Help Measuring 'Total Visitor' Number To Site

Thanks Meganom

I think I am clear on the definition of a 'Unique Visitor', say during a month. 

Yes, I concurr with your definition, but it is the knowledge of how many times an individual visitor comes back during a month which is crucial.

Using the illustration in your message, if somebody comes back 10 times in a month we need to know this - not just that they visited once during the month. In theory, this means we are 'missing' 9 visits (inclusive of the pages they visited, the information they viewed, whether they made an enquiry after their 3, 4 or 9th visit etc). As well as their page views during their visits, we need to know how many times visitors came back to view the information.

This not only has implications for clients who have complex information (which may require multiple views before an enquiry is made) but also in finding out which visitors (who may be key targets for clients) came to the site over the course of a given time period and how many times they came.

In an ideal situation, we would need to know:

A visitor from www.customerdomain.com visited the site on the 1st of the month - this of course identifies them as a 'monthly unique', but also a 'new visitor' because its their first time ever visit to the site. It's important to know of brand new, fresh visitors to the site.

If visitors from the same domain then come back during the month, we need to know how many times they re-visit- logging them as 'repeat visitors'.

This kind of information (along with other stats) has impact for the effectiveness of a variety of online marketing and offline attendance at events, PR activity etc

I think, from my enquiries so far, that this kind of intelligence is available, so I don't think I am asking for is anything ground-breaking.

Regards

On 11:18:33 17 October 2005 Meganom wrote:

 

I use different vendor, so I cannot be absolutly sure, but let me explain as I know it from my vendor experience. Unique visitors for particular time frame are ALL visitors that bothered to show up on your site during that time frame, regardless of how many times they viewed your site during that time frame or before. If someone, say, viewed 10 different pages on your site i.e. had 10 views during that time frame (as you note), that visitor is still counted as 1 unique visitor. The term 'unique' in this context means NOT that visitor is unique for that time frame (i.e. not visited before), but simply means that visitor was identified i.e. was distinguished by system from other visitors. 

New Visitors for that time frame shows the portion of Unique Visitors who have not visited your site before that time frame. Subtracting New Visitors from Unique Visitors will give you Returning Visitors.

Hope this helps.

On 09:07:39 17 October 2005 Markd wrote:

 

Meganom

A number of things here:

Firstly, I expect vendors to actually help me (firstly as an existing/potential client, secondly as somebody who freely admit to providing 'subtly incorrect definitions) in fully identifying information requirements. I believe, albeit in 'laymans terms', to have fully explained our - and our clients - requirements with regards to visitors to our current provider and others we are investigating.

Secondly, if you read my original enquiry, we ARE attempting to find out about visitors for a particular timeframe - such as a month, week, quarter etc.. Actually it is possible to distinguish between visitors who may have both 'unique' in a particular month (but have visited the site previously) and those who may well have been both 'unique' that month and 'first time ever' visitors. 

What we are struggling with our current system is to measure 'repeat visitors', again using a month as an example. If 100 visitors are 'unique' that month, but return 10 times after their first unique visit during that month, the subsequent visits are not accurately measured.

When a client comes to my company looking to buy services or for assistance, we try our best to help regardless of how 'subtle' or otherwise their knowledge (either of the matter at hand or any technical solutions we may provide) and do our best to help. This approach is magnified by a factor of a 100 if they are an existing client and provide us with business.

On 22:29:57 16 October 2005 Meganom wrote:

 

There is no problem at all. You just give subtly incorrect definitions. Your "New Visitors - those who have come to the site for the first time ever" should be suplemented with more specific "in this month", cause absolutly every visitor of your site has sometime come there first time, and there is no need for such over-agregate value. But with that addition, it becomes indistinguishable from your first definition of "Unique Visitors - those who have visited for the first time in that month", which is incorrect. Actually, the Unique Visitors is the very figure that you need, as one might expect. The only non-trivial thing here is how easily your vendors got confused over such a straightforward matter.

