Originally Posted by KlenTelaris

Well i spent a lot of hours researching counters and concluded following -none of them is 100% reliable.
And there is two types of server based counters - the one which uses their own tracker(by javascript,php include,whatever)and they need to be manually inserted into every page which you want to track.
Advantage of such script is how they have a lot of features.
Script like this are : piwik,crawltracker,crazystats,etc...
Other type is where script simply read logs of your httpd server(apache,nginx,whatever you use)but they are kind poor when it comes to features,and also while GA having problem with registering correctly hits,these one can show more hits then it's actual number.
Script like this are:webalizer,awstats

Normally i would recommend to use piwik,which is in counting world same what is wordpress in blogging world,but problem with it how is resource hug,so you need to be careful if you have a lot of sites - add site by site,see how piwik is behave,and then do further optimization,like settings optimization or even installing it on separate vps/server which wont be used for anything else.So it can work fine for you if you properly set it.

And when it comes to alternative to piwik,i used for some time crawltrack which is great because it doesn't store entire useragent of visitor into database,it store synonym instead,so for example if user browser is something like firefox 24 blabalbalglfdlgfdlgfdgldf gr9ut93ugf 9gdgrfd it will simply store as firefox and as result that will much less usage of database.But that script last release was in 2011 and i kind got impression how it have some bugs which is why i at the moment hesitate to use it.

So what i will do,probably take either piwik or crawltracker and modify them till they will work properly.
Or maybe even modify a trade script,since trade script already log same info as counters and that way i would avoid double writing of same data.