Google Analytics Length Of Visit Really Low: Any IIdeas

GA Says 75% of Visitors Leave Within 10 Seconds...

[Planet13]

Hi there, Everyone:

I just noticed that in google analytics the overwhelming majority of visitors (slightly over 75%) have a visit length of under 10 seconds.

Is this metric reliable? And what would the average be for most ecommerce sites?

When I look at visitor length in Stat Counter, it shows that around 70% of visitors leave within 5 seconds!

What happens if someone comes to a page and looks around for a while but then instead of going on to another page, closes the browser of types in a URL to some other site (or hits the back button to go back to a search engine)? Will that skew the Visit Length?

This is starting to worry me...

[enigma1]

Typically means it's not the page they're looking for.

You should be looking to increase the focus of your pages. Some pages will contain general content and you cannot do much about it other than restricting spider access (eg: privacy, tos etc). But the pages that are important for sales should be tuned to expose specific information about the items you sell.

So if the visitor searches for red widgets and ends up on your site because you have some generalized widgets, he will likely leave right away.

You should also check details of the exit pages because the bounce analytics show, is a general graph at the top and you need to see each page's details underneath. What's the bounce of the checkout page or shopping cart page for instance?

And yes if someone hits back on his browser that's the exit point.

Also there are no reliable bounce figures for ecommmerce stores assuming you have the right page focus, because the merchandise and/or type of store, changes pretty much everything. Could be 10% or 50% or 90%. What type of products you carry? is it a niche? something custom? or something way too common. But the pages that you rely on to make sales is what you should be looking for to improve. Check those for the bounce rate.

[Planet13]

Thanks for the tips, enigma1;

I will try to find out which are the ones with the lowest time on page. It doesn't look like there is an EASY way to do this in analytics though...

and of course, with google not showing keyword referrals for users who are logged into their google accounts, that is going to make improving the user experience that much harder...

[Planet13]

Hi again, enigma1 et al.,

Typically means it's not the page they're looking for.

Well, I am kind of baffled by this.

I have looked through the logs in StatCounter for visits that were under 5 seconds, and in my opinion, the pages that they land on ARE the pages that they are looking for.

For example, some visitors use the keywords "history of japanese widgets" and landed on my history-of-japanese-widgets.html page. I mean, it is an exact match for what they want. And the text is on topic. There are no banner ads at the top, no pop up ads, nothing like that. They have to scroll down below the fold to even see anything that remotely looks like an ad.

It's not the most beautiful page in the world, but I would imagine it would take someone at least 5 seconds to decide whether it is a well written page or not.

Another example, someone apparently searched for our site by typing in my-domain.com in the google search box, and landed on our home page. Our domain name is kind of long and unique, so it wasn't like they made a typo.

Now why would they type in the domain name of a site if they didn't want the home page of the site - especially since in the google SERPs it is pretty evident that the link is to the home page of the site.

This is really a mystery...

[lucy24]

Can they even track how long visitors spend on a single page? Last time this subject came up, the impression was that most javascript-based analytics programs don't say. (It's easy to do, but not by means of a single piece of code within the page body.) They only track when someone arrives on a page, so any single-page visit counts as zero seconds :(

[enigma1]

If they search for the exact domain name they may not be interested on the site's content. You would think if someone looks for examples he would type-in example not example.com. Or he may go directly to example.com if he's been there before.

For the other page is this something you see on various pages with many visits. Exact keywords match and they leave in few seconds? Any problems with the html and different browsers? Is the page content unique or it contains some text too common the visitor already knows about it.

I also don't know if the statcounter tracks re-entries. Sometimes I would read a page, then search again and re-enter the same page and then I will leave right away as I already read the description. It's not something that happens often though.

Can they even track how long visitors spend on a single page?

Assuming js and cookies are accepted by the browser, then yes. The first request won't be able to track anything, because it's the earliest the cookie can be sent from the server but with ajax you could make various requests from the client to the analytics server once the page is loaded and therefore tracking cookies could do it.

Now on the server end if the code never sees the ajax request, it will not record anything and may imply the page was never loaded successfully. The few lines of js analytics code you integrate on an html page can load other code to do the job but of course each js analytics program maybe different and not coded in the way one would expect. For instance if they record the very first click but don't verify it with the exit you won't know if the page was properly loaded.

[AnkitMaheshwari]

[services.google.com...] As per video GA do not counts the time for last page visit by user.

