Ok, one technical question here:

Did you remove the google_ad_channel line from your ads, or did you change it to:
google_ad_channel ="";

I wouldn't suggest removing the 'google_ad_channel =""' code as this would probably be a TOS violation. I'm fairly certain that the coding without custom channels still includes this snippet by default.

As an aside, I hadn't been using channels (URL not custom) for a while, but I reactivated the channels after adding ads to a few more pages yesterday morning.

It's still too early to tell, but this resulted in a series of non-targetted ads (and even other-language ads) yesterday with a CTR dip.

Maybe I'll see something more promising today.

No, if you don't have any custom ad channels the google_ad_channel = line isn't in the code, so you can just remove it.

Mark, I thought so too. But I just went and looked at a back up of some old Adsense code from January, and the ad unit includes the line

google_ad_channel ="";

so it would seem the Adsense code generator creates it both ways depending on some unknown condition.

With no custom channels currently set up on my account, if I go create an ad unit now, it does include that line.

I stopped being lazy and cut and paste the code directly from the Adsense site:

--------------
<script type="text/javascript"><!--
google_ad_client = "[omitted]";
google_ad_width = 728;
google_ad_height = 90;
google_ad_format = "728x90_as";
google_ad_type = "text";
google_ad_channel ="";
google_color_border = "[omitted]";
google_color_bg = "[omitted]";
google_color_link = "[omitted]";
google_color_text = "[omitted]";
google_color_url = "[omitted]";
//--></script>
<script type="text/javascript"
src="http://pagead2.googlesyndication.com/pagead/show_ads.js">
</script>
---------------

It's definitely there. Question is, do you have the chutzpah to mess with it? :)

Looks like Fred beat me to it!

D'oh, my error. Apologies for that. It does indeed keep that line in the code. I must have been thinking about the alternate_ad_url line.

[edited by: MarkJH at 4:33 pm (utc) on Aug. 1, 2006]

But no, Mark. I'm almost positive that when this all started about removing your custom channels, I went to the Adsense page and after removing my channel info, did a test ad unit creation, and at that time it did not contain the line (although doing so now, it does).

Interesting-- almost like they responded (albeit proactively, not retroactively) to this whole custom channel experimentation bit.

Well that line certainly did not get created a couple of years ago before channels were introduced (I even checked, just in case). So its presence is not absolutely required. But I agree it's better to use the code exactly as supplied by Google.

Has anybody beside me noticed that there is no space between the equal sign and the quote mark on ONLY the ad channel line?

I've checked all of my code and it is always that way. Anybody with expertise in javascript is welcome to comment on whether or not that is meaningful.

Example:
google_ad_type = "text";
google_ad_channel ="";

The difference is definitely not meaningful to javascript.

Whether or not it's meaningful to G when they're examining your code for possible tampering (which is against TOS) is known only by them.

So, at the Google end, is there any functional difference between us including the channel parameter (even if the value is empty) and not including it?

I don't know much about databases and javascript.

I generated new ads to replace the adchannel ones and yes, the code line was there.

Ann

Back again posting in my "silly" thread.
Well folks it is no longer a test. Things are still climbing.

Ended the month about 400 dollars over June, greatly enhanced averages and the higher payout per click is still holding up and rocking back and forth to a higher number.

I won't be telling my actual figures anymore, other than the one above, as I have been made fun of in another thread that I can no longer find...I have no idea what went with it but that is ww's business and not mine---'nuff said.

I hope those who tried my idea had good luck also.

Ann

I hope those who tried my idea had good luck also

Unfortunately it's not good news here Ann. Just had my lowest CTR and eCPM in 3 months yesterday. Things have been getting increasingly worse since I started the test. If you look a few posts back I predicted this would happen to me because my theory is once channel information is removed then the algorithm uses the whole of your unchanneled AdSense history to decide your payout per click. As my few months with AdSense were extremely poor (many zero earnings days) this is now driving down my CTR / eCPM. To me this is a ridiculously flawed algorithm.

