AdSense and Google Analytics

My failure story

[kidd]

Hello:

I've been a WebmasterWorld.com reader for a long time now, and the comments and stories published here have helped me out a big deal in the AdSense success on my website. That is the main reason why I decided to tell you about a past experience, not only to share my 2 cents but also to hear your comments about it.

Before becoming an Adsense user I became and Adwords publisher, and I decided make a request for a Google Analytics account. My invitation came almost two months ago, but it was until a little while that I actually decided to use it.

First I was not sure on using the account, my intention was to use it on the site I currently have the most earnings in Adsense, but after reading about this and hearing comments on some people telling that there were no worries I said, "well, ill try it one week".

I logged on to my Analytics account and made the correct setup then I added the JavaScript code to my whole website, which wasn't a real deal since I've been using templates to control the whole layout for a while now. I was eager to see the first results of Analytics and know a little more about my website users.

The 24 hours came, and I saw my stats, this was great! I spent almost an hour looking at every aspect possible of the stats, then I decided to look at my Adsense earnings. And there was a surprise.

My earnings where half as the previous day, I thought that maybe this was just a little hiccup in earnings, this tends to happen sometimes. But then the next day they were worse, and so the next day and the next day.

Also, I started to notice that my ads where less atractive, there were some ads in hebrew while my site is in spanish, and almost all of them where completely off-topic.

That is when I decided to give it a rest, I removed completely the Analytics code praying for my website to return to normal. By the next I removed Analytics my earnings started to improve, and by the second day I got back to normal, my ads as well.

I don't think this is a coincidence, since my income decreased at the same time I added Analytics, and returned to normal at the same time I removed it. But I would like to hear your comments.

Here is the summary:

Day 1 (Analytics added)
Day 2 - 50% decrease in ads
Day 3 - more decrease
Day 4 - even more
Day 5 - worst day in a year (Analytics removed)
Day 6 - 50% increase
Day 7 - Back to normal (I can breath now)

Cheers!

[Erku]

Hmmmm, that's interesting. What do you think of this everyone?

[jomaxx]

Lots of people use both and I don't recall any previous reports that Analytics has led to a significant increase or decrease; so I am initially skeptical.

It doesn't tell us very much to say your earnings were reduced. Are you talking about your CTR, your EPC, and/or your number of impressions served?

[koan]

I know it would probably be hard to do, but in the name of science, maybe you could repeat the experiment a second time to dismiss other factors that could easily make this a coincidence.

But yeah, I'd have a hard time doing it again myself.

[chocorol]

That's really interesting. I've been using Analytics for a while, even before I started using AdSense.

I'm going to remove my Analytics code for a week from all my pages in order to see if there are any changes in my earnings.

I'll keep you posted.

[kidd]

My page impressions didn't have a change, my CTR neither, what had a downfall were the Earnings and the eCPM.

I want to point out that I narrated what happened, and that my conclusion was that it was to weird a coincidence.

And well, Im not to eager to try it again... :P

[Alexstar]

I had some experience similar when I started using Analytics and implemented the code at the top of the pages. Traffic dropped which i guess due to the page loading delay due to the codes at the top. When I moved the code to the bottom, my ad impressions increased, but it never came back to the original level. I still keep the analytics codes for some reason but probably I can try to eliminate it because my host is now giving Urchin stats on their on.

[edited by: Alexstar at 5:12 am (utc) on Feb. 14, 2007]

[icedowl]

There is a somewhat related thread: [webmasterworld.com...]

I'm also removing the script from one site for a minimum of 1 week just to see what happens. Maybe I'll keep it off for longer even if no difference is noticed as that at least makes the page size smaller.

[bumpski]

First I'll ask a performance question:
Has anyone viewing this thread moved the urchin code to dynamically load and execute after the OnPageLoad event? Thus reducing it's impact on pageload performance.

Some explanation:

I'm looking into site page load performance, and in the past I removed Google Analytics code shortly after installing it, because it was degrading website performance significantly. Of course any page load time degradation would probably impact earnings.

BUT recently after reading posts here (Thanks Webmasters all) I noticed the instructions for urchin now indicate putting the code at the very end of the <BODY>, where originally when Google introduced Analytics the instructions said to put the code in the <HEAD> section of the page;

A HUGE difference!

But still, even by installing the urchin script at the end of the <BODY> section, it can really affect page load time. (Browsers seem to give precedence to loading script source code) I guess this could even be in the standards.

So anyway dynamically loading urchin after the onpageload event would assure it doesn't affect page load times. I'm hoping there's some brave soul out there who's done this already?

[KenB]

It seems to me that the page loading issue may be the most plausible explanation to the problem. Moving the code to the bottom of the page would be a significant help. I'd also look at other on page factors that could be optimized to offset the reduced loading time caused by GA (eg. optimizing/removing images cleaning up HTML code).

The reports generated by GA are really valuable for helping one tune, tweak and market one's site.

[Fish_Texas]

Folks, I believe you've solved a problem for me.
Thanks for posting these goodies...

We have 2 sites...one with low bounce rate and one with high bounce rate.
I checked everything to discover the cause to no avail. This had me stumped.

