Forum Moderators: martinibuster
For example, a click on an ad for digital cameras on a web page about photography tips may be worth less than a click on the same ad appearing next to a review of digital cameras.
[edited by: markus007 at 8:08 pm (utc) on April 1, 2004]
...a walking shadow,
a poor player,
that struts and frets his hour upon the stage,
and then is heard no more.
It is a tale told by an idiot,
full of sound and fury,
signifying nothing.
Shakespeare
The maturation of certain algos parallels human development and maturation. Therein lies the rub.
My teenage children hold forth that they know of what they speak, yet they know little by experience. The education of even very bright youth appears inextricably tied to learning by trial and error.
The AdSense algo, my children, my life....it's like deja vu all over again.
Google is making incorrect assumptions about the value of "information" clicks in a category where the record shows that most users are looking for advice on where and how to spend their money.
I agree it would be nice to know how this situation factors in.
Would it typically be 'big ticket' items that requires additional human sales contact?
Perhaps they have a 'big ticket' function in their algo?
I can see them not wanting to comment in this thread about earnings (they would be bombarded), but this in particular is something I would like seeing official comment on.
Many weeks ago I was having some sort of problem with AdSense reporting. Out of frustration I re-applied to Ad Sonar and the next day I was accepted.
The payout with Ad Sonar was so low I was only using them on a very few number of pages. In this thread someone mentioned that, with the AdSense changes, Ad Sonar was working better for them now. Now with the Channel data AdSense allows, possibly my impression of the low payout was more topic related than anything. It will just take more time to figure things out.
I was considering reinvestigating my relationship with Ad Sonar and now, out of the blue, I have received an email from Ad Sonar saying they want me to let them know when it would be convenient for them to call me so to discuss how to improve my site's CTR.
At minimum it seems I will end up giving Ad Sonar some more traffic. AdSense is making it clear that I need a back up advertising network. Ad Sonar actually seems to be trying to win my favor and business.
YES! THANK YOU!
In my site's niche you dont just "click and buy", you make contact with the service provider and send out a snailmail application... learn more... talk to them over phone and then using a snailmail check!
NO WONDER I'M GETTING LOW EPC... they think all my clicks are useless...
Come on google! Not every business model involved "click and buy"...
Some of those affiliate sales are immediate (people clicking on a hotel link and placing a reservation), while others may require correspondence with a human being. Does Google have any way of knowing when sales result from the latter type of inquiry?YES! THANK YOU!
In my site's niche you dont just "click and buy", you make contact with the service provider and send out a snailmail application... learn more... talk to them over phone and then using a snailmail check!
NO WONDER I'M GETTING LOW EPC... they think all my clicks are useless...
Come on google! Not every business model involved "click and buy"...
Right. But, clicks are the only way Google and the publishers get paid.
This is why I think that Google and Yahoo is heading towards:
1) offer contextual ads only (Yahoo does this now) and
2) offering ads in context across a network on a PPV instead of PPC.
And I think the PPV network is going to look something like what you see on the Personalized Google Search now in Google Labs.
I think it's done on a per-advertiser basis, then data is compared with publisher data. Explanation:
Advertiser A has not implemented conversion tracking (Coca Cola)
Adveriser B did (Web hosting company)
Advertiser C did (e-Widgets-online selling company)
Advertiser D did not (didn't bother)
Forget about A and D, and work with B's and C's conversion data.
Work with B's data first - and first of all, compute B's overall average conversion rate, throughout all publisher sites. Let's say it's 1%.
Now we realize that B has a conversion rate of 2% at publisher1.com. Give publisher1.com a good point.
But B has a conversion rate of 0.5% on publisher2.com. Give publisher2.com a bad point.
Now work with C's conversion data.
C has a 10% average conversion rate on publisher sites.
C has a 10% conversion rate on publisher1.com. Leave publisher1.com alone, don't give any points. C has a conversion rate of 7% on publisher2.com. Give publisher2.com a bad point.
At the end of the day, using conversion data from two advertisers (B and C), we have given 1 good point to publisher1.com and 2 bad points to publisher2.com
Publisher1 will have a better EPC than Publisher2 right now, until we process the conversion data again and see if the situation is different. Publisher2's clicks are less valuable than Publisher1's, since overall, they converted less well.
