Traffic drop after analytics troubles

[DirkC]

Hi

For two weeks we had an artifical low bouncerate & high pageviews/visit in our Analytics reporting. The day we corrected the bug in Analytics - our bouncerate & pageviews/visit returned to normal levels - however we saw our search traffic go down massively(-50% in sessions).

The bug in the Analytics was caused by a second Analytics tag which was included in an external script which interfered with our own tag.  The drop in traffic occurred just after the removal of the second script (which was only on our site for two weeks). We didn't touch our own tagging - and there were no technical changes on the site during this period, and there is no seasonal trend which could explain the sudden drop of traffic. We double checked our tagging - and the analytics tag is present & working on all the pages of our site.

On the organic traffic report from analytics you can clearly see the when the troubles with analytics started & ended (artificial low bounce rate) - and that the traffic drop starts right after the reporting issue ended. Webmastertools also indicated a lower number of views/clicks, but not to the massive 50% drop.

Is it possible that Google uses the measurements from Analytics for it's SERP's? Or should there be another reason, and where should we start looking? Appreciate your help!

e5GmoMM

[DirkC]

To be honest - I knew the media agency did a change - so I thought the script was removed. Checking the source again shows that they still have the tracking tag - but with a modified tracking code - the modified script is here http://bit.ly/1qmtdWg - our site is http://bit.ly/1D0weRQ

I don't see interference on the measurements, but if I look at the Google Analytics reference, I'm not sure if this is the proper way to implement two tracking codes on a page, but I am in not really an expert in javascript or advanced analytics configuraton.

[Linda-Vassily]

Pageviews will double and bounce rates will drop any time you have two tracking codes on a page (each pageview is seen as two, thus automatically making it not a bounce). Did you just remove the extra Analytics tag, or get rid of the whole script?

[DirkC]

It was a script from the company which is selling the ads on the site (a medium sized media agency in Germany). It included a Google Analytics tag from another account - which was called during the loading of the page. Not sure if it was the script itself, or the way it was called, but there certainly was interference. I noticed using the Analytics tag assistant that the tag which was used was not ours, and when the page was loaded it first became green (our tag), than blue, and then green again (external tag).

Direct result of the script was that time on page dropped dramatically, but pageviews/visit doubled & bouncerate was cut in half.

[Linda-Vassily]

It doesn't look like a coincidence to me either; I just don't think it is a direct correlation of the increased bounce rate and the decreased traffic. What was that external script that added the extra Analytics code?

[DirkC]

The (search) traffic didn't increase when the problem with analytics started (of course the page views did) - as you can see in the graph. There is a drop the moment the stats (bounce rate & page views/visit) change in the negative sense.

Quite possible it's a coincidence, and that something else is causing the drop, but I don't really see what it could be.

[Linda-Vassily]

I would think that if one could substantially increase the traffic of a site by artificially decreasing bounce/increasing pageviews using a second code, we'd see a lot of double-tagged sites out there...

[DirkC]

Always possible - however we never did any linkbuilding, and we never used external SEO agencies that could have done it. The link profile seems pretty normal, no links from spammy domains like blogger, blogspot, or other fishy domains.

The profile is quite comparable with similar sites within our company, and we don't see any changes at this date on any of the other sites.

[KevinBudzynski]

It's probably more correlated with the Penguin 3.0 refresh that have been rolling out. I'm guessing that it's just a coincidence w/the GA issue.

[DirkC]

Thanks - will check the weblogs as well.

I know that the normally there should not be a link between Analytics & SERP's - but it's the second time that I notice that a glitch in Analytics is linked to a sudden drop in traffic so I am getting a bit suspicious about it (the previous time was on another site 2 years ago)

[KevinBudzynski]

According to Matt Cutts, the use of GA doesn't hurt/help your site in rankings. Do you have access to your web logs? If so, do an analysis there and see if it replicates what you saw on GA.