What CTR % should I prune ads?

Assuming account-wide smart pricing rumors are correct

[androidtech]

If we assume that the rumors of smart pricing being applied account-wide are correct, at what % CTR do you start pruning ads from low performing channels? I'd like to hear from those of you that have had success pruning ads from pages with low CTR's and have had their earnings rise because of it.

[wyweb]

I don't believe there's any standard you can apply to all websites. My high CTR might be terribly low for your site(s), or vice versa. What I did was remove ads from low performing pages - pages that were low performing in relation to other pages on the same site.

[LeChuck]

Couldn't your pages with the lowest CTR be the ones that convert the best?

[ElvisFan]

I did a major stats count of all my channels and low performing pages on all my sites over a period of two months... it was a hard slog...

Once I removed ads from all low visited pages (under 10 visits per month)my CTR & CPM increased to the status quo before Smart Pricing hit me on about the middle of September.

Fingers crossed, so far this month my bottom line has increased by two fold.. and hopefully will be okay for the Christmas rush... LOL

[wyweb]

Couldn't your pages with the lowest CTR be the ones that convert the best?

They certainly could.. just as your pages with the highest CTR could be the ones that convert the most poorly. The fact of the matter is that we don't know.

What we do know is that in removing ads from poorly performing pages (again, in relation to other pages performance) CTR overall will likely increase. If CTR is indeed a factor in the smart pricing algo then this is bound to be a good thing in the long run, no?

[hunderdown]

Personally, I only take the code off if the CTR, relative to the site average, is really low--say less than a sixth. I can do something else on those pages. I also look at what clicks are bringing in. If the clicks are good value, even if there are only a few, then I will try harder to increase CTR.

It's going to depend on your site. Maybe your site average is 20%, and anything worse than 2%, in your judgment, isn't worth it. But if your site average is 5%, you would want to set the cut-off a good bit lower.

From recent discussions about smart pricing, it's now evident that characteristics of the visitor and the ad are factored in, so it's not just conversion data (IF available), CTR, and who knows what else.

[NoLimits]

CTR is irrelivant if Google is telling the truth in their blog.

It's CONVERSIONS that you need to focus on (conversion rate to me is indicated by EPC...

high epc = conversions are good
low epc = conversions are low

Google claims CTR is not a factor (and it shouldn't be)

[TheDonster]

On my site I had a leaderboard at the top directly where the heat map suggested. Then I put a banner ad completely at the bottom just in case visitors decided 'Now what?' when they finished reading the page. The bottom banner had a horrible CTR compared to the LB, (obviously since it was below the fold) but it still brought in money each month. I finally decided to remove the bottom banner to see if earnings would increase. The result (again in my case) was that CTR remained steady for the LB, but my EPC increased three times. Overall earnings increased slightly. Your results will definitely vary.

[hunderdown]

No Limits, you didn't read what Google said very carefully. The actually said "CTR is not related to advertiser ROI". That is NOT the same thing as saying that CTR is not a factor in Smart Pricing. To the best of my knowledge, they've never said that. Conversions can only be factored in if the advertiser reports conversions. Google still has to do something if they don't....

CTR may not be directly factored into Smart Pricing, IMO. But something related to it clearly is, or what a number of us have seen happen to our earnings when we remove low-performing ads wouldn't happen.

Also, it's not accurate to say that high EPC equals good conversion, and low EPC equals poor conversion. Some clicks start out being worth more than others. If a site is in an area where advertisers bid no more than 10 cents, EPC will never be more than whatever their share is of 10 cents, even if they convert perfectly. Conversely, if a site is in an area where advertisers are bidding $5 or so, EPC is going to be pretty good, even if the site converts poorly.

[wyweb]

I really have no idea whether CTR affects smart pricing in the overall cosmic scheme of things or not. I wasn't saying that it did. I'm basically saying that I don't have a clue what does affect smart pricing, other than conversion rates, which is a mysterious, elusive statistic I've not been granted access to.

What I am saying is that by removing ads from poorly performing pages I increased my bottom line. A residual effect of this is that CTR went up. Are the 2 related? I have no idea. It would not surprise me in the least if it's determined at some point in the future that they are.

[NoLimits]

Nobody really knows - or they aren't saying anything if they do.

This has been a large factor in the last few weeks... we have no clue! AS leaves us in the dark. A lot of publishers have cut down on AdSense as a result of their inability/unwillingness to share necessary information.

Why wouldn't they want us to know how to improve our conversion rate...? Advertisers want more conversions - we want more conversions. Let us optimize to achieve this damnit.

[david_uk]

I track all banners with channels, and then look at the results every week or so. If a channel is a poor performer then I remove the banner. I do use ctr, clicks and bottom line earnings in the decision to remove a channel, but my main metric here is eCPM.

Currently I'm at the state where I know what pages will / wont work on adsense, and use other banners or most often no ads on those pages.

There is a plus point about not having ads / different ads on pages in that it helps to limit ad blindness. There may be an indirect positive in doing this in that visitors may click more on the banners that *do* work for you.

[hunderdown]

Sage advice, David. Throwing ads on every page and every possible location on every page can be counter-productive.

You can also have one adblock, and vary the location where it appears....

[ronburk]

a) use common sense If one of your pages is devoted to saving money by knitting your own potholders, and that page is mainly displaying ads for $5,000 professional ovens, then common sense says (if the search queries used to arrive there say nothing to the contrary) that that traffic (regardless of CTR) is likely not very well qualified. Instead of just thinking about dropping AdSense from that page, however, you should think about further qualifying the traffic yourself. Replace AdSense with your own ad -- a pointer to another url devoted to professional ovens. Common sense says that if they click on that, and then click on an ad for a $5,000 professional oven, they have a much better chance of being qualified buyers.

b) test Try a revision that remedies what you suspect is putting ads in front of poorly-qualified visitors. Test for a month and see if revenues are up, down, or not significantly different.

This is another case where it helps to also be an AdWords advertiser. On the AdWords side, you learn that you can sometimes keep that low-converting traffic from clicking on your ad by featuring the price in the ad. Advertisers get 3 brief lines of text to filter out unqualified traffic; it ain't easy. Publishers have infinitely more options to filter out unqualified traffic -- if they understand the problem and make the effort.