Now, of course, we only used two advertisers (out of 4), and two publishers. Google might be using all advertisers (who use conversion tracking) and all publishers, to calculate "ecommerceness" of each publisher.
Well, that's just an idea, and I might be completely wrong.
The morale of my story is that the fact that Coca cola won't sell cans on internet doesn't matter. Google just won't use this advertiser's conversion data since there are none. And at any rate, Coke doesn't sell cans on the internet - but that's the case for all publishers who run Coke ads, so it will not (should not) unbalance things.
With the data they apparently get and record, Google has enough of it to calculate the real ROI-ness (I.e. value of clicks and/or leads) of each publisher.
This is good - GREAT - for advertisers, IF the discounted costs are propagated to them. I have an Adwords account and so far the price of clicks didn't change at all (it might change in the future, who knows how much data they need to collect before charging people less - a touchy business decision). It CAN also be good for publishers if, this way, Adsense gains more confidence from the advertisers and they pour more money in it (less defaults, more competition for publishers' inventory, ultimately higher CPC and probably EPC).
Let's hope it plays out good for everyone.
I am now getting less than 5 cents a click. Time to reinvest the inventory elsewhere.
only? i may have is why advertisers not seen lowering of CPC
from where we are looking
1 advertisers paying same
2 publishers getting less
steve
I honestly believe if publishers could see evidence of number 2 they would not be so upset and would see it as a good move for long term viability of adsense
They really should have a seperate bidding system for search and content ads.
The only automated system that can calculate how well ads are performing is called 'the free market'. Advertisers don't need help judging how much to pay they are smart enough to figure it out for themselves. It's just that at the moment they are forced to pay the same for two clearly different systems. This means they are probably paying less for search ads and more for content ads which is what Google is trying to correct.
There is so much more to getting a click through to your site than just an online purchase. I have advertised new community sites when they first went up to get people to join. Many business don't want you to buy straight away. The only thing you can say for certain is that advertisers want you to visit their site after reading what they say in their adverts.
My site has not been particularly affected as all the ads are for products people buy immediately online. I was however planning on building a site about an industry that doesn't operate in this way as the products are very expensive - they are willing to buy expensive ads though (although google won't value them as such).
This has to stop. It makes the game far to complex for advertisers - someone who just wants the visitors to download a form is going to get very cheap ads and those selling directly online are going to subsidize them pay paying more than a unregulated market would allow.
Publisher1 will have a better EPC than Publisher2 right now, until we process the conversion data again and see if the situation is different. Publisher2's clicks are less valuable than Publisher1's, since overall, they converted less well.
They cannot track or even know all conversion models and not everything can be handled by an algorithm.
Take a lead generation example. What happens if Publisher2 sends more motivated leads who tend to call rather than fill out a form. And what if those follow through with conversion better than somebody who does fill out a form. This IS the case with me. I get so many different types of leads, I do track conversions, phone calls are much more motivated buyers that do NOT get tracked. You cannot make all-encompassing statements.
The only automated system that can calculate how well ads are performing is called 'the free market'. Advertisers don't need help judging how much to pay they are smart enough to figure it out for themselves. It's just that at the moment they are forced to pay the same for two clearly different systems.
Not two different systems--more than half a dozen, by my calculation.
You may not think that SERP ads have much in common with "content ads," but what do the following "contnt ads" have in common with each other:
1) Ads on niche or special-interest editorial sites
2) Ads on general news/entertainment/portal sites
3) Ads on forums
4) Ads on e-commerce sites
5) Ads on parked domains
6) Ads on gmail
There's likely to be far greater variation in ROI between, say, an ad on DPReview (a digital photography review site) and a gmail message than there is between an ad on DPReview and a parked domain or a gmail screen. So just having a separate bidding structure for "search" and "content" woudn't make much sense.
At the end of the day, advertisers will want to bid on whatever matches the ROI. Paying the same for forums and ecommerce sites will end up with a bid price somewhere in between the acutal value of a bid for each type.
A simple first step would be divide search from content as they already have it separated by their opt out options. Ideally they would have lots of options and have reviewed all the sites to decide which category they should fit in.
I just don't think they can create a system based on the data they have available that can accurately detect what the bid price should be. It undermines the whole idea of having an auction.
someone who just wants the visitors to download a form is going to get very cheap ads
I believe that this is absolutely not true in the current state of information we have, except in one case - all of this someone's ads by miracle appear only on very badly converting sites. Then, that someone will get cheap ads - but bad conversion, so cost / conversion is likely to remain the SAME.
