get changed to a new way of doing things at the 'plex.

I still thank Google for the things they have enabled me to acquire with the earnings from them. :) Just bought 2 matching leather loveseats, one of them a motion (2 recliners in one), a matching Range and refrigerator (black and silver) as well as the property I now own. Quite an accomplishment for an "old" woman. :)

Ann

Just to be clear, does the official version of the theory state that Google don't know which page a click comes from, or that they do know but throw away the information?

[...Just to be clear, does the official version of the theory state that Google don't know which page a click comes from, or that they do know but throw away the information? ....]

They always know that but they may select to ignore that.

I think there is some evidence to say that Google doesn't record or store the url a click comes from, unless a channel is created:
If you have a domain name as an url channel, impressions and clicks wil display against it for the index page and all its 'children'
However, if you then go and create an url channel for one of those children, you do not get the past impressions and clicks for the new channel, but only future ones.
This leads me to think that Google does not record the 'raw' data, but only appropriate summaries.
I assume this was a storage size issue, as it would get tera or even peta -big very quickly.
This would mean channels would be a way for them to track our data, on the assumption that we are breaking our pages into sensible related groups.
Seem reasonable?

hey ann,
just wanted to know, what is your stand on "removing adsense from non-performing pages"?

You have done that?
OR have adsesnse on each and every page?

thanks

right now I am giving them as many page views as I have. I tried removing ads off of poorly performing pages and it didn't help me for more than a day or two.

Ann

Just confirming....so you are putting adesense on as many pages as possible, wheter it makes money or not? Right?

I have already does all the steps in your test:
Remove custom channels
Remove adds tracked by custom channels
Keep only url channels

Now I am wondering wheter I should remove adsense from pages that are making little or no money or put adsense ass well as affiliates on those pages.

Anyone, on what the effect of this might be?

Thanks

Ann, I don't think the idea is too far fetched. I thougt I was the only one who saw a difference when I added or removed channels. In my case, my income actually increased when I added channels that happened to perform well. Who knows?

Sounds like it may be worth a try :)

Ann

After reading through this entire thread I will make a few observations...

To me the biggest impact was reducing the number of ad units. This would automatically boost your earnings per click because higher priced ads get most of the exposure. With more units you have chances for lower priced ads to get clicks.

Without having any kind of control on traffic, or anything else, you can't really determine that the removal or change in the URLs has anything to do with the increased income. There are simply way too many factors to take into account, many which changed or are continually changing to say it was because of the removal of channels.

From what I can gather CTR is also up, which can't have anything to do with removing channels, and if CTR went up significantly at the same time that the number of ads displayed went down, well that is even more amazing, as you have less opportunities for clicks per page. So I can surmise there are a lot of factors here that come into play and there is no really solid evidence from anything posted in this thread that removing channels will beat smart pricing, or that it has anything to do with Smart Pricing.

I appreciate Ann sharing her experiences, and glad she is making more money, that is all good. If others want to expirment as well, I say go for it, as who knows what might be discovered. I just don't think anyone should jump in unaware thinking that somehow evidence of a smart-pricing workaround has been presented here. Besides just normal day-to-day factors, adwords changes, and everything else, the changes Ann made were significant and numerous. It is very likely none of them have anything to do with Smart-Pricing.

As long as people are aware of that before they go dumping their channels... it seems good to me...

I tried this for 14 days with no fruitful results :(

Two weeks into the experiment and after a half way point resurgence eCPM has tanked again. I'm thinking this is due to seasonal factors i.e. zero competition in my sector and advertisers are bidding less. Updates to follow.

Fraklic and all,

you are probably right and channels had nothing to do with it. It could have been just an anomaly where all the changes together helped.

At this point in time I have to say I just don't know for sure. Seems if channels was the part of my experiment that worked it would also have worked for all the rest...Huuum, you guys have me thinking again. Back to the drawing board.

My income is about right for August and running ahead of July like last year. At least whatever the reason I seem to be back in the ballpark earnings wise.

Ann

Thank god I stayed in the boat :)

Thank god I stayed in the boat :)

If you haven't got something good to say...

Ann I'm happy I tried the experiment for several reasons:
1/ I could sleep easy at night knowing that I had tried it :-)
2/ I as quite happy to conclude to myself that Google had not done a silly thing like factoring in use of channels when determining earnings. That was a big plus.
3/ I have really cut down on my tweaking and stat checking and have this month concentrated more on what's important, the happiness of my site visitors. Mega bonus.

So in conclusion I'm quite happy I 'jumped ship'.

Yippee,

Don't be so negative.

If I had Not tried ALL my changes and just stayed in the boat rocking along, like some scared people have a habit of doing, I would not have managed to turn my situation around to more than quadruple my earnings from the low they had fallen to in order to have a fighting chance of getting a positive growth pattern kick started again. doesn't matter whether it was with Google or some other "partner".

What counts is trying because throughout life the winds of change are constantly blowing and those who resist are always swept away. If one can learn to "work within the flow" eventually that person will be a winner (survivor).

Channel data may not have been the "real" reason but at least Trying helped my peace of mind as well as my income.

So Yippee and others like you, go ahead and stay in the boat while waiting and watching braver souls take all the risks so you can then jump in and "pirate" the positive results. When your boat drifts away don't forget us. :)

Ann

well, I lost out and I wouldn't change a thing either. I'm glad I tried. Putting the channels back on my site got my thinking about placement and colors and, more importantly, I have a massive A/B testing experiment running site-wide.

So I'm pleased to have played a small part in this experiment.

They'll be laying plaques to us in a hundred years' time...

"Those channel pioneers..."