Bounce rates - a very important metric

[stuarty]

Not sure how many of you use analytics but you really should be.

Analytics from Google is a very good way of determining the "productivity" of your website. It's very complex but can be used in a simple formto give a snapshot of where your site is "at".

One of the most important areas is something called the "bounce rate". This is known as a "metric".

Whilst it's one of the most important - it's also one of the easiest to confuse.

What it basically means is someone came to your website, saw the page, did nothing and left. That's called a "bounce". Generally a high bounce rate is not good - say 60-70% bounce rate means people are not liking your site.

HOWEVER....there's a lot more to it than this. We have done a trial using our own website and measured this in a split test.

Our bounce rate sat at around 60% using test A and 35% using test B.

Our website runs as a blog we can turn the front page into 5 full length posts in one big long read. We can also turn it off into snippets for the same 5 posts.

Bounce rate goes hand in hand with another metric called "time on site". This is the time that a person is looking at your web page.

Basically when we had the 5 full articles on the website the TOS was quite long and very few click throughs. The opposite proved much more effective, reduced bounce rate, improved click throughs and improved overall readership of secondary pages.

The moral of this story is as follows...

If you have a front page that has quite a lenghty introduction on what you do, it may be a better idea to split this up into a snippet which can be clicked through on. This will effectively give you two bites at the cherry to hook your visitors using a call to action. (ie a form, a click here to buy button).

For example - make your front page intro is only one paragraph leading to a click through to the full article. Add a small contact form on the front page. When the user clicks through to the full article make sure you still have the same contact form on the new page - this is effectively imprinted in the back of their minds twice and acts as a subliminal instruction to click and buy or contact.

It's worth a try and you may find your website enquiries will improve.

 

[PeterHoggan]

Interesting article stuarty and I agree with what you say. However there is a problem in the way that Google analytics, or any page tagging analytics system for that matter, measure bounce which can slant results when A/B split testing.

The problem is a browser will not carry out a server request if a page is already cached therefore the JavaScript is not called unless the page is refreshed. In essence this can create many false bounces. For example, a visitor landing on page A moving to page B then back to page A and exiting will be recorded as a bounce if page B was held in the browser cache. In my experience the best way to analyse bounce rate is via log files.

 

[Boxby]

But if a website provides the customer with all the information that they need on the first page that they visit, they will have a very high bounce rate, but still be a very effectively performing website.

 

[stuarty]

But if a website provides the customer with all the information that they need on the first page that they visit, they will have a very high bounce rate, but still be a very effectively performing website.

So how do you differentiate between a bounce and a satisfied visitor. Was it a case of...I can't be bothered reading all that crap and they bounce or was it I'll read to the end and close the browser or open a new tab? In both cases its a bounce because.....Analytics programs can't calculate the length of time of the last page in a session. In single page websites or single page sessions, this becomes the last page and Google's analytics can’t compute how long was spent on that page. So it becomes a bounce.

This is precisely why it's best to split your page into snippets with links (or hooks) leading on to the next page

 

[stuarty]

which can slant results when A/B split testing.

The problem is a browser will not carry out a server request if a page is already cached therefore the JavaScript is not called unless the page is refreshed. In essence this can create many false bounces. For example, a visitor landing on page A moving to page B then back to page A and exiting will be recorded as a bounce if page B was held in the browser cache. In my experience the best way to analyse bounce rate is via log files.

Not sure what you're saying here Peter - it's the javascript that records the bounce and reproduces it in your analytics data. If the script doesn't run then it can't record a bounce so it can never be a false bounce? Infact, if the script doesn't run no data will be recorded at all.

Regarding the split test in our site - all our pages are dynamic so there won't be any cached pages

I'd agree totally about analysing log files but the average site owner doesn't get access to these and you need a server log parsing application to view these. However....this doesn't solve the problem of bounce rates. Think tabbed browsers, killed sessions, page timeouts. People often browse the same site accross multiple tabs. If someone opens a new tab then the originating page becomes a bounce when the browser or tab is closed. If it times out after the default 29 minutes then time = 0 secs which also equals a bounce. From raw log files you only know how long a page session lasted - you can't determine if it was a bounce or not because you don't know if the user clicked and went for lunch or opened a new tab of the same page.

For the average site owner - google analytics is as accurate a representation as most can understand. We have both raw logs and GA and the accuracies between both are actually very close.

 

[Boxby]

So Analytics programs can't calculate the length of time of the last page in a session. In single page websites or single page sessions, this becomes the last page and Google's analytics can’t compute how long was spent on that page. So it becomes a bounce.