On 21:39:15 16 September 2005 Markd wrote:

 

Hi all
I wonder if you could offer us some advice on a problem we are having measuring a 'total' number of visits to a site.

Currently, we are using a 'tagged' statistics system - adding Javascript to the web page which sends info on a 'real time' basis to a server. It's OK on most things, but doesn't seem able to give us a 'total' number of visitors for a given time frame, say a month.

So what we get is:

Unique Visitors - those who have visited for the first time in that month

New Visitors - those who have come to the site for the first time ever.

But what it doesn't seem to log is any subsequent (repeat) visits by those who may have come back after their first (Unique) visit in that month.

The implications of this are that if a site gets 100 Unique Visitors in a month, and those visitors come back 10 times during the same month, approx. 1000 'repeat visits' are not logged at all. Obviously, this makes a 'total visitor' number impossible to measure.

Is this inherent in all 'tagged' based stats programs? I have contacted the Tech Support of the makers of the package we are using in case we are misinterpreting the stats, but they aren't the best Support we have encountered and are not being very helpful.

I know that Web Trends 'on demand' package can break visitors down in this way, but it's really expensive, so any other recommendations are welcome. 

Most of our sites are B2B and information based - so there is not really a need for the advanced e-commerce/purchasing analysis that I have seen in some stats packages.

Hope you may be able to help.

 

 

 

All good points, but...

Our clients need to get information both on the number and status of their visits and the number of pages they view on each visit.

It is of little use to know that you had 10,000 page views in a month, but not to know how many visitors (and their visitor 'status')  were responsible for viewing this information.  For example, without intelligence about visitors how can you know how these 10k views were viewed?

Were they viewed by relatively few visitors who came to the site many times in a month, what proportion were viewed by visitors on their first ever visit to the site and how many were viewed by visitors who were unique in that month but had visited the site before. 

How can you effectively structure your site content, marketing activities etc. if you only know about how may pages were viewed, but have little or no knowledge of the visitors which viewed them?

Surely, 10k views by 1k visitors surely has drastically different implications to 10k views by 20k visitors - that's before you take into consideration their 'visitor status'?

Am I really wide of the mark in expecting this level of visitor intelligence? Perhaps there are other 'wise people' out there who could comment.

Thanks again.

On 17:51:43 17 October 2005 Meganom wrote:

 

There should be a distiction made between pageview and visit. Visit is a set of pageviews reasonably close in time e.g. pageview that come in say 5 min may be considered part of the same visit, but those separated by days belong to different visits. How visits are distinguished is up specific vendor/software, and most simply do not distinguish visits and pageviews.

The information you ask for may indeed be available. For example, my vendor provides Sessions report that shows detailed information on all pageviews/visits of all visitors for specified arbitrary time frame grouped into browsing sessions (i.e. visits). But if trafic is intense, such detailed report would be too bulky to be informative, so you would need to have some agregation. My vendor has agregate report that counts pageviews/visits as well as visitors, so that you can see that during some time frame say 100 visitors visited your site and made 1000 pageviews, so that on average each made 10 pageviews. If vendor separate pageviews into visits, than you may have number of visits, say 200, so that on average each visitor made 2 visits, 5 pageviews each.

If you want to integrate orders into that analysis than orders also need to be reported.

It may well be even that your specific information needs cannot be met by existing software, but often may be aproximated with views per visitor or visits per visitor or detailed session data for each individual visitor or a like...

On 12:05:41 17 October 2005 Markd wrote:

 

Thanks Meganom

I think I am clear on the definition of a 'Unique Visitor', say during a month. 

Yes, I concurr with your definition, but it is the knowledge of how many times an individual visitor comes back during a month which is crucial.

Using the illustration in your message, if somebody comes back 10 times in a month we need to know this - not just that they visited once during the month. In theory, this means we are 'missing' 9 visits (inclusive of the pages they visited, the information they viewed, whether they made an enquiry after their 3, 4 or 9th visit etc). As well as their page views during their visits, we need to know how many times visitors came back to view the information.

