awstats vs. wusage vs. webalizer

How to account for variance in stats when using the same data?

[jetnovo]

I've shifted a site from one host to another. The old host used Wusage stats, which worked fine. The new host simultaneously runs both Webalizer and AWStats.

The problem is that on the new site, the stats (in particular, "unique visitor") from Webalizer don't match those of AWStats for the same period. There's roughly a 10% difference in the figures.

Compounding things further is that the Wusage stats from the old host seem to be nearly double what the numbers are for AWStats and Webalizer on the new host.

Obviously the client is quite concened as he spends a lot on his online marketing budget and has to report on "unique visitors" to the company directors. And on th face of it, it appears that in shifting to the new host his traffic has nearly halved.

Can anyone suggest

a) why the difference between AWStats and Webalizer, given the same set of data?

b) how can I explain the drop in traffic because of the change of hosts?

One further-complicating fact, I asked the client to provide me with his records of "hits" (since I thought maybe that would be more objective than "unique visitors"), but he only kept "unique visitor" stats from the old host - so I can't compare "hits" records.

[cgrantski]

Hiya, this has been talked about elsewhere here, though I don't think for these exact stats programs.

The main sources of difference are usually things like different definitions of visitors, or different ways to identify visitors, or different ways to count visitors.

One of them could be counting spiders as users, while the other filters them out and just tries to count humans.

One could be using cookies to identify a visitor while the other relies on IP address or the concatenation of IP address and User Agent.

One of them could be figuring out "uniqueness" within a 1-day period (3 visits by same user in one day = 1 unique visitor) but not within a week (same person does 3 visits 2 days in a row = 2 unique visitors) while the other program could be recalculating on a week period (same person does 3 visits 2 days in a row = 1 unique visitor). (go back and substitute month, quarter, year in the above for all the possibilities).

And so forth.

I assume your old hosting outfit did not save any log files either, not even in archives or backups? You could get an exact comparison by letting somebody crunch a week of logs from each place.