Forum Moderators: martinibuster
I understand that G allows 200 channels. How do you manage and interpret that number of channels without spending massive amounts of time doing so?
Thanks
If you want detail you are better off running a separate tracking script on large, and/or dynamic sites.
Freq---
What I have found is that I can put most pages in one of two "default" channels and keep only about 15 active channels, which track my home page, five or six very active pages or small groups of pages, and several pages I am working on, trying to improve.
Freq---
For existing/old pages, we decided which pages were not 'significant' to our income, and don't have a URL channel for them. In general, most of our revenue comes from the pages we track.
Tracking on an ad unit basis rather than a URL basis might provide you with more useful optimization information as well. You can figure out which placements, colors, ad formats, etc. work best on your site and experiment.
-ASA
Why you recomend this? especialy by topic?
With absolut the same page layout, ad placement, colors, different themes can have very different CTR.
My main site has several theme oriented sub domains.
So my channels are like
theme1.my-domain.com ..............usual a 300x250 in the navigation
theme1.my-domain.com/.AdLink ......a 120x90 AdLink in the navigation
theme1.my-domain.com/.Content .....if space, an aditional ad in the content
The naming convention with /. brings this channels first in the list for all channels and URL channels of this subdomain
In addition to this, I use URL channels to watch new reportages
theme1.my-domain.com/2006-reportage-name
Because there are only 200 channels, I will have soon to give up URL channels to watch older less earing reportages.
What's known is this: Smart pricing IS applied account-wide. Google has said so. There may also be a site-wide factor, though Google does not acknowledge this. And there may even be a page factor, but that's speculative.
hope this helps. everything is done with php functions, so it's easy to use. every page on the site is labeled into a group or most popular with tags that match adsense channels. takes time to setup, but once it's done, it's very easy to use.
The major categories have 2 or 3 channels each, depending on how many ads are on the page, most have just two.
This works pretty well.
And you can quite easily measure performance for ad sizes and locations, etc, doing this.
I use my web logs to tell me the top pages, and then experiment with ad placement on them. I pretty much know what pages get top hits, and where ads will or won't work as a result.
The problem with ads on every page is that many of them may get little, or no clicks but add to the overall page impressions. Having a low overall ctr may affect smart pricing. We don't know if google uses page or site or account wide ctr in smart pricing - in all probability all of them to some extent or other.
Therefore it's my feeling that having ads only on pages they will work is going to
a) reduce ad blindness on pages where the ads do work
b) optimise overall ctr for the site (which will feed back to better epc via smart pricing over time)
c) make the whole shooting match manageable, therefore easier to spot trends and make, and measure adjustments that will hopefully increase income.
Having a low overall ctr may affect smart pricing.
I would tend to say instead: "Having a low overall CTR may indicate poorly qualified traffic for the products being displayed; poorly qualified traffic will affect smart pricing, either now or in the future (as Google gets better at making it reflect actual return on investment for the advertiser).
Look at your low CTR pages and the ads displayed therein. If common sense says that content isn't going to attract motivated buyers for those products (e.g., page is about where to find the cheapest gas, and ads are for Ferraris), then consider removing the AdSense ads from those pages, perhaps replacing them with links to content that displays ads more suited to that particular audience.
It is, of course, possible to have low CTR pages in the midst of well-qualified traffic. Some products inherently have a low CTR (very expensive, very infrequently purchased, etc.). As usual, the best answer is found by testing and retesting.
Having a low overall ctr may affect smart pricing.
On the other hand, having a high overall CTR may affect smart pricing, too, if's the result of over-optimization (e.g., blended-in ads disguised as navigation links) that leads to frequent clicks but poor conversions.
For what it's worth, I use AdSense on sections of my 5,000+-page site that have extremely low CTR and ePCM, but I haven't seen any evidence that smart pricing is hurting me--probably because the few people who actually do click on ads in those low-CTR directories (such as photo galleries) are likely to be interested in what the advertiser is selling.
On the other hand, having a high overall CTR may affect smart pricing, too
Yes, that's been my experience. But not as a result of over-optimisation, but of not blocking MFA's. I believe MFA's have a better ctr than other ads in part because the people that place them are actually quite good at writing eye catching copy! I would say that regular advertisers (certainly in my sector) are lacking imagination when it comes to writing eye catching ads :(
Removed the MFA's, ctr went down but everything else went up and I earn considerably more as a result.
It appears (based on my stats) that smart pricing seems to like a stable ctr over time, and in my case that seems to be about my age / 10. I don't try and achieve any set figure - that's simply where the CTR generally happens to end up naturally.
I believe MFA's have a better ctr
I belive the opposit.
A visitor real interested in a shopping tour can click many ads. Ads on the page, ads in the AdLinks.
MFAs are real stoppers.
So a visitor on a shopping tour makes on a web site with a well maintained URL blocking list maybe 5 clicks, on a web site without URL blocking maybe only 2 clicks.
(EFV:)On the other hand, having a high overall CTR may affect smart pricing, too
(David_UK:)Yes, that's been my experience.
That's an interesting idea, so I've taken a look at my data this morning, comparing EPC with CTR across all our (200 URL) channels for the last 5 months, putting it into a trend curve in Excel (polynomial, up to order 6). The hypothesis is that if too high or too low a CTR hits smartpricing, then the shape of the trend curve should show a peak (eg: at around 10% CTR, say). However, the curve is flat. I then put the data into my stats package, to do a polynomial regression, and the R2 statistic was .01 (the closer to 1.0 the better the fit with the curve).
So, accepting the undoubted caveat that this is only my data, there wouldn't seem to be statistical evidence of the type of relationship between EPC and CTR being discussed.
That some people might see a trend that disagrees with this in their data can have one of two causes:
No doubt the debate about CTR and EPC will rage on, and I accept that it is possible I'm wrong and there is a relationship. But at the moment the evidence suggests that the explanation is more likely to lie in point 2 than point 1.