This not only has implications for clients who have complex information (which may require multiple views before an enquiry is made) but also in finding out which visitors (who may be key targets for clients) came to the site over the course of a given time period and how many times they came.

In an ideal situation, we would need to know:

A visitor from www.customerdomain.com visited the site on the 1st of the month - this of course identifies them as a 'monthly unique', but also a 'new visitor' because its their first time ever visit to the site. It's important to know of brand new, fresh visitors to the site.

If visitors from the same domain then come back during the month, we need to know how many times they re-visit- logging them as 'repeat visitors'.

This kind of information (along with other stats) has impact for the effectiveness of a variety of online marketing and offline attendance at events, PR activity etc

I think, from my enquiries so far, that this kind of intelligence is available, so I don't think I am asking for is anything ground-breaking.

Regards

On 11:18:33 17 October 2005 Meganom wrote:

 

I use different vendor, so I cannot be absolutly sure, but let me explain as I know it from my vendor experience. Unique visitors for particular time frame are ALL visitors that bothered to show up on your site during that time frame, regardless of how many times they viewed your site during that time frame or before. If someone, say, viewed 10 different pages on your site i.e. had 10 views during that time frame (as you note), that visitor is still counted as 1 unique visitor. The term 'unique' in this context means NOT that visitor is unique for that time frame (i.e. not visited before), but simply means that visitor was identified i.e. was distinguished by system from other visitors. 

New Visitors for that time frame shows the portion of Unique Visitors who have not visited your site before that time frame. Subtracting New Visitors from Unique Visitors will give you Returning Visitors.

Hope this helps.

On 09:07:39 17 October 2005 Markd wrote:

 

Meganom

A number of things here:

Firstly, I expect vendors to actually help me (firstly as an existing/potential client, secondly as somebody who freely admit to providing 'subtly incorrect definitions) in fully identifying information requirements. I believe, albeit in 'laymans terms', to have fully explained our - and our clients - requirements with regards to visitors to our current provider and others we are investigating.

Secondly, if you read my original enquiry, we ARE attempting to find out about visitors for a particular timeframe - such as a month, week, quarter etc.. Actually it is possible to distinguish between visitors who may have both 'unique' in a particular month (but have visited the site previously) and those who may well have been both 'unique' that month and 'first time ever' visitors. 

What we are struggling with our current system is to measure 'repeat visitors', again using a month as an example. If 100 visitors are 'unique' that month, but return 10 times after their first unique visit during that month, the subsequent visits are not accurately measured.

When a client comes to my company looking to buy services or for assistance, we try our best to help regardless of how 'subtle' or otherwise their knowledge (either of the matter at hand or any technical solutions we may provide) and do our best to help. This approach is magnified by a factor of a 100 if they are an existing client and provide us with business.

On 22:29:57 16 October 2005 Meganom wrote:

 

There is no problem at all. You just give subtly incorrect definitions. Your "New Visitors - those who have come to the site for the first time ever" should be suplemented with more specific "in this month", cause absolutly every visitor of your site has sometime come there first time, and there is no need for such over-agregate value. But with that addition, it becomes indistinguishable from your first definition of "Unique Visitors - those who have visited for the first time in that month", which is incorrect. Actually, the Unique Visitors is the very figure that you need, as one might expect. The only non-trivial thing here is how easily your vendors got confused over such a straightforward matter.

On 21:39:15 16 September 2005 Markd wrote:

 

Hi all
I wonder if you could offer us some advice on a problem we are having measuring a 'total' number of visits to a site.

Currently, we are using a 'tagged' statistics system - adding Javascript to the web page which sends info on a 'real time' basis to a server. It's OK on most things, but doesn't seem able to give us a 'total' number of visitors for a given time frame, say a month.

So what we get is:

Unique Visitors - those who have visited for the first time in that month

New Visitors - those who have come to the site for the first time ever.

