Page 1 of 1
Serenedipity Statistics
Posted: Mon Aug 07, 2006 5:09 pm
by temercer22
I am trying to figure out how Unique Visitors are determined. Is it based on unique IP addresses and therefore guarentees no duplicates, right?
Re: Serenedipity Statistics
Posted: Tue Aug 08, 2006 10:19 am
by garvinhicking
Hi!
No, they are determined via a HTTP Cookie! IP address can be ambiguous if people use a proxy (like AOL users).
Best regards,
Garvin
Posted: Tue Aug 08, 2006 3:37 pm
by temercer22
OK, thanks! But is it safe to assume that each unique user truly is unique?
Posted: Tue Aug 08, 2006 3:42 pm
by garvinhicking
Hi!
Sadly, in web statisticts nothing is every truly safe
Even with cookie tracking it might be, that a browser is rejecting cookies. This would then make the statistic software believe that everytime he accesses the blog, he is a new user.
Or somebody might fake his cookie ID, which would also skew the statistics.
This problem is inherent with all web-tracking software, it also occurs with AwStats, Google Analytics and whatever. You just cannot track every visitor in a unique way, because HTTP is a stateless protocoll.
Best regards,
Garvin
Posted: Tue Aug 08, 2006 3:49 pm
by temercer22
That makes sense.
However, as I have a client to answer to, what would you estimate is a safe percentage to identify as unique users? Do you think I can say 80% of our reported number of unique users truly are unique?
Posted: Tue Aug 08, 2006 3:55 pm
by garvinhicking
Hi!
Usually, what one does is to analyze both unique visitors PLUS the Hits. If the Visitors value goes near the amount of hits, you can assume that many of your visitors reject cookies.
Also you can't make a broad assumption as goes to percentage, because it is based on your visitors. If you run a news page for the common public, you can usually assume that 80% of your visitors are using and allowing cookies, and thus you can trust the numbers of unique visitors by about 80%.
But if you, say, run a security page about trojan viri, your users might be very special and 80% of them reject cookies, which would make the whole numbers not add up.
In the end, never trust a statistic you didn't fake yourself. The numbers are often up to interpretation, and you need to match it with your experience of your visitors (by using polls, browser statistics, additional Google Statistics or AWStats).
Best regards,
Garvin
Posted: Tue Aug 08, 2006 4:15 pm
by temercer22
Yeah, our AWStats reports 672 unique visitors, while Serendipity reports 579. Not a huge difference, but enough of a gap to raise an eyebrow.
Posted: Tue Aug 08, 2006 4:18 pm
by garvinhicking
Hi!
Ah -- AWStats does not set cookies, it calculates "visitors" based on IP and on the duration of their visit.
So, actually, something around 600 should be realistic.
Best regards,
Garvin