[enigma1]

Yes but if I understand it right, they also exclude the one page access from the ga calculations and listings.

[Planet13]

Hi again, Everyone:

First, I forgot to mention that I am NOT using the asynchronous GA code. Will that make a difference at all? I am using the previous GA code (not the old urchin code).

@ lucy24

Can they [stat counter] even track how long visitors spend on a single page?

That seems to be my impression, too, that they AREN'T tracking single page visits.

From the Stat Counter FAQ page:

Visit length is determined by the movement of visitor from one page to another so if visitor visits just single page then visit length displayed is 0 seconds. Also if visitor visit couples of webpages and moves to a different site then we cannot record how much time he spent on last page and hence time displayed is in several hours.

@ enigma1

For the other page is this something you see on various pages with many visits. Exact keywords match and they leave in few seconds?

Yes, this happens on many pages with MANY visits. Oddly enough, the pages with the highest traffic, and lowest bounce rate also have the highest number of visits under 5 seconds. And overwhelmingly, the visits are using keywords that related to the page content, and they are coming from countries where the main language is English (which my site is in).

Any problems with the html and different browsers?

I have tried on different browsers and asked some friends to see if there are any problems, and we haven't had any. It looks fine on my Android phone using Dolphin browser, but I haven't tried any other browsers on Android. That will be my next step. But smart phones and tablets make up less than 10% of visits, according to GA.

Is there any site out there where you can plug in a single URL and it will tell you if your page will have problems loading on various browsers?

Is the page content unique or it contains some text too common the visitor already knows about it.

The page content is unique (at the time it was written at least - now there are many articles on wikipedia and ehow that have "stolen" the material and just spun it, and cite my page as a reference). I would believe that it would take someone AT LEAST 5 seconds of reading to determine whether

[Planet13]

@ enigma1

Yes but if I understand it right, they also exclude the one page access from the ga calculations and listings.

From the video linked above:

For average time on page, bounces are excluded from the calculation.

...In other words, any time on page of 0 is excluded from the calculation

...for average time on site, bounces remain a part of the calculation...

So does this means that in GA, the Visitors -> Visitor Loyalty -> Length Of Visit metrics have NOT been weighted the same way that the Average Time On Page metrics have been weighted (i.e., having bounces and exits removed from the calculation)?

Even with that weighting, my metrics are NOT spectacular: 72% Bounce Rate and 1:08 Average Time On Page for my Top Content.

[enigma1]

I believe the exit pages are recorded. But the exit pages are not in the bounce or time on page calculations. That's for ga.

Now one minute on the important pages means the visitors did spend some time looking at the content. What's the content's length on these pages, I mean whats your expectations how long they should stay on the same page? Also what's the next page target? Is it the shopping cart, a checkout page? And see the bounce rates of the shopping cart or checkout pages. You should expect very low bounce rates on these.

Then check the page loading speeds. Are the pages loading fast enough? You could use some anonymous proxies from different geolocations, if you target different countries, for testing purposes, see if there are latencies. GWT displays the times the bot takes to parse pages, check the average and check for load spikes if there're any. Analytics also has similar figures. Are there any latencies recorded?

[onlineleben]

Could you just split your unique articles into 2 pages, so that the have to click to go to the second one which triggers a new GA load and records the end of the first page visit? - You just move the problem to an other page but you would get feedack about how long people really stay on these articles (at least on their first page).

[Planet13]

Could you just split your unique articles into 2 pages, so that the have to click to go to the second one which triggers a new GA load and records the end of the first page visit?

Well, that is a definite possibility. Unfortunately, most of my traffic is from organic search, so I am a little concerned about how well I will be able to get pages to rank after I split them up...

I will definitely think about it though...

[lucy24]

I wouldn't do it. They'll just read one page or the other, whichever one comes up in their search. Minus a bit for the people who searched for two or more words in conjunction, and the words ended up in separate halves of the article so they never arrive on your site at all.

[Planet13]

Thanks for the opinion, lucy24.

Gives me more to chew on...

[RhinoFish]

if you're using certain redirects (that drop session or referrer data), they might also mess up page-to-page hand-offs, yielding a high bounce rate.

back in August, G changed the way they handle sessions:
[analytics.blogspot.com...]

good read for understanding these issues. plus, it'd be interesting to see if your high bounce rate changed when they changed session handling.