I have a 1 page website, yet GA is recording bounces and time on site?

[smithaa02]

How does this work?

Officially from some google sources and various opinions around the web...the bounce rate is merely the number of times a visitor didn't go more than one page deep into your website.

Yet, on my one page website, google analytics is recording all different types of bounces for the different days. Some are yes, 100% bounce days, but most days are not, and vary on average from 80 to 40 percent.

How is this possible for a one page website?

The other puzzler is that google is recording time spent on my one page site. Again the values vary and obviously for some visitors, GA doesn't have a clue because it records a time of 0 (for visitors of time 0 I have a 100% bounce rate) yet on other days google does record precise and lengthy times on site.

From the googling I did, the official explanation for time on site is that it is GA measure of request time of the last page loaded minus the time the first page was loaded. Yet on my one page website...this should not be possible so there must be another way GA is working?

Maybe if google through other GA accounts/the google toolbar/future searches would know definitely if user X was no longer on my site because there were using another service google was keeping an eye on?

Does anybody know how GA really measures bounces and time on site?

[walkman]

Does anybody know how GA really measures bounces and time on site?

Don't know but they recently changed it too, to add to the confusion. If you have a one page site, I don't see how bounce can be anything other than 100% or 0%.

[johnblack]

May be a page refresh skews the bounce rate? Two pageviews but one visitor?

[lucy24]

Time on page is not hard. You can do it yourself with javascript: run your time-checking functions "onload" and again "onunload".

[tangor]

Google is the bean counter of the internet. Created the biz so they own it. :)

(Heck, I don't know! walkman has it right in real world terms!)

[lostdreamer]

+1 for john black

I also see this on some of our websites, pages used in mail campagnes where the bouncerate should be 100% since they have no other links.
When I used a piece of javascript to check what people where actually doing on the page, the hits that had a "Time On Page" set seamed to be hitting F5.

[Dijkgraaf]

The problem is that you can't really measure either how long someone is actually on the page or how many pages they visit accurately 100% of the time.
Time on the page requires either that JavaScript is enabled or make the assumption that if they load another page on your site (and this can include the original one) that they spent all the preceding time on the first page. So if they have JavaScript disabled and either visit the page twice in a short period or reload the page it will have some false results as johnblack points out. Those you are seeing a time on page in GA of zero are probably those that do disable JavaScript or at least block GA.

Identifying that it is the same visitor is also not 100% reliable if there are measure in place such as blocking cookies, giving no User Agent or a uninformative one and changing the IP address (some proxy servers will do all of that). So sometimes someone who does visit multiple pages will seem in the stats to be multiple visitors visiting only one page each.

[perfectionj]

Effective content always reflects bounce rate of your website. The thing which are considered to increase bounce rate are:

1. Make your content impressive.
2. Attractive designing is compulsory.
3. Keep engaged your readers with the widgets like "related posts" or "most popular posts".

[lostdreamer]

@ Dijkgraaf:

Those you are seeing a time on page in GA of zero are probably those that do disable JavaScript or at least block GA.

Those that have JS disabled or GA in their blocklist are the ones that DONT show up in GA ;)

How would you run the GA script with JS turned off or the GA domain in your blacklist ? ;)

[Dijkgraaf]

Those that have JS disabled or GA in their blocklist are the ones that DONT show up in GA ;)

GA doesn't have a noscript tag that uses an web bug image? I'm surprised, most other tracking systems do.

How would you run the GA script with JS turned off or the GA domain in your blacklist ? ;)

As above, you don't run the script but use a web bug image ;-)

[lostdreamer]

I checked to be sure, but it seems GA doesnt have a noscript code (there are hacks for it in PHP / python etc. but not in pure html)

Also: Though this would work with JS turned of, someone with the domain in their hosts file, or simply a plugin in firefox like noscript will never be visible in GA.

[smithaa02]

Wouldn't it be possible for google to keep track of this if they had a profile (yeah this seems conspiratorial) of each user on the internet based on their profile/ip/cookies/gmail account/google account. So if they visited another site on the internet that also had GA or if they returned to the SERP's, google's internal db could say users 323423 just went from widgets.com/pagea.html to example.com/pagex.html...therefore logically, the time on site could be no more than the time difference between the two?

[lostdreamer]

and what if we have multiple tabs / windows open?
Time on site can allways be more then the time between two seperate page calls.

And... ehm.... I'm pretty sure google is doing everything it can to keep up-to-date profiles on all internet users... Just like Facebook and all others who try to get webmasters to put 3rd party JS files on their websites.