CPM bids versus CPC bids

I can't believe the data

[bumpski]

I use many URL channels and custom channels to analyze site Adsense performance. The new interface provides the new statistics of CPC bids versus CPM bids. The report shows a ratio of 10 to 20 to 1 of CPC bids to CPM bids (Ad Impressions) for my sample site. This cannot be correct. If you select the Page View version of this stat the ratio is ridiculously high.

I've always questioned Adsense earnings for CPM ads as being non-competitive with CPC ads on low traffic sites, and this stat may be a clue.

I think this site has far more CPM ads than are being reported by this new stat. Perhaps Adsense is not including Text ads that have CPM bids in the CPM stats?

If you use URL channels on a good selection of your pages that have relatively low traffic at least in this example you can clearly see from the old statistics that there must be many CPM ads. Simply by looking at page earnings for pages with no clicks. I would say for this site that ratio of pages with earnings and no clicks to pages with earnings with clicks is easily 4 to 1 on any given day. Therefore I would expect the stats for CPM bids to CPC bids to be at least 4 to 1, no where near what this new BID statistic is showing.

In the past with many channels it was easy to see CPM rates of 0.33 cents for page views with no clicks. This can be done by sorting by earnings. One trick when you do this is to realize that pages that have absolutely no earnings are listed in the results in the old interface in reverse alphabetical order. Again in the report you would see pages with 0 ecpm and 0 earnings list between pages with non-zero ecpm and the reverse alphabetically sort page's URL channels. The way to see this easily is to have lots of channels with low traffic pages OR you can also look at your data first thing in the morning (Pacific Time!). This is when you can look at pages with zero clicks but with earnings and ECPM displayed. These must be CPM ads ?

I will say at some point in the last two years CPM rates somehow jumped by a factor of 10. As I said I used to see rates well below a penny versus pennies these days. Quite a change.

Anybody else look into this new stat?

Does it make sense in your case?

[jetteroheller]

Just looked at my stats. Placement targeted ads are coming back.

They are about 3% of impressions and 2% of revenues.

Some month ago, they had been nearly not existing.

[Bddmed]

The problem is for reports where a single page view generates statistics for multiple rows: Ad units, Ad sizes, Custom channels, Ad types, Targeting types, Bid types. For instance, if a page contains ad units with different ad sizes / ad types / etc., then only the ad size / ad type / etc. of the first ad unit will get the page view. The other gets no page views, hence show invalid page views, page CTR, and page RPM.

[google.com...]

Maybe this clarifies a bit of what you're experiencing.

[bumpski]

Even when I look at "Individual ad impressions" the CPM bids versus CPC bids simply make no sense.

I'd appreciate an opinion on my conclusion that pages that show earnings without clicks must be CPM. I simply don't know what else they could be. I have no targeting enabled.

These CPM bid stats simply make no sense and may explain why my sites made 40% more earnings back in the day when Adsense was kind enough to turn off CPM ads.

Clearly Adsense's ability to statistically analyze low traffic pages is overwhelmed by the requirement that their algorithm respond quickly to advertiser's bid changes which can occur on an hourly basis. CPC ads on these low traffic pages produce way more earnings per month or year than CPM ads.

And I'll repeat the assertion that CPM ads displace the better paying CPC ads on low traffic pages. An assertion I've made in other posts. Google's eCPM (effective CPM) calculations for CPC ads still need a lot of improvement for low traffic pages.

[ken_b]

OR you can also look at your data first thing in the morning (Pacific Time!).

No. It takes too long for stats to equalize or settle out to be able to do this on the day in question. No same day stats can be trusted to be absolutely accurate. That's probably why same day earnings are refered to as "estimated".

pages that show earnings without clicks must be CPM.

That's my understanding

IF

you are looking at stats from at least the day before if not from the previous month, or better yet after we've been paid for the month in question.