| 8:34 pm on Nov 9, 2005 (gmt 0)|
I have been thinking about doing this as well. I have something like 40 channels in the site, ctrs varying from 1% to 4,7%
However there are a few channels that have a lot of impressions and low ctr, one being a link unit in my site's footer that has very very poor ctr.
Should I remove? My overall ctr would be 2-3 times better if I remove them.
| 8:43 pm on Nov 9, 2005 (gmt 0)|
You should see the effect in a week, if it's going to happen--smart pricing updates weekly. I've done this three times and have seen it work three times. Overall earnings went up each time.
However, this doesn't seem to work for all sites, and it's also possible that you axed too many channels. When I did this, I removed code from pages that either were getting NO clicks at all or very few, and low-value clicks. If you take the code from pages that are bringing in some income, then that loss might cancel out any beneficial effect from improved CTR.
Re removing code from a channel in a footer, that's a somewhat different case, since you're talking about pages with multiple ad blocks and only removing one block. I would CERTAINLY take the ad block you have in the footer off. That's generally not a good location for an ad. Try that, wait a week, consider other options.
Don't do too much at once.
| 8:49 pm on Nov 9, 2005 (gmt 0)|
I'll try removing the footer ad and report in a week or so how the change affected.
| 6:44 pm on Nov 15, 2005 (gmt 0)|
Well it's been about a week now. I removed the footer ad which had poor ctr.
I also did some tweaks on the ad colours and switched placement and units on some ads.
By removing the ad overall clicks of course dropped. My tweaks brought overall ctr up slightly (+0,2 - 0,3%), but the clicks remain less than what I had before.
The result? +30% growth in ecpm and going higher by the day. Overall revenue is up about 1/3.
What does it mean? Is it my tweaks or the removal of the footer ad? Hard to say. I'd assume the second option though.
| 7:57 pm on Nov 15, 2005 (gmt 0)|
Thanks for reporting back, Thez. Less is sometimes more....
| 7:59 pm on Nov 15, 2005 (gmt 0)|
All the advice to the effect that less ad blocks can mean moe income is just plain wrong. You'll just earn less and less by removing ads from "low performing channels"
| 8:29 pm on Nov 15, 2005 (gmt 0)|
I suggest you read up on the "smart pricing" threads here, that's the issue we were testing. My case quite well proved that removing low performance channels resulted in higher ecpm per page and thus a lot higher overall revenue. Like I wrote, I'm getting 1/3 more total revenue with less clicks since I removed the footer ad.
| 9:00 pm on Nov 15, 2005 (gmt 0)|
osewa77, he is just plain right. Thez removed an ad block from the footer of multiple pages, and earnings went up significantly. Since he made some tweaks at the same time, maybe you don't accept that example.
Here's another, somewhat different. I have no more than one adblock on most pages of my site. Last year, I had a channel that was contributing 10% of impressions but had very few clicks. I removed the code. Fewer impressions for the site as a whole, same number of clicks, so a better CTR--and earnings went up by about 20%. There's nothing else that can account for it. Similar experiences on two other occasions.
I've read the AdSense Blog and JenStar's posts on Smart Pricing, and unless I missed something, AdSense has never said that CTR isn't a factor in Smart Pricing if they don't have conversion data. The closest they've come is to say that CTR isn't related to "advertiser ROI"--which is not the same thing at all.
It's possible that CTR is correlating with some OTHER factor that is included in Smart Pricing. I don't know. Until I see it proved otherwise, I am using channels to identify pages and areas where I either need to improve CTR or take off AdSense code.....
| 9:59 pm on Nov 15, 2005 (gmt 0)|
> I've read the AdSense Blog and JenStar's posts on Smart Pricing, and unless I missed something, AdSense has never said that CTR isn't a factor in Smart Pricing
I know that nothing is known for sure about this (hmm that sounds like a Donald Rumsfeld quote) but using CTR as a factor in a pricing algorithm is not a bad idea as it will send a signal to publishers to improve poorly performing ad blocks, something that would be good for the CTA programme in terms of revenue. Whether Google does this with SmartPricing?
| 10:57 am on Nov 16, 2005 (gmt 0)|
Yesterday I had +70% revenue compared to same day last week with about same clicks. Almost double!
How much more proof does one need :)
| 11:17 am on Nov 16, 2005 (gmt 0)|
>Yesterday I had +70% revenue compared to same day last week with about same clicks.
>How much more proof does one need :)
Thez, what you are doing and this thread are very useful, and I may end up making more money as a result :-). But I do still have some questions before considering your experiment as 'proof'.
The 70% increase, in itself, is not particularly meaningful because it depends on what the "standard deviation" is of your site. If your standard deviation is, in effect, 10% of your average earnings, then your proof is very compelling; if it is 100% of average earnings then the 70% change could simply be a coincidence.