But what it doesn't seem to log is any subsequent (repeat) visits by those who may have come back after their first (Unique) visit in that month.

The implications of this are that if a site gets 100 Unique Visitors in a month, and those visitors come back 10 times during the same month, approx. 1000 'repeat visits' are not logged at all. Obviously, this makes a 'total visitor' number impossible to measure.

Is this inherent in all 'tagged' based stats programs? I have contacted the Tech Support of the makers of the package we are using in case we are misinterpreting the stats, but they aren't the best Support we have encountered and are not being very helpful.

I know that Web Trends 'on demand' package can break visitors down in this way, but it's really expensive, so any other recommendations are welcome. 

Most of our sites are B2B and information based - so there is not really a need for the advanced e-commerce/purchasing analysis that I have seen in some stats packages.

Hope you may be able to help.

 

 

 

 

 

Thanks for your help Alcochrane

Initially, we just want to be able to know how many visitors came to a site more than once - not necessarily equating this to the number of pages viewed, although this is the natural progression after gaining visitor information.

Ideally we want to see that in a month a site got 100 'new visitors' (those who are first time ever visitors), 200 'Unique visitors' (which may incorporate both first time ever - as they are 'unique' - plus others who may be unique in that month) and 300 repeat visitors (those that came back more than once in that month, after they registered as unique/first time ever).

It would be great to then look at these visitor classifications and see which pages they viewed on each visit. We could then determine how many times, say a particular product page, was viewed or how many times an enquiry form was viewed before/if it was actioned.

I can say with 100% certainty that our current solution doesn't provide the visitor measurement we would like. I know this because, after repeaded trans-Atlantic phone calls and numerous emails, they have told me it doesn't do this. 

What it does provide is a loose measurement of 'repeat visitors' which uses range of visits for repeat visitors. It may say that 1 visitor came back 2 - 10 times, and this scale goes up to 50 - 100 times. 

Personally, I don't think that this provides nearly enough accuracy in information. For example, it seems to produce a huge gap in accuracy if all you can measure is that a site gets 100 visitors who came back between 50 - 100 times. The margin of error here is too great.

Most of our clients are B2B with no 'e-commerce' functionality. What I find really puzzling, with all the 'bells and whistles' for measuring e-commerce on stats packages, is that the information we require is seemingly so hard to provide. 

I fully accept that my communication of what I need may be far from lucid :), but I would have thought that this kind of information on visitors is fundamental....?

On 08:56:23 18 October 2005 Alcochrane wrote:

 

I don't think that you are wide of the mark on this one:

You should be able to get your tool to do simple reports showing the number of page views per visit (eg 20 visits had 1 page view, 20 visits had 2 page views, etc).  These may get aggregated in your tool (1, 2-4, 4-10, 10+ page views per visit) .

I think what you are asking for is to see if certain sections get more page views on second, third and fourth visits.  Whether you can segment down your above report into one that contains repeat visitors only is a different matter.  Equally segmenting this down into content sections should be relatively simple.  I suspect that the tool will only be able to distinguish between new visitors and repeat visitors however, hence you will only be able to look at the first visit and all subsequent visits.  Again, this is something your vendor should be able to answer for you.

Equally there should be a way of showing the number of visits per visitor over a certain period of time (eg 20 visitors had 1 visit, 20 visitors had 2 visits, etc).  If your tool is powerful enough you may be able to segment this again.

These methods of segmentation (or filtering) will almost certainly be available with your tool (even if they are not obvious).  You just have to discover how you use them and present them into meaningful reports to your clients. 

also, bare in mind with cookie deletion rates as high as 30% per annum. That these numbers must be taken with a grain of salt...

On 10:04:27 18 October 2005 DeriJones wrote:

 

Amazing how complex seemingly simple concepts can become...!

"Ideally we want to see that in a month a site got 100 'new visitors' (those who are first time ever visitors)"

That's pretty clear - your software gives you that OK.