Is there anybody else who has seen their CTR / eCPM plummet after the test?

Yes, Ann, many people make fun of ideas that they don't understand. I guess it's part of human nature. Sorry that it is that way, but it's their loss.

For the record, since I removed my unused channels eCPM has gone up just over 10%, which is an increase capable of supporting an average family in the United States. I don't undertand it and I'm certainly not banking on it, but it has appeared to have worked for me.

Fraklic and dataguy,

Fraklic,
I'm sorry about your situation but you could try going in and streamlineing your site: Removing all pop unders and flashy banner ads, check out the heat maps, put adsense (just one linkunit) on every page you can use under the TOS. Giving more page impressions could serve to make your site more appealing to the bot and draw some better ads in. Maybe that will help.

I did the same but ONLY after I was sure the test worked and was still working. On the 31st Google shorted me a few thousand impressions as that is my biggest visitor day of the month, every month. Yesterday I think they released about 3 thousand of them. Maybe more tonight I hope. They are probably checking them out...that;s fine, it should serve to reinforce my accounts (sites)worth.

Dataguy,
Thanks for the kind words. I am so glad you have showed some improvement. Keep up the good work. :)

Ann

Thanks Ann. My site has been streamlined from the start, one ad unit, no pop ups or flashy ads. Ad unit is large rectangle above the fold and centre of screen.

Apparently Google doesn't rate my site as a quality site. But I know it's quality and so do my visitors. That's all the matters to me.

Good. Here's hoping they get their act together. :)

Ann

Just thought of something. That large rectangle may be driving clickers away. I no I never stick around to read the sites that throw the ads in my face, maybe you have had better luck with them but the ad I use is the adlinks and it is blended in so as not to be annoying..

No that shouldn't be it Ann as CTR has been stable for months now just with that ad unit. I'm hoping it is just a screw up from Google's end Actually I'm pretty confident it is.

What do you think of the theory that when you delete channels that Google starts factoring in old unchanneled data to determine your CPC?

Actually, I think that they go by what your site is doing now and go forward with that. Old data would not be of any use to them but having to reevaluate the "old way" could just level the playing field between the quality sites and the MFA's. I think the MFA's would lose big time in that scenario..

Just my opinion for what it's worth. :)

I wish I knew more so I could help out.

Ann

I tried it for ten days, but it didn't really seem to do much for me - but my "busy season" ended on July 5th. Next year I'll try to remember it around April, when things start picking up for my main site, and see how it works when I have some significant traffic coming in.

It did cause me to clean out some old garbage channels, and rethink how I wanted to re-do them anyway, so it was still certainly a useful idea for me.

Back again posting in my "silly" thread.

I'm sorry some people have felt this way.

I think the problem is that results are patchy, and nobody can come up with any real idea of why it does, or doesn't work. Other techniques we have, we are aware of why they might work, and what type of pages might benefit or not. IMHO the way forward here is to try and come to some concensus of what factors may be involved, or what sites that this works on or not have in common. Once we have some idea as to what sites this might work on I think we have a chance to try and figure out why it might work (or not).

I've been trying it for nearly a week now and not seen any variations that I wouldn't consider normal. I removed all channels and replaced all ad blocks with newly generated code.

I have only the one site on my account (the sites I tinker with are on my wife's account). I've always deleted unused channels because it makes the interface a bit tidier, and I've kept my ad blocks only of they work and been blocking MFA's for over a year. I can't think of anything else pertinent to mention.

Has anyone noticed anything in common with other sites that work / don't work at all?

nobody can come up with any real idea of why it does, or doesn't work

I thought some non-idiotic thoughts on the subject scrolled by.

• Could stimulate reassessment of your content, leading to different ads being displayed (resulting in either better or worse CTR). Nobody seems to be trying to rigorously track changes in ads being displayed, so this is hard to rule in or out.
  
• The stats gathered via channels may be one input variable to SmartPricing, so changing your channels may affect your payout (for better or worse EPC).
  