This is precisely why it's best to split your page into snippets with links (or hooks) leading on to the next page

But, as I understand it, bounce rates don't adversely affect rankings.
And if your customers are finding all the information that they need on your home page, why make them work harder?

Bounce rate is a tool, not a rule. High bounce rates can't be simply taken to mean poor websites.

If your website works, why would you change it simply to improve a number on your website stats? Because stats just simply stats. They only mean what you want them to mean.

Websites are for visitors - not for analytics.

 

[stuarty]

But, as I understand it, bounce rates don't adversely affect rankings.

True but I never mentioned ranking and I'm really not sure why you would mention this?

And if your customers are finding all the information that they need on your home page, why make them work harder?

Ok...how about if all newspapers and magazines were only one page long? If all websites were one page then how would servers cope? What about bandwidth? How would the visitor feel if they had to read 5000 words or wait for 100 images to load before they got what they wanted? It's a lot more work for the visitors this way than it is to provide structured links to secondary pages. Think about a blog archive or a product catalog.

Bounce rate is a tool, not a rule. High bounce rates can't be simply taken to mean poor websites.

Bounce rates aren't tools or rules.They're "metrics" or measurements. I never said bounce rates meant poor websites. What I do say is that its an "indication" of how productive or non productive your website is. High bounce rates as a rule are "generally" poor performers in terms of sales/conversions.

If your website works, why would you change it simply to improve a number on your website stats?

What do you mean by your website works exactly? Do you mean as long as its available its working? Does one sale mean it works?

One very sound reason why people want to improve the number in their stats is because it equates directly to £s which keeps people in business.

Because stats just simply stats. They only mean what you want them to mean.

I'd love my bounce rates to mean a mansion on a beach in california with a ferrari parked outside and a pub next door

I'm afraid the data isn't just statistics - it's absolutely brilliant information that can be used to identify problems on your website which can be used to make your site more productive in terms of sales.

Websites are for visitors - not for analytics.

Yes but in order to obtain analytics data you need a website in the first place. Websites are also for owners.

 

[Scottish Business Owner]

I've just checked these figures for the forums. Bounce rate is 1.37% and average time on site a fraction under 7 minutes. This strikes me as quite good but are forums traditonally like this anyway?

I'd like to understand a bit more about what this means and whether I should be doing any work to get the bounce rate down even more or get the average time up.

 

[stuarty]

I've just checked these figures for the forums. Bounce rate is 1.37% and average time on site a fraction under 7 minutes. This strikes me as quite good but are forums traditonally like this anyway?

I'd like to understand a bit more about what this means and whether I should be doing any work to get the bounce rate down even more or get the average time up.

Yeah - it's pretty normal for forums. Low bounce rates occur from login redirects, click throughs to topics but even the lurkers click through. Behaviourly, forums are enticing and people are keen to see replies to their threads. Equally, email reminders increase click throughs and keep bounce rates very low.

This proves the point that very effectively that if your home page has nothing but links then these effectively become "calls to action" forcing a click through.

Rather than focus on reducing your click rate - (which probably can't be any lower) I would look at other metrics by drilling down to see what pages are your best performers. This way you can target advertising or any other monetizing features that you have.

This is the great thing about analytics - you can focus on finite details or the site as a whole.

If you want I can take a look at your analytics and do some eval for you. Pm me the details if you can .

 

[peteark]

Bounce rates are a good tool, there are many out there

People who say this is good, that is bad, like Boxby, who could never back up their theories with a reasonable explanation. These are the ones to ignore, they clutter up many a good forum, offering nothing but unsubstantiated tripe.


On the other hand Stuart knows a thing or two and backs this up with good explanations, one to follow

 

[peteark]

If a person or company wanted to work on their sites sub pages, bounce rates are a good metric, it highlight pages that are popular with the public, from ones that are not.

Ego often hampers people's ability to perform well online, they disregard good quality data, in favour of what they themselves believe, web designers make the same mistake in the business world, by measuring their performance on visual representation not improved conversion ratios.

 

[Gordon N]

Ego often hampers people's ability to perform well online, they disregard good quality data, in favour of what they themselves believe, web designers make the same mistake in the business world, by measuring their performance on visual representation not improved conversion ratios.

[Awoooooga! Sweeping statement alert!!]

Did you miss out the word 'some' before 'web designers'? According to your signature that also counts you in too!

I measure my performance on customer satisfaction, and meeting (and where possible exceeding) the requirements of the 'web design' brief.

