| 6:51 am on Oct 26, 2005 (gmt 0)|
It does not surprise me at all.
Google are only wanting to please their shareholders -- not us stupid publishers.
| 7:03 am on Oct 26, 2005 (gmt 0)|
What a load.
I can't wait until YPN fixes their targeting and MSN rolls out.
| 7:05 am on Oct 26, 2005 (gmt 0)|
I believe I will try my middle site on YPN and compare the two of them.
| 7:15 am on Oct 26, 2005 (gmt 0)|
Interesting that Jenstar mentions that poor statistics from AS prevent publishers from optimizing (improving) the conversion rate. I think one of the major tasks for the AS team is to improve statistics for publishers very soon. Otherwise we all will keep on fishing in very dark waters.
[edited by: mzanzig at 7:16 am (utc) on Oct. 26, 2005]
| 7:16 am on Oct 26, 2005 (gmt 0)|
Oh boy, this thread is going to be here for a while.
| 7:41 am on Oct 26, 2005 (gmt 0)|
I reported this activity a week ago at DP, this article confirmed it!
"Has anyone noticed such behavior? This happened 2 TIMES, still a coincidence?
On 10-05-2005, I added a new channel, CPC immediately dropped the next day by HALF! Two days later, I removed the new channel and miraculously it returned to normal!
Same thing happened 2 days ago, only I haven't removed the channel yet because the site is now live!
Is this a coincidence? But happened 2 times? I mean it really doesn't make sense, I make new sites to earn more revenue, not lower the revenue for my other sites. I also don't see how this would benefit google. Has anyone noticed this behavior before when adding new channels?"
| 7:59 am on Oct 26, 2005 (gmt 0)|
Why can't they tell us everything that has been delivered from their servers.
|I think one of the major tasks for the AS team is to improve statistics for publishers very soon. |
The directory example.com/example/example has the following statistics. You can sort by any ad parameter(i.e background color, size, contextual only, image only, site-targetted), ... you get the idea.
Unlike the size of the internet as a whole today, they have shown they can manipulate the data on this scale. Why not do it? They didn't buy urchin just to offer it as a product, did they?
| 8:03 am on Oct 26, 2005 (gmt 0)|
Does anyone here fear, as I do, that these "glitches" are nothing more than a convenient excuse to increase earnings without providing accountability? There may be good reasons to be so secretive, but one of them are to keep up those ridiculous earnings like they just got. They've got to get those earnings off someone's back.
| 8:18 am on Oct 26, 2005 (gmt 0)|
I have sites earning dollars a click and other earning cents a click. And I have one site earning cents a click where page impressions are high and click throughs very low.
Are you saying that if I took Adsense off that one site I would earn more money? And that my other sites are somehow handicapped by that one site? Surely not.
| 8:24 am on Oct 26, 2005 (gmt 0)|
I feel it is on an per account basis. I thought my thinking was ridiculous at first but it seems like a Google employee confirmed it!
| 8:27 am on Oct 26, 2005 (gmt 0)|
No that's not how you can tell. It has to do with conversion stats related to the advertisers with ads on your site.
Maybe somehow ecpm might give you some kind of idea. But probably not.
You know what a funny aside is? If they're screwing this up for publishers, that should mean smart pricing is as dumb for advertisers. BOTH sides should be pissed.
| 8:33 am on Oct 26, 2005 (gmt 0)|
So it's not something I can influence then - at least not by dropping what I perceive as being my worst performing site.
| 8:35 am on Oct 26, 2005 (gmt 0)|
You can influence it. However they aren't going to give you the stats to do so. You'd have to guess. The best way to be sure is trial and error. One week at a time or so (going by Jen's blog post).
Or you could go by standard assumptions. Certain types of sites convert better than others. Dump the ones that convert worse.
| 8:41 am on Oct 26, 2005 (gmt 0)|
Clark - I think I get it now. Thanks for taking the time to explain this.
By conversions do you mean people who click and then buy or act in a way the advertiser requires? And that any Smart Pricing is based on these conversion rates?
