smart priced -> daily, weekly, monthly?

bad question. if we are more thinking about it, we will get answer...

[moonkey]

Smart pricing is based on conversion ratio. If clicks don't lead to any kind of conversion, we're smart priced, so that mean advertiser will pay lower per click, and we will receive lower revenue from those click/s than regular.

How smart priced works? We're not sure, but if we are more thinking about it, how conversion ratio is calculated?

Let's define examples:
1) Site A, 10 000 visitors daily, 1500 clicks average
2) Site B, 1 000 visitors daily, 150 clicks average
3) Site C, 10 visitors daily, 1 click average

Advertiser has ad about dog food, and he owns online pets food shop.
Site C leads to this shop by 1 click, the clicker hasn't convert anything (he hasn't buy anything). For one week, there will be 7 clicks generated and let's say no conversion. If this account should be smart priced on just 7 clicks? I'd say no. There had to be more than 100 clicks to determine what's conversion ratio if there is any. So for some trafficed sites, they can be smartpriced right away, the others may don't, I think it could wait for months until someones account can be smartpriced.

I am sorry for writing all this.. but I can not explain the other ways like this. Hope you understood me :)

what do you think about this? let's bet if there'll be more questions regarding smart priced period...

[europeforvisitors]

Smart pricing is based on conversion ratio.

It isn't that simple.

According to Google, smart pricing is based on the likelihood of conversion, in which--again, according to Google--the type of content plays a role. So Google doesn't need to wait for actual conversion data (which may not be available anyway) to determine a starting point for smart pricing.

Obviously, any conversion data that does become available can be used to adjust the level of smart pricing for a given site or account.

Also, there's been a lot of confusion on this forum about what a "conversion" is. A "conversion" isn't necessarily a transaction, because not all advertisers are e-commerce sites. Depending on what an advertiser has chosen to define as a "conversion," a conversion could be a request for more information, a registration, a newsletter subscription, or some other business action that doesn't involve an immediate sale.

[djulien]

Well explained.

So by the same token, the likelihood of conversion must be gauged by some sort of manual once-over to determine site quality.

Is it really as subjective as that?

[Green_Grass]

It is a major mystery. Basically it is vague on purpose so that advertisers have an incentive to advertise on the publisher network.

It goes up and down randomly, based on some vague unknown formula and we publishers should basically try to live with it.

Just be happy when suddenly EPC goes up.. and try to work on a monthly average.

Also maybe QS of the website may also determine EPC values. QS may take into account type of content and this could be used by Google to 'justify' its smart pricing to itself. I am not convinced, but ofcourse , I am basically happy with Google or I would not stick with them.

[jimbeetle]

It's very unlikely that there's a manual component at all to this as the sheer number of publishers and number of publishers' pages would make it virtually impossible to scale.

Adwords advertisers are able to include a piece of tracking code on a page, presumbably one that can only be reached if the user performs an action that is considered a conversion. It's this data -- as far as we know -- that is used to build the historical liklihood that a click will convert.

Whatever ways Google massages the conversion data or combines it with other information is all a matter of speculation, sometimes farfetched. No one outside of Google knows.

<added>Green_Grass, good points on QS, one I don't consider too farfetched. ;-)</added>

[edited by: jimbeetle at 3:23 pm (utc) on Aug. 2, 2006]

[moonkey]

djulien you're right. that's what I am saying.

europeofvisitors, I am aware of all those conversion, newsletter, sign ups etc. not only shopping, that was just an example.

even i am more focused about "what clicker do on the advertisers site", I'd say there's so many options to have track them all. It isn't so simple as just copying urchin tracker code on the advertisers page.

I think the conversion is based on how much the visitor is really interested in product/offer. This can be monitored how long has he reading that site, how much pages he has visited. This visitor should be marked as a potential costumer in front of all those who close their window before the site has even loaded or right after window loaded. This is just my opinion.

[hunderdown]

moonkey, even for a site that has lower traffic, smart pricing could update weekly--it will just be more up and down. Some weeks it goes up, because the publisher had good clicks the previous week, and other weeks down....

[europeforvisitors]

So by the same token, the likelihood of conversion must be gauged by some sort of manual once-over to determine site quality.

As JimBeetle says, manual reviews aren't scalable. When smart pricing was announced, Google used two examples to show how content could affect the likelihood of conversion:

- A page of photo tips (lower conversion)

- A camera review (higher conversion)

Google has many types of data at its disposal, including the data that you mentioned in your most recent post. It's likely that smart pricing gets tweaked for individual sites or accounts as more data becomes available, but type of content and other on-page factors certainly could be used to determine an initial smart-pricing discount.

[mt1955]

Does anyone here know how granular "Smart Pricing" is? Account? Site? Channel? Page?

[jomaxx]

Supposedly Google has confirmed that it's done at the "account" level, but I'm actually disinclined to believe that because it's patently unfair.

NOT to publishers, I hasten to add. If Google used a weighted average based on how each page contributes to the account's income, then it would be completely fair, and in the long run the result would be more or less indistinguishable from doing smart pricing on each page individually.

No, it would be unfair to advertisers. Say a publisher had one killer website and one dud with low-quality traffic, in whatever combination led to the account having an "average" rating overall. Then advertisers on the dud website would be overpaying, and advertisers on the killer website would be getting a nice discount where none was really justified.

[Scurramunga]

Basically it is vague on purpose so that advertisers have an incentive to advertise on the publisher network.

Sounds plausible

[mt1955]

Say a publisher had one killer website and one dud with low-quality traffic

So then is it the idea that sites getting traffic by tricking the visitors into coming there (MFA's, link farms, junk pages etc) are probably going to have low quality traffic - fewer conversions - and so paying less will discourage those practices?

Anyway, what then - if anything - can a white hat webmaster really do to influence high vs low quality traffic? A banner on each page saying "Be Gone All ye Non-convertors" would probably not work. :-)

[djulien]

Quality inbound links, outbound links to highly page ranked sites, and plenty of middling to excellent content (in a nutshell) should draw increasingly legitimate traffic over time. It worked for me, at least.

Just make sure that you at least target a specific audience with your choice of keywords and stay as much within theme as possible.

You're already there if you're using white-hat methods ... I'd assume that traffic via the more disreputable of PPC advertisers would result in lower quality traffic. I've always used Adwords for my campaigns, personally.