"200 'Unique visitors' (which may incorporate both first time ever - as they are 'unique' - plus others who may be unique in that month)"

That's also straightforward.

"and 300 repeat visitors (those that came back more than once in that month, after they registered as unique/first time ever)."

(is it important for you if that 'first registration' was in the month, or previous?)

Anyway, this is the bit you're struggling with it looks like.  Humm, if you only have 200 unique visitors in the month, then you can't have 300 repeat visitors... there's a max number of visitors of 200.

Instead, it sounds like what you want here is a measure of the number of *visits*, rather than visitors.  Clearly visitors who come back repeatedly mean that the number of visits is always greater than the number of visitors.

Knowing the number of visits/month, plus the unique visitors/month means you can calculate the average number of visits per visitor etc.

Or am I missing the point here...

Deri
www.scivisum.co.uk
web portal KPI testing

Not to muddy the water, but I have also had a query from a client today about logging the actual activity undertaken by a visitor to their site.

The want to be able to know:

that a visitor identified by a domain, eg. www.visitordomain.com, entered the site on a particular page and then took a certain path through the site before exiting.
They would like to do this for each visitor that comes to the site, again in a given time period such as a month, week etc. 

As well as identifying the visitor domain itself, they would like to get a country of origin (either by TLD or by looking at the language used in the browser).

At the moment, our current stats package only lists this info under 'Latest Visitors' - showing only the last 200 visits.

Can well understand that to do this for a high traffic site would create large information files - but is this something which is 'standard' on other packages?

Thanks for all you help on this - I hope others are benefiting from our enquiries!!

On 10:25:37 18 October 2005 Markd wrote:

 

Thanks Deri

No you're not missing the point - it's my choice of 'illustrative figures' which are confusing you!!

As soon as I typed it I realised that '300' would be misleading, as was my use of the term 'regisration' - by which I meant the first time the visitor was logged by the stats system, not a formal registration to use the site or parts of it!!

In a nutshell, it is 'visits' that our clients are interested in - specifically repeat visits which would indicate a number of returns to evaluate information. Obviously, if we could tell that 'visitor A' came back 20 times in a month (which would tell them if particular target companies/domains are visiting)  that would be great, but just to know how many repeat visits were undertaken would be a start.

<<(is it important for you if that 'first registration' was in the month, or previous?)>>
Taking into consideration my (misleading) use of the term 'registration', we want to be able to choose a time period - say the previous month, week, day or quarter - and analyse the visits by 'new', 'unique' (for the time period being examined), and 'repeat visits' (or visitors) to the site. 

Again, this had direct impact to 'spikes' in activity after attendance at events, online marketing campaigns and other activity.

Hope this clarifies things - thanks for your help.

On 10:05:57 18 October 2005 jbovard wrote:

 

also, bare in mind with cookie deletion rates as high as 30% per annum. That these numbers must be taken with a grain of salt...

On 10:04:27 18 October 2005 DeriJones wrote:

 

Amazing how complex seemingly simple concepts can become...!

"Ideally we want to see that in a month a site got 100 'new visitors' (those who are first time ever visitors)"

That's pretty clear - your software gives you that OK.

"200 'Unique visitors' (which may incorporate both first time ever - as they are 'unique' - plus others who may be unique in that month)"

That's also straightforward.

"and 300 repeat visitors (those that came back more than once in that month, after they registered as unique/first time ever)."

(is it important for you if that 'first registration' was in the month, or previous?)

Anyway, this is the bit you're struggling with it looks like.  Humm, if you only have 200 unique visitors in the month, then you can't have 300 repeat visitors... there's a max number of visitors of 200.

Instead, it sounds like what you want here is a measure of the number of *visits*, rather than visitors.  Clearly visitors who come back repeatedly mean that the number of visits is always greater than the number of visitors.

Knowing the number of visits/month, plus the unique visitors/month means you can calculate the average number of visits per visitor etc.

Or am I missing the point here...

Deri
www.scivisum.co.uk
web portal KPI testing