I described in [webmasterworld.com...] how to calculate standard deviation. To 'finalise' your proof, would it be possible for you to work out what the standard deviation is for earnings on your site, and say what it is as a percentage of average earnings. This would be greatly appreciated, thanks, as it will put the 70% change in perspective, showing whether it is more likely a coincidence or a result of your removing low performing channels.
Secondly, when you removed "low performing channels", exactly what did you mean by "low performing". My limited understanding of smart pricing suggests that problematic channels would be ones with a relatively low EPC. Although low CTR or eCPM can often be associated with a low EPC, they do not necessarily mean the page is poor performing. This is because you might be getting a very small number of clicks, but if they all convert then they bump up your smartpricing performance.
So, what were the exact criteria you used to decide which channels were low performers?
| 11:18 am on Nov 16, 2005 (gmt 0)|
Removing ads (or showing fewer ads, to reduce page-load time) on pages with few clicks seemed to reduce income for me. *I did not measure revenue per click or per page.*
I may try a more subtle scheme, but my experience suggests that crude measures may well not work.
| 11:42 am on Nov 16, 2005 (gmt 0)|
I moved the low eCPM pages to YPN, adsense eCPM increased but that was due to CTR, EPC stayed the same. =(
| 11:57 am on Nov 16, 2005 (gmt 0)|
I agree with hunderdown. I figured this out many months ago. I removed ads from pages that get a lot of traffic, but low CTR and I saw my EPC and bottom-line earnings increase dramatically with the same amount of traffic to the other pages. I believe there is a threshold for CTR both low and high that helps determine where "smart pricing" comes into effect. Too high and you are not building enough value into the click. Too low and you are only causing impressions to take place....
Of course this is all conjecture on my part, but removing ads from low performing pages is something that does work.
| 12:26 pm on Nov 16, 2005 (gmt 0)|
I have a site about widgets which has recently matured and is getting much more traffic. It consists of a 4 ad Leaderboard under my top nav, the content and another 4 ad Leaderboard just above my footer. I have been getting more clicks and the bottom line earnings have increased, but I have noticed that CPC is lower than before.
So if what you are saying is true, if I remove the bottom one, I will probably get less clicks but the clicks I do get should pay more per click?
What if I change the Leaderboards to 2 ad banner sized... that would also cut the number of ads shown by half. Would that have the same effect?
| 12:47 pm on Nov 16, 2005 (gmt 0)|
I don't have an excel so I can't really calculate the deviation like you suggested. But from google adv. reports, I can see the difference for averages.
Compared to 1. Nov - 9. Nov before I removed the footer, and to 10. Nov - 15. Nov there is:
- 47% increase in eCPM
- 27% increase in average daily revenue
From last 3 days average revenue is 45% higher compared to last week. Clicks have increased about 10%.
The footer ad performed really badly, only 0,2% ctr and 1/10th of my overall ecpm.
| 12:48 pm on Nov 16, 2005 (gmt 0)|
"So, what were the exact criteria you used to decide which channels were low performers?"
Just the low performance in ctr and ecpm.
My number of daily clicks is way over 1,000
| 1:00 pm on Nov 16, 2005 (gmt 0)|
Thez, thanks. With that number of clicks the revenue increase is likely to be significant.
I think I'll try removing some low CTR and low eCPM pages and see what happens. Even if we lose revenue as a result, we're not going to lose a lot.
I'll start with ones that are also low EPC pages, because I'm not sure about the impact of low-CTR/low-eCPM/high-EPC pages.
| 3:28 pm on Nov 16, 2005 (gmt 0)|
mrSEman, have you been tracking the performance of the different ad blocks with channels? If you haven't, you should. Thez figured out that a block positioned at the very bottom of the page was getting virtually no clicks.
Trying different size adblocks is a somewhat separate issue and perhaps worth an experiment....
| 6:17 pm on Nov 16, 2005 (gmt 0)|
I track all of my ads via channels and always remove poorly performing ones or replace them with something else. With channels that are doing "so-so" I test different units and placements until I reach a good result. I can always squeeze a few extra dollars per ad by doing this.
It's the sum in the end of the month that counts, not what you make in a week or the last two days.
| 6:49 pm on Nov 16, 2005 (gmt 0)|
|It's the sum in the end of the month that counts, not what you make in a week or the last two days. |
Indeed. And there are so many things that affect performance. I've seen pages on my site shift from being one of the top five to having minimal earnings, or vice versa, just because I did a site redesign and the path to that page changes....
The big lesson is to analyze what's happening on your site(s) and keep trying to find ways do to better. This is just one aspect of that.
| 7:18 pm on Nov 18, 2005 (gmt 0)|
Well my giddy aunt.
Earlier this week I removed ads from pages reporting low CTR, eCPM and EPC. Recorded page impressions and clicks fell accordingly. Earnings dropped a bit, but not out of the normal range.
Then, at the end of Thursday the SmartPrice algo was updated on my account. When I come home today earnings are about at least 30% higher than my best expectations - well above the normal range for a Friday in mid/late November.
