Removing Adsense from Low Performing Pages

an unwanted experiment....

[spaceylacie]

For a long time I have been considering removing Adsense ads from a site of mine that has performed badly in the past but has a large number of visitors. With 10 to 20K ad impressions a day, I figured even with a very low CTR it was worth running some ads. Another site of mine gets about the same number of visitors but has at least 10X the CTR.

Well, the site with the low CTR crashed last week and I have to now switch to a dedicated server to handle visitors. The site contains downloads up to 1 meg each and most visitors download many files per visit so the site uses much more bandwidth per visitor than the norm. Anyway, the site went down during the switch, but guess what? My income didn't go down. My other site on the same account, the one that gets 10X the CTR as this one, started making more money right away, making up for what I lost from the low CTR site being down.

So... I am thinking... when I get the download site back online, I may replace Google ads with ads for my other site, the one that gets a much higher CTR. The sites are both about the same general topic, widgets, of course, so the ads would be topical.

Any advice? Do you think having G ads on my download site is just dragging me down with it's low CTR, earning-wise, and I should instead re-direct as many as possible to the higher earning site? I don't mean actually re-directing them, but instead putting ads for that site where they can be seen more readily, in the G "hot spots", where my Google Adsense ads used to be.

[dirkji]

That's interesting, let us know if the effect lasts. I wouldn't interlink the sites if I were you.

- Dirk

[spaceylacie]

It's possible to link without blatantly showing the link to SEs. If I do replace the G ads with my own ads, I may "hide" the link. And, of course, I would not make them look like G ads.

I am watching stats closely to see if the trend continues.

[zett]

Now, this is indeed a very strange situation. Interesting.

Just a thought...

Apparently the visitors/clicks of the site that went down have no value to Google AT ALL, so I would not send the visitors to the other site (so that they can drag that site down as well). I'd rather use the site as a playground, testing other ad networks, maybe CPM stuff? Or put up a link saying "this space for sale".

[spaceylacie]

The site came back online tonight, G ads still in place and my overall ecpm immediately tanked. It has been down just over a week. So you agree, zett, this site is dragging me down? I don't know exactly why G doesn't like these visitors/clicks, but apparently they don't. The main thing I can think of is that the visitors are mostly non-US residents, compared to my other site which is mostly US residents. Also, that site is about getting stuff for free, while the other is more geared towards sales. With both sites, the visitors are the same gender and age group. But, different income levels.

I may just take your advice and use it as a playground to try out other options. I don't see the point in continuing to run Adsense on the site. My overall account seems to be getting penalized because of it.

[MikeNoLastName]

OF COURSE your OVERALL eCPM is going to tank if the CPM for the downed site was inherently lower than the other (up) site (duh - since you're averaging the two together). That doesn't mean the down domain is useless. It may still be making you significant money independent of the other domain. What you need to look at however is whether the specific CHANNELS for the up domain ALONE were affected. Hopefully you had enough URL channels defined on the up domain before and after the down time to get a statistically significant sampling to determine this. A week, should just be sufficient if the traffic is high enough. If the PPC and eCPM for the up channels ALONE were affected negatively while the other was up and positively while the other domain was down, THEN and ONLY THEN, might you consider to vanquish the down domain ads.
On the other hand, I too have noticed overall advantages to "weeding" out certain individual high traffic pages on ALL domains which had particularly low CTR/eCPM and directing visitors to closely related well-paying pages.

[edited by: MikeNoLastName at 9:06 am (utc) on Sep. 13, 2007]

[joelgreen]

IMO this is somewhat similar to Anne's test (there is a thread on the WW, removing channels increased revenue). But I believe this is short-term behavior, and your earnings may decrease soon.

[dirkji]

Removing channels is completely different than removing ads.

[joelgreen]

Removing channels is completely different than removing ads.

Yes, but both scenarios seems to "shock" AdSense, which is leading to unexpected earnings increase/decline.

[spaceylacie]

Mike, the overall eCPM for all channels on the second site also went down as soon as the first site I mentioned came back online. Say it made X number of dollars before the first site crashed, after the crash it made X + 20% a day on average. The first site, the download site, only makes up about 10% of my total earnings for the month and the second site I mentioned makes up about 75%(the other 15% is spread out among some other newish sites).