On a more serious note, how would you suggest I measure conversion ratios on an information only site (other than a contact/feedback form),such as a tradesmans site or place-of-interest site etc? In my opinion not every site is built purely to be monetised, some are purely to provide information to the visitor.

 

[peteark]

Gordon I am sure you have the nounce to understand I was referring to ecommerce based sites.....maybe not....lol

I network with plenty of designers, they do not take the defensive role, they know there are many problems in their sector, saying that I would hazard a guess there are more in the SEO world.

I wrote an article today that is sort of an expansion of my post, makes me wonder about number 19, guess that will wait for another day. Search Engine Friendly Sites

 

[Gordon N]

It wasn't defensive, it was humour - hence the wink.

 

[peteark]

Humour.....I like that, sorry I missed the wink, 8 hours skiing is the only excuse I can offer

 

[Anita]

I'm not a designer, nor SEO expert, but I've followed this thread with interest. I've actually seen the positive effects an improvement in bouce rate has on business, down to the pounds and pence. I've worked with a number of clients who were dependent on online conversions and were experiencing very high bounce rates. Using SEO expertise to hone in to a more targeted audience together with a better front page layout (as described by Stuarty) and dare I say copy that appealed directly to the self interest of the visitor (just as important in my view) the site bounce rate of clients decreased significantly and conversions rose which ultimately resulted in more sales. I appreciate that this observation is nothing new to the experts in this field in the forum, however, if you are a business owner who has only ever been involved with your own site, I just wanted to add my perspective that from experience your bounce rate is indeed a very important metric!

 

[Boxby]

True but I never mentioned ranking and I'm really not sure why you would mention this?

You make the assumption that websites simply convert online. Not all do, very many don't. There are plently of sites out there whose sole purpose is to inspire, the right type of customer to pick up the phone and call.

So, if your website is achieving this, then the only possible negative effect from having a high bounce rate would be if google looked at your bounce rate and interpreted it as a "generally poorly performing website".

And that's why I mention it.

Ok...how about if all newspapers and magazines were only one page long? If all websites were one page then how would servers cope? What about bandwidth? How would the visitor feel if they had to read 5000 words or wait for 100 images to load before they got what they wanted? It's a lot more work for the visitors this way than it is to provide structured links to secondary pages. Think about a blog archive or a product catalog.

Again, you work on the assumption that all conversions must happen online. I am not advocating that you put your entire website content on one page, all I am saying is that just because the customer visits ONLY the home page, this shouldn't be automatically and simply interpreted as a "generally poorly performing" website. You need to think about the website individually, and the purpose that it's suppose to provide.

Bounce rates aren't tools or rules.They're "metrics" or measurements. I never said bounce rates meant poor websites. What I do say is that its an "indication" of how productive or non productive your website is. High bounce rates as a rule are "generally" poor performers in terms of sales/conversions.

Disagree, because its based on the wide assumption that the sale/conversion must happen online. It's simply to simplistic to state this.

What do you mean by your website works exactly? Do you mean as long as its available its working? Does one sale mean it works?

Does it bring in the level or work that you require it to? Simple as that. If you need a website to generate 3 leads a day, and it sits in the top 3 on google, even if it has a bounce rate of 70% you cannot describe that as a "generally poorly performing" website. It's a website that works.

I'm afraid the data isn't just statistics - it's absolutely brilliant information that can be used to identify problems on your website which can be used to make your site more productive in terms of sales.

It can, but if you look at statistics and simply follow assumptions and general "rules" and interpretations without thinking out what they mean to YOUR specific website, and YOUR specific customers, and what they use your website for, then you don't identify problems, you create more.

Yes but in order to obtain analytics data you need a website in the first place. Websites are also for owners

And that my friend is the WORST assumption that any website owner can make. Websites are for customers pure and simple.

 

[stuarty]

lol - yeah right!

 

[peteark]

You make the assumption that websites simply convert online. Not all do, very many don't. There are plenty of sites out there whose sole purpose is to inspire, the right type of customer to pick up the phone and call.

If a visitor reads a website and picks up the phone is this not classed as an online conversion??? There are tons of lead generation sites out there, they all convert online, if someone reads the site and contacts the owner.....Boxby last year you asked me how to sort out your own website, keywords and link building, now you are an expert in the field of on and offline conversion.......stop mate your hurting my ribs

 

[stuarty]

Now that I've stopped laughing....

You make the assumption.........

Didn't you read this below earlier?

People who say this is good, that is bad, like Boxby, who could never back up their theories with a reasonable explanation. These are the ones to ignore, they clutter up many a good forum, offering nothing but unsubstantiated tripe.

You know that expression "quit while you're ahead" don't you?