One swallow does not make a summer, of course. I'll need to do some proper stats in a few days time. And we have had Freak Fridays before. But so far so good.
| 3:20 am on Nov 20, 2005 (gmt 0)|
My earnings have also suddenly doubled this week.
I have changed nothing.
I'm pretty sure google said CTR has no relation on smart pricing.
However, I imagine the reason things are looking better, is that you have less ad real-estate to offer, and hence are only showing the more valuable ads.
Chances are you are showing better ads on average, that are better targetted and more likly to convert. So the conversion rate is better, hence you start getting paid more for the ads.
I don't think this is any kind of google consipiracy, just the nauture of statistics.
Just my thoughts.
| 5:20 am on Nov 20, 2005 (gmt 0)|
|I'm pretty sure google said CTR has no relation on smart pricing. |
Nope. They said that "CTR has no relation to advertiser ROI." That is the most they've said about the CTR issue, unless I've missed something.
But I don't think I have. If you can point to a less ambiguous statement by Google, I'd like to hear it.
Until someone does, I'll continue to believe that Google has not ruled out CTR as a factor in Smart Pricing.
Also, 21_blue took the ads off low-performing pages. He did not remove one of two or three adblocks off of a page. So what he did did NOT make only the top-value ads show.
If we are going to get anywhere discussing this, we've got to read what other AdSensers post carefully, and read that Google says carefully. Otherwise misunderstandings can turn into accepted fact,,,,
| 6:42 am on Nov 20, 2005 (gmt 0)|
So what % of your revenue did you guys cut out? My footer ad block represents 2-3% of my site's revenue.. is that too much to chop? Or too little to keep?
| 2:19 pm on Nov 20, 2005 (gmt 0)|
>So what % of your revenue did you guys cut out? My footer ad block represents 2-3%
>of my site's revenue.. is that too much to chop? Or too little to keep?
The pages I removed represented about 20% of page impressions and 5% of revenue.
Freaky Friday was followed by Super Saturday and Sunday is starting to look Special as well. Visitors have been down in the last week or two, but the increase in EPC has made up for this in the last two and a half days. But two and a half days do not a trend make, so we'll have to see how things pan out.
It will be interesting to see what happens to EPC when the SmartPrice is updated at the end of this coming Thursday If SmartPrice is calculated over the whole week, then as I removed these pages on Tuesday (meaning that last Thursdays EPC change will have been based on a part week) then hopefully EPC will increase again (?).
However, I'm also making change to my primary landing pages to reduce their clickthrough rate and encourage genuine widget-seekers to explore the site more. I anticipate this will also increase EPC, but it's a bit like changing two pegs in the game of mastermind: any further change in EPC won't be attributable to any one factor.
| 5:53 pm on Nov 20, 2005 (gmt 0)|
So here is the question.... what matters, is it raw CTR? Or page CTR?
The only ad block I have that looks like a candidate for removal is the one on my main page in the footer. It gets a ton of impressions, but only a few clicks, and it contributes 2-3% of revenue overall. This block is clearly lowering my overall CTR, however it is improving the CTR on my main page.
If I chop it and it is page CTR that matters then smart pricing should punish me (but only to the tune of 2-3%). If it is raw CTR that matters I should get a big reward.
Unfortunately the mastermind analogy is going to hurt.. we are getting near to Christmas and year end and I am anticipating rising revenue regardless, so it will be hard to know whether it is smart pricing or simply increased competitiveness.
| 6:03 pm on Nov 20, 2005 (gmt 0)|
martingale, in my experience, it is site-wide CTR that matters, because I observed this effect on a site that only had one adblock per page. I took the ads off one set of pages, site-wide CTR went up, and so did EPC--and overall earnings. I've seen this happen three times.
There may also be page CTR factors, I don't know.
Note too that the number of impressions that the removed ads generate is, of course, important. If they generate 10% of impressions then removing them will have much more of an impact than if they only generate 1% of the impressions.
You may just have to try it. I'd suggest trying it for a full week or longer--since Smart Pricing doesn't update dailiy. And try to choose a fairly normal week, which will be a bit tricky this time of year.
| 6:42 pm on Nov 20, 2005 (gmt 0)|
>it is site-wide CTR that matters, because...I took the ads off one set of pages,
>site-wide CTR went up, and so did EPC
Hunderdown, were those pages also low EPC (in comparison with the rest of the site)?
So far I've only been removing ads from pages that are low on everything - CTR, eCPM and EPC. However, I wonder if it is EPC that really counts. I realise that there is an argument that Smart Pricing operates on a site-wide basis. But in the Adwords blurb Google say that the Smart Pricing algo takes into account "many factors", of which the performance of the site is presumably just one.
So, a relatively low EPC page on your site could indicate that the advertisers are getting discounts due to the "other factors". These other factors could, in turn, be knocking down your account Smart Pricing assessment.
If that is correct, it suggests that low CTR/high EPC pages are OK, because the 'other factors' aren't coming in to play.
| This 40 message thread spans 2 pages: 40 (  2 ) > > |