Forum Moderators: martinibuster
Anyhow, because my last 10 months' stats were highly changing, it helped much in this experiment.
Here is the bottom line: in months when my page impressions were 5 times as much as other months: the average monthly EPC was the same. When my CTR was 0.2, and now when it went up to 4.2, the average monthly EPC is still the same.
Time to put back adsense to my low CTR pages/viewers.
From now on, I'll make way for site-targetted and the new onsite CPM ads.
What does change is the CTR and when I check it's almost always because of poor targeting by Google and an influx of MFA ads. I just block the MFA's & MFO/Y's and within 6 hours the CTR starts rising again.
This could be coincidental as AS's own algos may adjust due to poor historical CTR.
Anyone want to hazard a guess as to how real time their system is?
I started to optimize Febrauary 2005.
No change in EPC, but dramatic higher CTR.
Big G prides itself on its algos, so i'm sure they take many factors into consideration when calculating your smart price for a click. Off the top of my head: Time of day, geo-location of visitor, time of year, how well the ad performs on the Advertisers side, what color level of threat the nation is at, the price of oil, NASDAQ, etc. Trying to figure all of that out is just too much. The only thing we need to figure out is who pays more AS, YPN, MSN, etc.
How do you block MFA's & MFO.... I would be interested to test your theory on my site... thanks
This is interesting, but it doesn't settle the debate. If, for example, the advertisers on sallam's site were reporting conversion data, then Google would rely on that. Conversion data obviously trumps the various substitutes for it that Google might use when they don't have it.
There also may have been other factors influencing EPC, so that any Smart Pricing changes were obscured. Or a 4.2% CTR might be below the point at which CTR starts to be used as a factor...
The test, for me, remains this: take the AdSense code off some really poorly performing pages. If CTR isn't a factor in smart pricing, then earnings should stay the same. If it is, earnings will go up. This has happened three times on my site. It may not work on every site. Quite a few things go into smart pricing, and many of them are out of our control.
And re the question of why low CTR might be considered a sign of poor conversion: low CTR could indicate that the site content, the ads, and the visitors are not meshing well. If visitors are generally not seeing the ads that interest them, and thus aren't clicking, it isn't unreasonable to assume that they also aren't as interested in them when they DO as visitors to a higher CTR site. Google could even test that, and perhaps they have. All they have said about this issue is "CTR is not related to advertiser ROI" and that leaves some room for them to actually use CTR in some other way.....
The conclusion that I am coming to is that the Smart Pricing algorithm is more complicated than I can figure out. On my site, it does seem to be a bad idea to keep ads on low-CTR pages. But building content and growing traffic are more important, of course.
I think some people are wasting too much valuable time analysing instead of creating new content
Every merchant calculates his expenses and his income.
Creating content costs time and money.
Last round trip for new content was 1200km.
Entrance on fairs is for me free, because I have a press card.
So I have to calculate expenses against possible wins by writing new articles.
How is the income by similar reports?
How is the general development?
More and more reports are interesting for me.
This year, I have only to risk the expenses of car travels.
Next year maybe other continents with far higher traveling expenses.
So all the statistics make the decision to risk some thousand $ for a long journey.
You could have a high CTR, but those ads are cheap, and your EPC will still be low. However you could have a low CTR, but the few ads that were clicked, were expensive and paid very well, your EPC is high.
Like someone said, focusing on content is a best bet.
How do you block MFA's & MFO.... I would be interested to test your theory on my site... thanks
Use the preview tool and check each low preforming page which you can tell from the channels in reports that you have already created. You did do this didn't you? Makes things much easier to track.
I always use auto, Canada, US & UK from the drop down list at the top right for my English language pages. If you have other languages then use the country for that language.
Go through the ads that show up and see if there are any you don't recognise immediately. Click the link (only inside the preview tool window, this is allowed) and have a look at the landing page. If it's no content and just AS or Over links then put a check mark on that ad block inside the preview tool.
After going through the rest of the ads and putting check marks on those you don't want click the "show selected URL's" link in the top left. You will then be given the URLs for those ads. High lite and copy them.
Open AS and go to "AdSense for content" and click the "Competitive Ad Filter" folder. Click EDIT and paste the URLs from the Preview Tool into the window. Now delete all www. and anything after the .com. Only leave the root domain.
In other words if the URL from the Preview Tool is www. widget.example.com/widget/phonywebsite.asp only leave example.com. Click save and in a few hours (usually about 4 or 5) those sites will be eliminated from appearing on your site.
Thez, I don't know if you've also seen elsewhere me suggesting that it isn't CTR we should focus on but "relative EPC". My reasoning is that where EPCs are consistently lower than the norm for your site they probably indicate discounts for the advertisers and therefore poor conversion rates.
Were the pages you excised also low EPC?
a)where your traffic comes from
b)how pre qualified it is
c)Search engine "relevance" (more traffic isnt always better! I now get about 25 percent more traffic after the latest jagger - click through fell by the same)
d)ad colours
e)ad positioning
f)number of blocks
g)content (pos review/piss take of product?)
h)ad copy
i)ad relevance
And more.
Now smart pricing SHOULD give the advertiser the same conversion / sale from all these possible scenarios.
He should get x number of sales per x dollars paid.
Regardless of if you have a genuine original useful content site, proffesionally written reviewing products, or a spammy MFA landing page with a mass of hidden/disguised ads that you click to escape from.
The idea is that it takes all of this into account and should make things fair for everyone.
So hiding ads, adding 3 ad blocks, and so on as advised by adsense will NOT increase income in the long term because your epc will fall. It has to because your clicks are less "meaningful" and will not likely convert. Their value will drop sooner or later.
The only one to gain is google with almost every page on the entire net covered in "ads by google" and has more real estate to put more cheaper ads on due to less competition for ad space!
Seems to me we should all have one single ad block, well written pages, very qualified traffic, etc. Then we will earn JUST THE SAME without having our pages looking spammy! And drive the price of ad space up!
I dont think the smart pricing algo is at all simple!
It HAS to take all of this into account. Does it though!
This could be spot on. So far, since making my changes to try and up my SmartPrice last week, although average EPC turned upwards on Friday, as did earnings, the "earnings per visitor" has stayed roughly the same. That is, if I'd done nothing, my earnings would probably have gone up because my traffic went up.
