Does taking Adsense off low CTR pages increase overall earnings?

[astrobiologist]

I've heard this a lot over the years: that leaving Adsense ads off pages that generate a low CTR will improve your bottom line. Is this one of the Adsense urban myths or is there some substance to it?

I noticed that the good folk at Adsense recently started sending out tips again -- more specific but also somewhat crptic. Mine was that I could be earning more if I allowed Adsense ads on the 43% of my site that is currently (according to their reckoning) Adsense-free. I'm not sure what this 43% is referring to because certainly over 90% of my (one-and-only) site's pages carry Adsense script. Maybe it refers to file size, or some other measure.

Which brings me to my point. I don't have Adsense on some very large (in terms of word count) pages on my site -- including my home page -- simply because they are so diverse in content that they result in poorly targeted ads and hence low CTR. But am I leaving "money on the table" or am I being smart by not having my whole site penalized for having a lower overall CTR?

[drall]

In my very advanced testing I have found this to be a total myth. Your just leaving money on the table.

[zdgn]

Ditto drall.

I tested not so long ago for an extended period, and overall earnings actually decreased over time with fewer ads on fewer pages and increased with more ads on more pages. In fact, increasing the number of ads/pages has never negatively impacted the overall earnings for me, but decreasing always has.

[purplecape]

I think that this belief is based by experiences very early on in the history of AdSense, when a few people reported here that a very low CTR on some pages seemed to have some influence on smart pricing--apparently, the theory went, the overall CTR of the site was looked at. By taking the ads off low-CTR pages, the site's overall CTR went up, and the site benefited by higher EPC. This was pre-2005 or even earlier.

Now the algorithm most likely is different!

But test it and see. Re your home page, I'd try either a small ad unit in a key spot or an adlink unit. Use the AdSense custom tags to tell the bot to ignore or emphasize certain sections of the page. I use this approach on my home page and it earns a nice part of my total earnings...

[wyweb]

I have a few whacked out conspiracy theories about the relationship between CTR and eCPM and how the former can impact the latter. I have no supporting evidence in favor of these theories other than my own observations so I tend to stay quiet about it.

As CTR goes up, I invariably see higher eCPM numbers. With the higher eCPM comes the higher earnings at the end of the day. It's anything but consistent but on the whole there seems to be a correlation.

I stripped Adsense off low performing pages long ago and my bottom line went up. Plain and simple. There could be other factors involved here. We're not given enough information as Adsense publishers to do any serious number crunching. It's basically cause and effect. A/B testing. Change something and see what happens. Change it again and see what happens then. Given a significant enough sample over enough time you're able to make what you think are informed decisions. And then you run with it.

I'm a firm believer in shaking things up when they get stale but I'm also well aware of the fact that what works for one site won't necessarily work for another. All you can do is test.. and then test some more. When you reach optimum performance - leave it alone and go have a beer.

My opinion is that, yes, removing adsense from low performing pages does have a positive impact on my network of websites. I'm not particularly interested in piimping my sites out completely anyway so there's that advantage as well.

Your milage will probably vary....

[alika]

I have tested this as well.

In our case, there was an initial bump up of revenues, but everything fell eventually -- revenues, eCPM, EPC and CTR

But see how it goes with you, and really test it.

[dataguy]

I've proven this myth to be true over and over again. I'm not sure why it doesn't seem to hold true for everyone.

I'm not going to try to change anyone's mind, what works for one might not work for another. In a world where we are prohibited from talking specifics, it's rare to be able to compare notes to see what the difference is.

I'll leave you with one example if you want to try to explain it away.

If you've ever had a site which hits the home page of Digg, you'll experience what is referred to as 'the Digg effect.'

Every time my site has experienced the Digg effect, traffic rises because of that one page receiving traffic, but Digg traffic rarely clicks on ads. Applying the 'more is better' logic you would think that having a few clicks is better than nothing. Instead what you will see is your CPM for your entire site tank. In general terms, a site making $1,000 on a normal day from AdSense could be reduced to $600 for a day with 30,000 Digg visitors. More more ad impressions, less revenue.

This is a short, easy to point out example, but I see it on a long term basis too.

Sometimes I wonder if the difference we see from one site to another is from widely-varied CPM. For example, this myth might be true on sites with an average CPM of $10, but not true on sites with an average CPM of $1.

I don't know, but I'm willingly giving up a couple a hundred thousand potential ad impressions a month because it's how it works on my site.

[Chrispcritters]

I recently stopped serving ads to traffic from countries with (relative to my average) low eCPM rates. This resulted in significantly fewer ad impressions. Revenue has basically remained unchanged. Overall I've seen a decrease in average page load times (reported by GWT), an increase in pageviews per visit. (I'm guessing more people would be willing to link to the site as well.) It's only been about a month, but I'll continue to watch eCPM rates for US traffic to see if there will be more than a minor increase.

[ken_b]

The only way to know if taking ads off low performing pages will work for you is to try it.

Just because it works or not for others means nothing.

[HuskyPup]

Slightly off topic.

For those who do remove their ads off these specific pages, just how do you do this?

Manually or do you use a CMS that allows this to be done?

I don't have a huge site but it does run into several thousand pages and the thought of having to hand code individual pages with/without AdSense code just wouldn't make senese for me.

I can add or remove AdSense site wide with a couple of .inc file uploads. I'd be interested to know how you intend doing this?

[wyweb]

For those who do remove their ads off these specific pages, just how do you do this?

I used a freeware editor called NoteTab Light for years. Upgraded a few years ago to the standard version because it has a spell checker.

It has a global search and replace feature that can edit 100's of pages at one time. Copy/paste/click/save and you're done.

Other HTML/text editors have more features but this is the one I started out on and I tend to stay with what works.

[wyweb]

Sometimes I wonder if the difference we see from one site to another is from widely-varied CPM. For example, this myth might be true on sites with an average CPM of $10, but not true on sites with an average CPM of $1.

And it could very well be dataguy. It wouldn't surprise me if there were half a dozen other factors that played into this as well. The bottom line is that we just don't know.

All of this only undersores the need for testing on your own individual sites. You're getting conflicting opinions here astrobiologist. It doesn't mean somebody's right and somebody's wrong. It only means that there are wide variances to Adsense returns to different publishers. What works for me will not necessarily work for you - and vice versa.

So you test. I'd do that anyway regardless of what other publishers said. I don't take anything blindly. I need to see it for myself.

[astrobiologist]

Thanks for the feedback, guys. Seems like it's another case of "your mileage may vary." As an experiment, I tried putting ads on a few pages that get plenty of traffic but low Adsense CTR. So far, so good. Some extra dollars from those pages and no discernable effect on the rest of the site's earnings or other metrics. We'll see.