Forum Moderators: martinibuster
How does G know if a click through converts?[
Short answer: They don't.
Also, remember that, for conversion-tracking purposes, a "conversion" doesn't have to be a sale: It can be an advertiser-defined "business action" such as an inquiry, a registration, visiting a certain number of pages, etc. (Not all advertisers are looking for immediate transactions; advertisers of high-end products and services are more likely to be seeking qualified leads than e-commerce sales.)
Side note: What the AdSense publisher earns per click is determined not only by the bid price and the "smart pricing" adjustment, but also by Google's compensation formula. We don't know what that formula is, or what factors it looks at when calculating the publisher's share of the advertiser's payment.
I've found some interesting stuff myself, such as Google's Conversion Tracking Guide, [adwords.google.com...]
Longer answer: They don't have to know if one particular clickthrough converts; they only have to know what the statistical likelihood of a conversion is, based on conversion tracking and any other research data that they may have available.
There is no way to establish "what the statistical likelihood of a conversion is" for any given page without having access to a statistically significant number of actual conversions to analize which they do not have.
The quality of ad copy and landing pages vary among advertisers just as the quality of publisher sites varies. It's all over the map in fact.
Also, are low conversion rates for sites in a given industry due to low quality content pages or low quality ad/landing page combinations?
IMO there simply isn't enough concrete statistical data available to Google for them to accurately classify a site (or a page) as "likely" or "unlikely" to convert, especially since so much depends on the ads and landing pages themselves (the quality of which varies greatly from industry to industry and site to site).
This could explain why a couple of my sites where "smart pricing" has driven down the EPC considerably draw frequent requests from advertisers to replace the AdSense blocks with their exclusive ads at a premium fee. To me this indicates that actual advertisers see those pages as more "likely to convert" than Google's "Smart Pricing" does.
I assume also all this MLM people have no conversion tracking.
Both of them are in my URL filter as far as I spot them.
IMO there simply isn't enough concrete statistical data available to Google for them to accurately classify a site (or a page) as "likely" or "unlikely" to convert
They obviously have a different opinion; last April, when they introduced smart pricing, they explained that prices of clicks would be adjusted "based on their expected value to the advertiser." It's unlikely that they came up with "expected value" by winging it. Maybe it was a case of garbage in, garbage out, but they must have been using some kind of statistical data to determine that expected value.
The question is a bit academic, of course, since it's their network and they get to set the pricing formula. A more interesting (and equally academic) question is why some publishers are now reporting big drops in EPC and others have held steady or gained. Smart pricing has been around for nearly 10 months, so what's changed recently? Have advertiser discounts been adjusted again? Is the pricing formula drawing on new types of data? Or do the changes reflect something besides smart pricing, such as differences in how the payout formula works for different types of content, traffic levels, profitability levels, etc?
Until they do that, it is just a gimmick for them to convince advertisers to spend more on contextual advertising and blow-off publishers who complain about payout.
As I wrote in another forum, I find it shocking how much blind-faith people put into "smart pricing."
They obviously have a different opinion; last April, when they introduced smart pricing, they explained that prices of clicks would be adjusted "based on their expected value to the advertiser."
Their opinion is just that: opinion (as is mine). The difference is their opinion is based upon incomplete data (I'm being charitable with that description). Mine is based upon hard statistics.
While it's true that it's Google's game (and at least for now I'm very happy playing it), the answer to the initial question is "They don't". They can and do make educated guesses but those guesses are obviously very wrong for many sites (at least two of which are mine).
As universally accepted - Google can track conversion only on a limited number of keywords and with a limited number of advertisers. I've run a couple of experiments. On non-Adsense pages where I know I have low conversion (advertisers provided figures) I removed all advertisers and ran Adsense. The Adsense EPC is higher on those pages than on almost identical pages attracting the same ads. And the difference is even more marked post Allegra. (Sample size: a few tens of thousands of impressions per month on a low double digit CTR). And, these are products that can be tracked to conversion.
