Forum Moderators: DixonJones
I could tie the unique visitor determination to the user agent as well as the IP address, but I don't want to set that up to correct a tiny margin of error.
- Grant
Would you consider a unique IP address to be a unique visitor for the purposes of site analytics?
Depends on exactly what those "purposes" are, how much traffic you get, and how evenly distributed the source of that traffic is.
If the purpose is just to make relative comparisons of unique visitor counts (e.g., "what's the ratio of this month's visitor count to the same month last year), that's OK. If the purpose is "I want to detect an attack and automatically block IP addresses", then that's not OK.
The lower your overall traffic volume is, the more likely it is that the approximation of "IP Address == Unique Visitor" will be fairly accurate. If you have a million visitors per day, it may be a fairly flawed estimate. If you have 100 visitors per day, there's a good chance it will be highly accurate.
Finally, consider the source of your traffic. If you sell something targetting AOL users, then your traffic source will be skewed towards a place that shares IP addresses to a very high degree. That could render the approximation of "IP Address == Unique Visitor" a poor one, even for relatively low traffic volumes.
Like many web stats, measuring unique IP addresses can be very useful and informative, even though it is unlikely to be a 100% accurate indicator of what you're really interested in (unique visitors). The trick is to understand when and why the inaccuracy arises, and interpret the numbers accordingly.