I have following situation I don't understand. I have asked it from Lithium support and they said everything is ok.
Sometimes we have a peaks with visits, we can see peaks in Admin and LSI Metrics but not in Google analytics.
Admin metrics says about the visits:
Google Analytics says:
So yes, there is difference between how Lithium and Google´s metrics works and I understand it.
Usually all the metrics displays almost same numbers, meaning that difference is about 10-15%. But then we have days when the difference is 35-40% and that is something I dont understand, if we have a peak with the visits difference between Lithium metrics and GA should be still about 10-15%?
What´s going on and have you seen same behaviour?
Do you know is it possible to get server log files from Lithium, to deep dive data and analyse these peaks?
For a while we were using Google Analytics and Adobe Omniture so had three analytics platforms (including LSI of course). I found that the daily variance was quite high between the three. Sometimes GA and Omniture would agree (+/- 2 to 3%) and LSI was waaay out, at others LSI and Omniture would agree but GA would be out.
In the end a decision was made to use LSI for Community Managers and Mods to manage day to day tasks but that all reporting to stakeholders about performance would come from Omniture, which was the main system used by our other sites.
Probably not the kind of response you were looking for but that is the approach we took.
Thanks @DanK, yeah that wasn't the answer I was looking for but good to know that we are not only one who have this problem.
Our analytics tool is GA but so far I have reported our numbers with using LSI/Admin.
I have used two different reporting tools on a Community before as well - my history is more with Omniture (as @DanK) mentioned. The new Community that I work for uses GA, but haven't dived into differences or anything just yet.
My situation is very similar to Dan's in that the company used Omniture for all other reporting, so we followed suit and did the same. There were always discrepancies, and some of the slight ones were brushed off with the capability of Omniture to remove internal numbers from the metrics. The analytics person at the time had said something to the effect of "native tools should always trump external tools for reporting" - in essence meaning that LSI is going to be the closest you are going to get to perfection because it is a platform tool. Anything else is third party to that, so major discrepancies should be looked at in this light.
I know this may not help, but you can at least know that it is definitely not an isolated problem.
Good luck in trying to dig around to find the cause of these spikes. If you do find out, let us know (if you can, at least). It definitely gets to be interesting as you dig deeper into the numbers and discrepancies.
Overall, is LSI good enough to report on page views and overall traffic compared to omniture and Google analytics? Need to know how much one can rely on lsi for page views and traffic.