How does Google know the conversion rates? Surely once someone clicks through they can no longer track their behaviour? Or are they guessing too?
| 8:57 am on Oct 26, 2005 (gmt 0)|
If adsense is important to you, do yourself a huge favor. Open an adwords account. Take a small budget and use it to generate some traffic for yourself. You will learn a lot about half of what adsense is all about. Adwords and Adsense go hand in hand. I did it myself for that reason only and the experience was well worth it.
They track ROI by giving adwords advertisers a special code to place on a page which denotes that the buyer "bought"...however you want to define that.
| 9:08 am on Oct 26, 2005 (gmt 0)|
That's the best bit of advice I have heard this year. Many thanks. Adsense is increasingly important to my business and it should have been common sense for me to realise that it would cost very little to see it from the Adwords sie of the fence. Thanks for opening my eyes!
I'm going to sign up now.
| 3:52 pm on Oct 26, 2005 (gmt 0)|
|How...can Google be so stupid as to count Smart Pricing on an account basis rather than at worst domain basis. |
Threads were started on this issue yesterday here [webmasterworld.com] and here [webmasterworld.com].
| 4:08 pm on Oct 26, 2005 (gmt 0)|
Jenstar, thanks for your info as always. However, about the following detail, there is maybe a misunderstanding:
|(...) that team member disclosed (...) Smart pricing affects an entire account. It is not on a per page or per site basis. |
On the WW post, the second phrase ("It is not on a per page or per site basis") isn't mentioned. Probably smart pricing affects an entire account, but this doesn't mean that it is not affecting the specific page and site as well.
In fact, observing stats for a long time, it seems that smart pricing affects mainly specific clicks, then also -in decreasing intensity- the page, then the site, and then the account.
I'm not completely sure, but removing ads from low-performing pages or sites with very low eCPM ($0, or not far from $0) appears to be a good practice, because probably it increases the average site/account conversion (even if eCPM is not an exact conversion measure, only indirectly related).
That is, unless you start getting successful CPM ads on those previously low-performing pages. But site-targeted CPM ads still need improvements, according to WW members.
And indeed, smart pricing seems to act with a delay of usually some days or more.
This is just my impression, and I can be very wrong, of course.
| 4:37 pm on Oct 26, 2005 (gmt 0)|
Are we deliberatelly being fed misleading information by Google to keep us all occupied wondering what is going on?
The general consensus of opinion is that the larger the site, the higher the quantity of visitors and page impressions...then CTR and eCPM both drop. It's been posted around enough times.
I see that on my two biggest sites in comparison to the satellite sites which have much higher CTRs and eCPMs.
If I were to drop the pages/sites with the lowest CTR and eCPM then my main money earners would be the first for the chop...or am I being a bit too simplistic?
After all, I'm only a country bumpkin who still hasn't gotten used to not counting in bushels:-))
| 4:53 pm on Oct 26, 2005 (gmt 0)|
The consequences of this idea are potentially so profound that I need to say this one more time:
Even if there is a single "smart pricing" number for an entire account, if it is calculated using a weighted average based on gross revenues then in the long run it will mathematically work out the same as if every page had its own number.
In other words, just because some phone rep from Google who is cold-calling potential new accounts happened to remark that smart pricing uses a single number per account, that doesn't necessarily imply that low-conversion sites can poison high-conversion ones, and it doesn't necessarily mean that this is the whole truth or that the assertion is correct at all.
| 5:07 pm on Oct 26, 2005 (gmt 0)|
|In other words, just because some phone rep from Google who is cold-calling potential new accounts happened to remark that smart pricing uses a single number per account, that doesn't necessarily imply that low-conversion sites can poison high-conversion ones, and it doesn't necessarily mean that this is the whole truth or that the assertion is correct at all. |
Rushing headlong into dropping Adsense from sites or even pages on the basis of this kind of idea might not be in a publishers best interest.
| 5:09 pm on Oct 26, 2005 (gmt 0)|
these valuable three adsense smart pricing threads running today enlighten as well as depress me.
- publishers have no chance in measuring ad conversion because google doesn't provide the needed metrics. ctr and ecpm in this context can be regarded as irrelevant or even misleading
- because we have no clue, we only can guess our underperfoming sites, which makes it a costly and tedious trial and error experience, if we experiment with dropping them
- we get penalised account-wide for things we have no influence on, especially incompetent advertisers, who can't manage our qualified traffic
| 5:16 pm on Oct 26, 2005 (gmt 0)|
Before posting I did look for other threads on the issue and didn't find them. One of them I did see later.
