Unique Visitors bug?

[Paul Reading]

What is wrong I have had 4 visitors on my site this morning but StatCounter records 19 and 4. Is this a bug?


http://www.cp-lighting.co.uk/images/bug.png

Note the dates:-

http://www.cp-lighting.co.uk/images/bug2.png

[webado]

If they have cookies disabled (or third party cookies disabled) they don,t get recognized as returning and every pageload will be counted as coming from a new visitor.

[Paul Reading]

I have checked this and aparently cookies are not dissabled.

[westmoors]

Well from what you are displaying I would say the visitor from Manchester, England who entered your site at 10:57 had cookies disabled. Their 16 pageloads plus three other visitors equals your 19 uniques shown on the summary chart. So where is the mystery?

[Paul Reading]

Is this a new thing? My ecommerce software has it's own counter and it says that I am getting between 40-100 unique visitors a day, Statcounter says that I am getting 300-500 per day over the same period. I find it hard to believe that there can be so many people with cookie blockers, but given the above it would appear that the Statcounter figure is the least reliable of the two counters.

If Statcounter is so badly wrong, has it always been so? Is it likely to be fixed or is that not possible, in which case how is it that my ecommerce software seems not to be effected by the problem?

At the end of the day I just want reliable statistics.

[Paul Reading]

What I have realised is that I have spent a lot of money trying to figure out how to get the 3-5 hundred visitors a day to buy, the conversion rate was something like 1/250-350 visitors. If the stat counter os wrong then it makes my conversion rate more like 1/50 2% which is much better. 2% is not as high as I would like but I do not know what to do about that.

[westmoors]

The summary stats are based on cookies. Many people do have cookies disabled or at least third party ones, which the statcounter one is when people visit your site. What do your detailed stats say? The detailed stats are collated by IP address and may give a better reflection on the number of 'unique' visitors. However, if a visitor does not have a static IP (eg AOL users), then each pageload will show as a different visitor. Between the summary and detailed stats you should be able to determine how many visitors you have had.

As for your ecommerce software, I cannot comment. How does it track? Does it track all the pages that your statcounter code does?

[Paul Reading]

I don't know how it works, but in the example given at the top of the page 4 actuall vistors are recorded as 19, this is close to the 4-1 difference between the ecommerce software and Statcounter. Do all counters work the same way as Statcounter?

[westmoors]

What you showed at the start of this thread is all from statcounter. The detailed stats (based on IP address) showed 4 visitors, the summary stats (based on cookies) showed 19 visitors. The detailed stats show the number of pageload each visitor made as 36, 1, 1, and 16. Simple maths tells me that the last of these must have had cookies disabled since 16 + 1 + 1 + 1 = 19. The visitor with 36 pageloads had cookies enabled as they only got counted as a single visitor in the summary stats.

When you say there is a discrepancy between SC and your ecommerce tracker are you only looking at the SC summary stats or are you analysing the detailed stats as well?

[Paul Reading]

All I am interested is in counting the number of unique visitors I manage to attract to my site. However, because of the cookie issue the unique visitors stats collated by stat counter grosly over estimate the numbers.

Do all counters like Stat Counter rely on cookies or is there a more reliable methos of calculating the unique figures. If it is a matter of countingthe ip addresses that must be bore accurate than the ciookie methos despite the fact that AOL mucks about with IP addresses all the time.

[westmoors]

I don't know about other counters cos I've only used statcounter. What I'm trying to tell you is that you cannot rely solely on the summary stats but need to look at the detailed stats as well. Working from both you should be able to determine the true number of unique visitors as I have already shown you.

Another scenario to explain further:

Summary stats: 7 unique visitors
Detailed stats (recent visitors): 6 unique visitors: 4 AOL visitors with 1 pageload each, 1 visitor with 6 pageloads and 1 visitor with 5 pageloads.

From this I would be able to deduce the following

AOL visitor: 4 pageloads, cookies enabled.
Non AOL visitor: 6 pageloads, cookies enabled.
Non AOL visitor: 5 pageloads, cookies disabled

Since 1 + 1 (cookies enabled) + 5 (cookies disabled) = 7 (summary stats)and 4 (AOL) + 1 + 1 = 6 (detailed stats). Hence 3 unique visitors.

[webado]

Counting distinct Ip addresses would certianly whittle it down to a lower number. Except for AOL and any other visitors who visit repeatedly but each time with a different Ip address - like dial-up user and any other not having a static IP, and these are the majority of internet users I think.

The cookie at least can identify those that are repeat visitors as long as they don't disable or delete cookies.


In the case of Statcounter there's also the fact that all detailed stats are based on the log contents which is fixed in size, containing only the latest X hits (100, 1000 or whatever your log size is). Then it cannot know what has visited before.

There's no 100% foolproof method.