Take the 2-minute tour ×
Webmasters Stack Exchange is a question and answer site for pro webmasters. It's 100% free, no registration required.

I'm working on an eCommerce site and there's a perceived problem with the bounce rates on product pages. The best performing page (in terms of views), tends to have a very high bounce rate, at about 80%. This page is also targeted with Ads, but in both non-paid and paid traffic, the bounce rate stays high.

What's weird about it is that the time on this page is about 2 and a half minutes. The others pages have a slightly lower bounce rate (between 50% and 75%) with visit times between 1 minute and 2 minutes.

I'm not entirely sure about how to analyse this data. I've checked all the possible technical problems (faulty GA code, JS race conditions, breaks in the chain), and whilst there were some things that need improving, no technical fix budged the bounce rate at all.

The other thing we noticed was that the product pages were getting lots of traffic for very generic and brand based search terms. In these cases the pages come up as sitelinks, could this explain a high bounce rate too?

So in short :

If the page view time is so high, should I be worrying about the bounce rate at all? What's the chances of someone visiting the page and then leaving to compare prices? And how do sitelinks affect bounce rate on eCommerce sites?

share|improve this question
    
are you sure your sitelinks campaign is targeting the right set of keywords? –  Prasad Feb 25 '13 at 4:15

2 Answers 2

up vote 3 down vote accepted

Based on my experience and my understanding of how javascript works, the Google Analytics code only records data related to the moment the user first loads the page and the javascript is run. So if a user only looks at one page, they would have a time on site of 00:00:00, regardless of how long they looked at a page. The GA code has no way of knowing when the visitor leaves.

Whenever I isolate my Analytics data to a short time period or other filter and I look at views of 1 visit and 1 visit/per page, the time on site is always 00:00:00. As a general rule, any condition that shows a bounce rate of 100% should also show a time on site of 00:00:00.

Of course, that means in general, the lower the bounce rate, the higher the time on site you will have, as you have fewer 0's bringing down the average. If you have pages with a high bounce rate and higher times on site, this suggests that the visitors that don't leave are averaging much higher page views. In other words, most visitors are leaving right away, but the ones that stay are staying very long.

To verify this, you can apply a filter in GA for Include > Bounce Rate > Less Than > 100. This will give you the true time on site for everyone that does not leave right away, and I suspect that if you compare your 80% bounce-rate pages with your 50-75% pages you will find higher times on site on your 80% pages. But there is no way in Google Analytics to view time on site for visitors who visit one page; they only bring down the average time on site for whatever section they belong to.

As to why this is happening on your site, I can only offer theories. Make sure you are filtering your own visits (usually by IP). I would only "worry" about the bounce rate in that if you can figure out why most of the visitors are leaving, you might be able to get them to stay. Compare the bounce rate by keyword; maybe you are accidentally ranking for non-targeted keywords and your bounce rate is high for those keywords, but lower (with lots of page views and time on site) for the keywords you want to rank for. It's possible that visitors are visiting the page and then leaving to compare prices, but there is no way to know. And I'm not aware of any data regarding sitelinks affecting bounce rate on eCommerce sites.

share|improve this answer
    
So, the average time on page is only an estimate? Even when I filter it like you suggest, the time on page is still about 2 and a half minutes. I thought Google measured bounces based on whether someone goes to another page on the site or not. This blog says that Google ignores time on page and level of engagement link . Has that changed since? –  James Pegg Dec 1 '11 at 19:57
    
@Takuhi It is not an estimate, it just misses the amount of time spent on the last page. If a user looks at 2 pages, time on site will register the time between the 1st hit and the 2nd hit. 3 pages - 1st hit and 3rd hit. But GA cannot measure how long user stays on the last hit if there are no further hits. So that's why only one hit shows as 0. You are correct Google counts bounces as % of visits that do not go to another page on the site. –  joshuahedlund Dec 1 '11 at 20:11
    
Ah, that makes sense! I guess I'll have to find out why people are leaving, and also why people are staying. Thanks for the help! –  James Pegg Dec 1 '11 at 23:23
    
Good points @joshuahedlund, I took a look at my site, turns out even I have a good 50% of such traffic! –  Prasad Feb 25 '13 at 10:37

Remember bounce rate is the percentage of visitors who only visit one page on a site and leave. Make sure your stats are correct and that you aren't count moves to other pages on your site or the display of an off-site shopping cart or the like.

Also make sure that your stats are excluding traffic from 'bots and other automated sources. Some of my sites get a crazy bounce rate from 'bots and SEO/PPC spying software. I try to exclude or ignore these stats when I notice them.

The page view time indicates that visitors are hanging around and checking out the product in question. However, the real question is if the page is converting, is it leading to sales? If so, don't worry about the statistical anomaly too much. However, if it isn't converting well, then you may want to examine why this is (price, unclear purchase terms, page design, etc.)

share|improve this answer
1  
Gogole Analytics should already be excluding traffic from bots and automated sources. –  joshuahedlund Dec 1 '11 at 17:32
    
@joshuahedlund - I've found it less than optimal in this regard, especially when dealing with sources that hide or spoof their referrer information. –  jfrankcarr Dec 1 '11 at 17:47
    
Gotcha. That's new to me. What tips you off to that kind of behavior tainting Analytics stats and how do you filter for it? –  joshuahedlund Dec 1 '11 at 18:38
1  
Since most of this behavior was coming from sites in certain countries in Eastern Europe and Southeast Asia and since I'm not targeting those markets at all I filtered all of this traffic out. I also filter out certain US based proxies that some of them use to evade such filtering. Of course, that's a moving target so I have to watch for weird activity spikes as well that indicate automated link spamming attempts or other such things. –  jfrankcarr Dec 1 '11 at 19:16
    
Surely the site as a whole would get a high bounce rate if it was a bot? The home page only has a bounce rate of 23%, whilst getting 18% of all visits. With conversions, should I ignore the conversion rate and just look at the total number of conversions? (My thinking being that a high bounce rate will skew the conversion rate anyway). –  James Pegg Dec 1 '11 at 19:45

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.