I hear you. But something is fishy in Denmark. I've been seeing it and saying it for quite a while. And this makes sense as the cause.
As for the math, I don't quite get how
computes? Are you saying that the high paying page losing ground will be offset by increasing the rate of the low paying page? Does it work out mathematically to be the same thing?
|Even if there is a single "smart pricing" number for an entire account, if it is calculated using a weighted average based on gross revenues then in the long run it will mathematically work out the same as if every page had its own number. |
| 5:23 pm on Oct 26, 2005 (gmt 0)|
Clark: Yes, that's exactly what I'm saying. It CAN be calculated so that one page earns less, the other page earns proportionately more, and it makes no actual difference to the bottom line in the long run.
But thinking about this further, the above works from the publisher's point of view, but from the advertiser's POV it makes no sense at all.
To use a simple example, say you have 2 equal sites, one of which is rocking while the other has a very low conversion rate. In ANY system where smart pricing is determined by the overall publisher account only, advertisers on the low-conversion site will get only half the discount they deserve, while advertisers on the high-conversion site will get a substantial discount where none is really justified.
I find it hard to believe Google would implement something so obviously flawed, that so obviously fails to do what it is intended to do. IMO this Google rep probably does not know or did not reveal the complex calculations that must go into making smart pricing work.
| 5:37 pm on Oct 26, 2005 (gmt 0)|
I have a solution... switch all your lowest CTR pages to YPN. I've done it with hundreds of pages now and my bottom line has gone up (AS + YPN earnings vs. AS only for those same pages). More importantly, on those pages where I still have AS, EPC has gone up.
I'm sure other variables come into play, but the bottom line is this: I moved low CTR pages to YPN, where I get a much higher EPC and roughly the same CTR. On the pages where I still run AS, EPC also went up. Win-win.
| 5:48 pm on Oct 26, 2005 (gmt 0)|
In Jenstar's blog entry about smart pricing, she refers to the post here at Webmasterworld as the source of her info, so many of these other posts at WebmasterWorld citing Jenstar's blog as validation of the first post are based on circular logic.
Personally, I'm not so sure about the premise in the original post. It does not ring true with my personal testing with smart pricing on my own sites. I have changed my income dramatically on one site many times with layout testing and have not noticed any significant changes in income on my other sites.
| 6:48 pm on Oct 26, 2005 (gmt 0)|
If you want come circular logic try this on for size...
Jen probably knows a lot of insider information that she's been told 'off the record', or sworn to secrecy via NDA, etc. However, once someone else breaks a story and it's already public information she's off the hook and can reference their story before expounding upon it further.
Based on what I've seen Jen's a pretty decent fact checker and I'm sure she wouldn't jump on a story about smart pricing unless she knew it had some truth behind it.
Therefore, at this time I'm going to assume the story has some level of credibility.
| 6:52 pm on Oct 26, 2005 (gmt 0)|
|OptiRex wrote: |
If I were to drop the pages/sites with the lowest CTR and eCPM (...)
Only if the eCPM is really, extremely, low (such as $0 or $0.x). Therefore, in the worst case of a failed test the lost revenue would be probably small. I've not tested all the possibilities enough, but there seems to be some positive effect on earnings only in this case. And with a delay of some days, not immediately.
A theory on this effect: If I'm not wrong, for AdWords advertisers, low-performing ads (< 0.5% CTR, I think) were dropped after, say, 1000 impressions (I don't know exactly). Now the system has changed and it's less simple, involving costs, etc., but perhaps this is related in some indirect way to the low-performing pages effect.
It's possible that removing ads from extremely low performing pages might be beneficial in some measure for the whole AdSense program (advertisers, publishers and Google). As said on another post, unless that a substitution -such as the new CPM ads- starts working better on those low performing pages.
| 6:57 pm on Oct 26, 2005 (gmt 0)|
I agree Juan - but only if we are talking in terms of conversions. CTR should be a non-factor.
And with conversion tracking being flawed (advertisers don't HAVE to report conversions) - the entire system is flawed.
| This 218 message thread spans 8 pages: 218 (  2 3 4 5 6 7 8 ) > > |