• Could be too small a data set to mean anything whatsoever, or could be completely coincidental (since reversing the changes is not being tested). E.g., my July was nearly twice my June, and I made no significant changes whatsoever. If I ascribed the increase to the insignificant changes, and others tried the same thing, some would (randomly) see "success" and some would (randomly) see "failure".
  

Ann, tell us specifically how your site has been doing on 1st and 2nd of August. Is it good or low CTR like many others?

It is Excellent! Earnings are up and all numbers across the board are up also.

Am seeing improvements in income over even a very good July 2nd to 31st.

Ann

Could stimulate reassessment of your content, leading to different ads being displayed (resulting in either better or worse CTR). Nobody seems to be trying to rigorously track changes in ads being displayed, so this is hard to rule in or out.

That's one thing I was looking for. IMHO, to change ctr dramatically you need to change something or other. The things that often help here are to change position of the ad block, or different ads being selected. In my case, I put the ad blocks in exactly the same spot, and I didn't see any changes in regular advertisers. This didn't surprise me, as the epc is fairly high on my site, and I can't see any advertisers paying more than the ones I see all the time. Why they used to remove them and place MFA's instead is another story - not for this thread. Maybe others that have tried this have noticed changes. Less MFA / scraper ads, better targeting etc.

The stats gathered via channels may be one input variable to SmartPricing, so changing your channels may affect your payout (for better or worse EPC).

I agree this is an obvious one, but the variation seems to be very large, meaning they would have to be taking this into account a great deal. And that's the point that's been bugging me as it doesn't really make sense to me. I personally feel that the changes in fortune are as the result of indirect consequence, rather than direct consequence as this is.

Could be too small a data set to mean anything whatsoever, or could be completely coincidental (since reversing the changes is not being tested). E.g., my July was nearly twice my June, and I made no significant changes whatsoever. If I ascribed the increase to the insignificant changes, and others tried the same thing, some would (randomly) see "success" and some would (randomly) see "failure".

I've put back some channels, as I've been getting a lot of 1c clicks the last few days - I need to try and find out where. I can't conclude anything from my expermiment and return of channels. My changes in fortunes the last few weeks seem to be dependant on the darned Google updates and outages.

I think that there is something in this, but I think that whatever happens is as a result of indirect changes related to channel removal - not as the result of channel removal itself. I think we need to accept that it does seem to work in some cases, and take stock. It doesn't seem to adversely affect sites unduly, so we need to look at what changes on sites that it works for.

What we need to assess is:-

• What metric(s) change for the better
  
• What ones stay the same
  
• What metrics change for the worse
  
• What changes are seen in ads shown
  
• Any other changes made to the site that may have affected stats
  

Any other changes people think might be happening.

So maybe if agree what points to try and focus on, patterns may become a bit clearer. At the moment it seems the thread is in the main surprise that it works, or not - not looking at why. That's the big question we really should be working towards now.

So is there any input from those it works for as regards to a) what we should be trying to analyse here, and b) specific changes they have noticed as listed above.

I do not know why it works and given the secretive nature of Google I have no way of saying.

I can only say that i saw an improvement immediately in the ads being shown and the payouts per click.

In the beginning I did just what I said about channels and added one ad to all the other pages that met the TOS to give additional pageviews in order to show usefulness and space for Googlebots benefit.

Later, after I saw the change was holding up I removed all popunders and banners from every other ad company because I got angry at the stupid ads they were serving up to my visitors. The results has been sustainable....why I could not hazard a guess other than the ones I said in here about channel info being used to rate or weight the account by using the lower paying data we were so carefully providing the bot and that was used against our sites, account wide, to bring in lower paying crap ads.

I haven't the slightest idea what the algo is or how it was supposed to work but I would be willing to bet the result it was giving was not the one Google intended. I think it was intended to help but someone jigged when they should have jogged.:)

I know that my numbers continue to rise, on a see saw basis, and my traffic is showing some great growth
plus people are staying around longer than they used to judging by the pageviews.

Ann

Ann,

Sounds like your improvement is due to removing the competition and letting google play its hand unobscured.