Low bounce rate site has GA on bottom...
High bounce rate site has GA on top under Meta Tags before <body>

GA script being moved to bottom as I write this.
It might also add to Adsense $$$
Thanks Again
Fish Texas

[Pengi]

I've always used Analytics - virtually since day one - I've not noticed any problem with getting targetted ads (apart from one period of about 2 weeks when they all changed to one sub-topic but is was still relevant to my site).

Of course if I'd never used Analytics I could be doing twice as well!

[jomaxx]

The page loading issue may be relevant for some, but not for the OP, who reported unchanged impressions and CTR and a reduction in EPC only.

[Jane_Doe]

I personally don't use Google analytics because of similar concerns - not the page loading but because I wonder if my stats aren't up to par I'm worried I won't get the best ads. I don't know if that is a valid concern or not, but since there are lots of stats programs out there I prefer not to take a chance and would rather limit the amount of information Google has regarding my sites.

[edited by: Jane_Doe at 10:39 pm (utc) on Feb. 15, 2007]

[Fish_Texas]

Jane Doe...
Any recommendations for tracking software?
I don't like using GA also.
Post here if permitted or send a Sticky

Thanks
Fish Texas

[joelgreen]

I prefer not to take a chance and would rather limit the amount of information Google has regarding my sites

I prefer GA account because it is trusted by Google, thus one would have better chances to be reinstated in case of ban from AdSense. Also there would be no need in sending server logs, just provide GA login. IMO.

[bumpski]

The page loading issue may be relevant for some, but not for the OP, who reported unchanged impressions and CTR and a reduction in EPC only.

The page stats will be the same whether the visitor chooses to hit the "back button" or click on a Adsense ad.

If the site is slow there is less time to look at the ads. With the urchin script at the top of the page it could block the Adsense scripts from running for a long time, seconds!

One thought I want to investigate further is using the "defer" attribute for scripts on both urchin/analytics scripts. Make sure the page is fully loaded then do secondary things like statistics. Or just invoke the urchin script with the OnPageLoad event.

I used analytics when Google first introduced it and the performance was extremely inconsistent. Again this could delay display of Adsense ads, and the slow page rendering results in the "back button" instead of interest in the ad that should have been displayed, stats stay the same.

Google moved the urchin script to the end of the page, there must have been a reason. A little note to Adwords/Analytic users would have been nice!

[walrus]

Aside from page load time, with the constant fluctuations in ecpm there is a good chance its just law of averages, some will experience a decline or increase anyways(smart pricing, spent ad budgets, inreasing puiblisher competion, etc.)that just happened to coincide with adding analytics.
I added them specifically so that Google could see the refering search strings because it would help them to rate the quality of my traffic.
For example I have visitors that arrive via searching for "widgets for sale in melonville" which is geo specific and more likely to convert than visitors that arrive via searching for "how to fix a widget" or "widget tips".
Without analytics G doesnt know know the search string, unless of course it was via G, but thats G not adsense. If they can see that you are getting quality leads (which is what it boils down too) they may be more likely to feed you better Ads because you have or are more likely to have a good conversion rate.
I could be wrong, but just may be an angle to consider.

[celgins]

Interesting.

I've used GA for a while now and haven't noticed any drastic changes in eCPM or ad relevancy.

[chocorol]

Well, it's been almost a week since I removed the Analytics code from all the pages of my site. I've got to say that I did not see any drastic change in EPC or in the bottom line.

I'm adding the code again, I miss all the info that GA can show of my visitors.

The only time I saw a difference in my earnings was when somebody targeted my site. My bottom line dropped more than 50%. I immediately blocked that advertiser which only made my site look ugly and somewhat spammy with those giant font ads.

[ronburk]

thus one would have better chances to be reinstated in case of ban from AdSense.

Speculation, based on Google statements, or based on actual data points?

[joelgreen]

Speculation, based on Google statements, or based on actual data points?

Just my thought, that is why I added "IMO" at the end of my previous post :) Provided my GA login, and they never asked me for server logs (when my account was suspended). Not sure if that can be considered "based on actual data" (do not even know if they analyzed GA data).
Also from GA they could easily see traffic sources. Think one would have better chances of being reinstated if Google sees traffic is mostly organic or coming from good sources.
Having GA account means I have nothing to hide from Google.

[Shaffer]

Hello,
that's really interesting.
I mailed my freind about it since he is reading all of the Google AdSense "contract" with users policy, terms of user, etc. Maybe they mestioned taking addmissions and it went out of hand. Maybe the said that, incoordinatedly, they would be allowed to beta test and account, and you were thier "pray".
Anyways, being from Israel, all I can say is that I am proud to see my home language appearing here and there, rather randomly. :D

Shaffer.

p.s
You can go to the policy and terms of use release notes and ctrl+f the following occurances: beta, addmissions, stuff like that...

[netmeg]

Like someone else in this item, I removed GA at first because it slowed the load time. After they said to move the code to the bottom of the page, I put it back and have been very happy with the results. My overall earnings across all my sites are up 150% to 200% over this time last year (and I think I added the GA code in October or November); I don't think GA caused my earnings to go up, but I don't think it's hurting any of my sites any. I haven't seen any differences in relevant ads at all.