So, yes, they are applying some factor in calculating % of payout per click. Whether it's Smart or not is another issue.
Perhaps they've realised that the enormous amount of data available on trackable products doesn't really relate to the non-trackable products. And, they're using something other than conversion data in their payout formula ...or scrapped the Smartpricing idea altogether and are applying this "other" factor across all publishers/sites.
Sorry, EFV, I know you've often argued that SmartPricing works and is what causes the statistical drops in earnings some people here report. It may give a warm feeling believing that your clicks do convert (whether it's for reasons of your ad blocks not being colour-and-design "integrated" .... or some assumption that content makes the clicker more likely to convert) but if you are getting a higher EPC than a similar site attracting similar ads it may be worth examining that the difference in EPC involves a factor outside of SmartPricing... as you stated:
A more interesting (and equally academic) question is why some publishers are now reporting big drops in EPC and others have held steady or gained.
External auditors sounds good but I can't see it happening anytime soon :)
They at least have AdWords conversion data from my inquiry pages. That's some pretty concrete data.
Yes, it is. But unfortunately your conversion data is offset by the majority of webmasters who don't provide conversion data to Google.
Sorry, EFV, I know you've often argued that SmartPricing works
I've never argued that smart pricing "works." Indeed, I've often argued that worthless traffic is worthless, even when the advertiser buys it at a discount. That's why, in my opinion, smart pricing isn't an adequate substitute for advertiser controls and/or a tiered network.
If smart pricing "works," it works to the extent that it saves advertisers from having to pay full retail for clicks that, in Google's estimation, have less likelihood of converting than clicks from AdWords on SERPs do.
In the long run, smart pricing also may "work" to discourage junk sites if Google assigns greater discounts to sites with certain characteristics. For example, it would make a lot of sense for Google to assign bigger discounts to "scraper sites," because--among other things--such sites drive up the costs of quality control for Google Search.
But I'd never argue that smart pricing and actual conversions are in perfect synch. Smart pricing's main value is in making the content network more palatable to advertisers.
Smart pricing's main value is in making the content network more palatable to advertisers.
Or to play a Public Relations role to both sides of the coins. I have some anecdotal evidence that scrapers are getting paid more than content publishers. I have no evidence I can present but a stong enough suspicion based on what I've seen that it's happening in at least one case. If that is largely true across all scraper sites there are two possiblities
1. The scrapers are converting better or
2. SmartPricing is a sham.
figment88 may indeed have struck on something. If Smartpricing was a sham - note the italicised "if" - then Google would be the one who loses the most by admitting it.
That's why, in my opinion, smart pricing isn't an adequate substitute for advertiser controls and/or a tiered network.
On this we agree 100%. Allowing advertisers to "opt-in" to specific sites or a category of Google hand-vetted premium sites would be an excellent idea. I like your often described "AdSense Select" idea. I would even be willing to pay a review fee if neccessary.
My posts might lead you to believe that I'm all for publishers at the expense of advertisers. That is not the case as I'm heavily involved in both areas.
My beef is with the "smart pricing" ruse that Google has implemented in an effort to:
1 - Increase the bottom line for investors (surely no one here believes that the entire smart pricing discount is passed on to the advertisers?)
2 - Appease the advertisers.
I have two sites (out of 6) that are given the smart pricing axe on a regular basis. I know it's smart pricing at work because none of the other variables (even working together) could cause such regular, almost predicatable drops in EPC when impressions and clicks are staying the same.
If this weren't the case I wouldn't have AdSense advertisers frequently contacting me asking to sponsor my article pages in place of the AdSense ads. If smart pricing was working as described, why would they willingly agree to pay a stiff monthly fee when they could simply use AdWords and get traffic from my sites on the cheap.
Advertisers should to be the ones determining the value of clicks from specific sites and pages, not some Google algorithm.