So you see what I am saying? I think I'd actually make more money if the download site didn't contain any Google ads. Unless, of course, it's just a short term thing, like a shock factor, causing the eCPM to go up to X + 20% for the other site during the time that the first site was down. In another week, maybe the shock factor would be over and it would be back to making X a day. I still have G ads on the first site while I figure out what I am going to do. All advice welcome.

[freelistfool]

I have a download page on one of my sites that gets a lot of traffic but virtually no clicks. The users hit the page from google find the download button and are gone.

Pay Per Click isn't a very good use of advertising space on a download page. My advice is to replace Adsense with Pay Per View ads as a test. You might find that you make more money with the Pay Per View and keep the extra money caused by the anomaly you're seeing with Adsense between the two sites.

[Jane_Doe]

It is difficult to perform any controlled tests with Adsense because Adsense income, advertisers and traffic fluctuate daily, so you always have multiple variable changing in any test which tends to muddy the results.

However, I have tried taking sites on and off again over a period of months to see if any trends emerged regarding the earnings of the sites remaining on the account. For me it didn't seem to make much of a difference to the earnings of the sites that remained with Adsense across multiple sites if I took other sites on and off the same Adsense account. But there are many variables in situations like this, so it is impossible to draw any definitive conclusions. My earnings may have been impacted if I'd left the other sites up for a longer time, or if the sites I took on and off were higher traffic sites.

For me my best guess is that smart pricing is not occuring across the board on all my sites, and adding or removing one site does not seem to impact the earning of the other sites. Your experience seems to differ, so you might try adding this second site on and off to the account over time to see if you can spot any definitive trends.

[linear]

Spaceylacie, you of course need to do your own test, but what you are testing worked positively for me when I did it (a couple years ago).

Just to restate it to avoid ambiguity, I removed the lowest performing ads (which amounted to pulling them off an entire site) and increased my earnings (not just EPC).

[mixart]

What about creating a seperate adsense account for your seperate site - then the two sites performance won't be connected in any way to the other.

[Content_ed]

Any advice? Do you think having G ads on my download site is just dragging me down with it's low CTR, earning-wise

Yes. There's nothing new about this, I can barely comment on any thread without mentioning it, and I got the idea here from other publishers years ago. It has always worked out for us that not running Adsense on pages that don't convert well results in increased earnings for our site. In our case, we just target an eCPM of $XX.XX for every page, and those that don't make the grade over time go ad-less.

Out of laziness and workload, I didn't follow that course of action this summer and our earnings were down, year over year, for the first time. I used the excuse that I wanted to see how Google would manage the site without "interference" from me, I'll be smarter next year. As is, we run Adsense on less than 25% of our page views. Putting more pages into the mix with lower eCPMs always results in quick loss of overall revenue. The only thing I'm not really confident about is whether it's the low eCPM or the low conversion (the two are usually but not always coupled) that hurts. I'm beginning to suspect it may be the low conversion.

[MikeNoLastName]

-What xx.xx would one suggest? We've always aimed for a number which includes 3/4 our our impressions and remove 1/4.

-Unfortunately as publishers we have no idea what our conversion factor is, or how it is being calculated. Say you have one high volume really sucky advertiser, who has a poorly converting site in general but has been buying 50% of your traffic for years (now paying 1/4th of what they were a year ago because of their constant degrading of you), and has perhaps saturated the market, while the other 50% as a group are overjoyed, but don't bother reporting their conversions. What can one do?

[joelgreen]

target an eCPM of $XX.XX for every page, and those that don't make the grade over time go ad-less

Had few pages having eCPM < $0.5 for more than 6 month. I removed ads from those pages. Few months later added ads again, just to check if something changed. eCPM is very good on those pages now (for about 6 months).

As a conclusion, pages which do not perform well now, may perform well or even outperform other "good" pages later (maybe "right" advertisers pop in, not sure).
YMMV.

[Content_ed]

As a conclusion, pages which do not perform well now, may perform well or even outperform other "good" pages later

True, and that's why active managment on the part of the webmaster is required. If your website is a large forum that only covers one topic or a huge encycolpedia that covers everything, it may not matter. But if you have several discrete topics, they are bound to show some seasonality, due to holidays, school years, the construction cycle, etc.

The trick is more deciding when to go live again than when to take ads off. We count on past performance for that.