Lower EPC for lower CTR?

Is it just me, or does this really happen?

[jonathanleger]

I had a site dedicated to "widgets", and the ads being shown were all 100% "widget"-dedicated because it's a niche market. I was getting an EPC averaging X. At this time I was getting Y page views per day. All was well.

I did some research into what kind of ads perform better, and generally just wanted to overhaul my site, so I did, with good results. Wereas I was getting Z clicks per day before, now I am getting (Z * 1.5) clicks. Yeeha, right? Wrong.

After overhauling the site, I am now getting (Y * 4) pageviews instead of only Y page views (it makes navigation better). This, of course, led to a lower CTR, since I had more pageviews (same number of visitors, though). Now my EPC has dropped from X to 2/3rds X. I am still earning more than I was before, which is nice, but why did the EPC go down?

Does CTR effect EPC? If so, why? Shouldn't ROI be what influnces EPC, not CTR?

[jomaxx]

I don't think so. I have a second site that I sometimes run AdSense on, increasing impressions 10X and driving down the CTR about 5X. I haven't noticed any effect whatsoever on the revenue from my primary AdSense site, although I'll have a better idea in a day or two when the new stats kick in.

[paybacksa]

I saw this as well... as traffic increased, CTR decreased, but it appears EPC also decreased. So I dropped ads from all low-CTR pages. Total impressions down, CTR way up, and appaent EPC up as well. It did NOT seem to relate to ad content -- I did a decent but not exhaustive examination of that.

[Broadway]

I had the same but opposite experience. I revamped the appearance of my AdSense ads (made them more visible) and boosted my CTR about 25 to 30% (same number of page views). Now my EPC is up, maybe as much as 40% or so.

Several times over these last few weeks it has crossed my mind that AdSense was rewarding my site with better ads because of the increased CTR.

Of course the other explanation, and probably the right one, is that the holidays are over. Advertisers are spending money again.

The added money is great but it just so happens that the extra income boosts me over a certain psychological level on a regular basis, thus making me a very happy person.

[blaze]

Google has conversion data for the advertisers you send traffic to... If you send higher conversion traffic then they will give you higher EPC ads.

[jomaxx]

blaze, what is your source for that assertion?

[blaze]

Just ask any Adwords subscriber. Heck, set up your own Adwords campaign. Good exercise for any AdSense publisher.

I'd say about 4 out of 10 adwords users use the conversion function .. more than enough to get reasonable data.

And trust me, if you had that data, and you had publisher a) conversion rate equal to that of regular AdWords and publisher b) who had a conversion rate that was a quarter of what AdWords could produce .. who would you send the high priced Ads to?

The Googleguys aint dumb, my friend.

They use this to determine fraud as well, btw. If you are sending loads of traffic that isn't converting this is how they figure out whether or not you should stay in the program.

[alika]

sort of optimization -- that the high paying ads are sent to those who convert? this is being done by banner ad networks (though CTRs and some other metrics that the networks do not disclose), so this theory is not really farfetched as far as google's implementation of adsense

[freitasm]

The theory may be correct. I've noticed that in days that my CTR is higher my EPC and overall CPM is higher too. This is even if the impressions are the same (of course, otherwise how CPM would get higher).

[jomaxx]

I wasn't asking because I thought the idea was necessarily farfetched. I think it's somewhat plausible that Google could eventually use conversion data as a fraud filter or to allocate ads.

The problem is that this idea has quickly gone from a what-if scenario to a theory to a belief to a "fact". It's not a fact, and in the absence of any real evidence I think it belongs back in the what-if department.

[alika]

with G, everything's a what if. they never disclose their algo. and the same can be expected with Adsense, how they tweak the campaigns, and how they combat fraud.

[Visi]

Personnaly I like the previous conspiracy theory that there is a cap on expected site revenues, or an allowable CPM based on averages. Clicks go up ...epc goes down. We naturally have all the supporting data to show this, after only a snmall adjustment of past data to fit our theory.

Just adding our 2 cents to the ongoing guessing game called google adsense:)

[paybacksa]

Npw that we have a couple of days of channel data, it appears (hardly conclusive evidence) that clicks on strongly-themed pages have higher EPC that clicks on less narrowly defined pages.

Might G charge more (thus payout more) for more highly targeted ads, as reflected by the conTEXT of the page?

[itisgene]

Well, I want to comment on the conversion tool.
As an adwords advertise, we spent quite decent amount of money and had internal tracking system (not google's - because we don't want to give away our sale info to them). Adwords advertisers have no way to tell the difference of the traffic from which sites performed better than the others.

It is just aggregated by campaign and all you can see is aggregated CPC. Do we increase or decrease based on the conversion? Absolutely. That's why we pull the plug on content ads... I guess the reporting was not correct or the quality of traffic was just horible. The conversion rate was 1/10 of search traffic's. At the same CPC why would you spend the money on content ads if you know they don't drive qulity traffic?

Here is another thought.

I have a few web sites as a hobby and run adsense on them. I HOPE I am sending qulity traffic to the advertisers. If they notice any low qulity traffic, they will turn it off as I did for our company. The EPC then will go down because major advertisers dropped out.

Well, if we get more money next year, we may turn it on again but for now the search traffic seems enough for us.

[blaze]

Absolutely, if AdWords advertisers drop out than general EPC goes down.

But why would google punish all AdSense advertisers across the board because just a few are sending non converting traffic?

While it's not 100% complete, AdWords advertisers do use Googles conversion tool, at least enough that Google can get a sophisticated statistical (generated over a group of advertisers over time, not just one or two) sampling of the quality of traffic clicking through.

They then calibrate the results against what their own Google SERP 'AdSense campaign' does in terms of conversions and use that as a barometer of whether or not to throttle / kick out low quality websites.

Just participate in an AdWords campaign and you will see how naturally Google leverages statistical sampling all the time.