What matters, most of the time, is cost per conversion.
By the way, conversion tracking allows to track Leads and Sales separately.
I'm not trying to be devil's advocate - those changes divided my own income by a lot - but I'm just trying to figure out the trends.
With the data they apparently get and record, Google has enough of it to calculate the real ROI-ness (I.e. value of clicks and/or leads) of each publisher.
Fair point and post, but what about this scenario:
Publisher has a site about big-fee services. Most ads are for relevant big-fee service companies - no conversion tracking is possible and/or done. A couple of ads are for half-as-relevant "click and buy" products. A few curious clicks on these produce no sales. Google rates that publisher as a poor converter. But in reality he's delivering super leads to those big-ticket companies.
Perhaps Google uses on-page factors to compensate, but...
But there does have to be some sort of discrimination with publshers. Just take two examples:
1. A highly informative information site in a niche field, dozens or hundreds of pages, a great education resource. And then somewhere on it are also adsense links to advertisers, if a person is inclined to buy. No trickery or foolery.
2. A site that comes up under a fairly lucrative keyword, with one main page that is really a trap page. The only way for a visitor to escape is to back out, or click on an adsense link which is the only link that is very visible.
If you were an advertiser, would you be willing to pay the same rates for clicks from both of these sites?
...with one main page that is really a trap page. The only way for a visitor to escape is to back out, or click on an adsense link...
an even better example would be GMail. I really don't see how users could be in a "buying" mood if they are reading mail from friends and family. Users might be clicking on gmail-ads just out of curiousity (e.g. to check out how much the new Lexus a friend bought is going for) or simply because out of boredom ("lets surf the 'net after reading all the email")
At the end of the day, advertisers will want to bid on whatever matches the ROI. Paying the same for forums and ecommerce sites will end up with a bid price somewhere in between the acutal value of a bid for each type.
Sure, and paying the same for search and content will deliver a bid price somewhere between the actual value of a bid for each type. All Google has to do is eliminate the "opt out" option for content ads, and the market will determine what the optimum bid for any keyword or keyphrase should be.
I see some pushing the idea that google will move to a pay per conversion model. Do you really think in the longterm that advertisers will want google to know its sales figures? That is why that whole model is in my opinion flawed in thinking. Google may think that this is the way to go, however privacy concerns will ensure this does not suceed in my opinion.
There's no indication that Google is moving to a pay-per-conversion model; it's just moved to a pay-per-anticipated-conversion model. Therefore knowing the sales figures or conversion rates of all advertisers isn't necessary: Google merely needs to have a large enough statistical sample to devise a formula.
a) my adsense revenue has too gone down, and
b) as an Adwords advertiser I am not very impressed with the conversion tracking they offer. I got far more conversions from adwords than their system indicated and the cost per conversion figure they indicated for me was WAY out. IMO they better improve that before they go making wholesale changes based on that data....
If it is based on accurate data I can see how this would be an excellent system to base payouts on.
They had to develop this information somehow, somewhere or someplace...else the alternative it is smoke and mirrors behind a revenue shift.
I am still waiting to hear how the adwords users have seen a dramatic shift downwards starting April 1, indicating that this is a passthrough event not a money grab by google. (think about 400 posts ago I asked this?)
In summary have to agree with he others that advanced warning wasn't present on a major impact to googles partners. This is probably the most disturbing issue. Always knew they could jerk around the pay scales and was willing to take the risk, however doing it with no notice is truly unprofessional and lacks in trust building at the least.
I was going to give it a month's data to see what happened, but if this continues the change will be made before the week is out.
And it will probably involve direct deals with the companies whose ads have been carried on my site. I'm making them way more money than is reflected in the current AdSense payout, and we'll just cut Google out of the picture if this keeps on going like it is.
MQ
The main problem with the changes seems to be the random ads that showed up. Today i have had web design ads even though my site is nothing to do with web design at all.
After doing some thinking it appears us as publishers were too hasty in accepting such an open-ended agreement from Google. They held all the chips, did not disclose payment formula and made the relationship very one-sided. A good contract protects both parties, not just one.
Once publishers start leaving en masse, and/or news of thousands of unhappy publishers leaks into mainstream media (not just web sources) things may change.
