Google Analytics Magic - Part One

This is an ongoing series titled, "Google Analytics Magic."  This is part one where I will explore my favorite custom report - a customized keywords bounce report.  In part two, I will go on to answer some of the questions sent in from this post and then explore how to use the Advanced Segmentation feature in Google Analytics. 

I've been having a lot of fun in Sparked. I recently offered to share some of my favorite Google Analytics tips and reports with an organization there. I then figured this might be useful to a wider audience, so this crazy blog post was born.

BloomingBetty has agreed to let me use their data as an example. BloomingBetty is a new blog focused on defining success for women. I highly recommend checking it out. Full disclosure: Lori is a good friend and I have been advising her pro bono.

This will be a multi-part series on using Google Analytics.

What I love about Google Analytics, beyond being free in most cases, is that advanced insights can be found without having to be a programmer or even an analytics expert.

My Favorite Report

Anyone using Google Analytics is likely familiar with the dashboard. It's the default view. 

The key to Google Analytics is to always go at least a little bit deeper than the default view. But where to go can be daunting.

In this post, I'm going to share one of my favorite reports.

Every expert tells us that the reports we choose need to be based on the objectives and key measures that matter to us. I totally agree and I shouldn't have a favorite report, but I do. 

It is one of my first "go to" reports when I start working with any new client.

It is a customized Keywords report. I'll show how to create a customized keywords report at the end of this article. First I will start by exploring the standard "Keywords" report already available to everyone in Google Analytics. 

I get to the standard Keywords report by clicking on "Traffic Sources" in the left hand nav and then selecting "Keywords".

  
 

A report much like the following one should appear.

(click to enlarge any screenshot)

The Keywords report shows the top search terms being used to get people to my site and what happens, in aggregate, once they get there.

I can see, for example, that people typing "blooming betty pike" into a search engine to get to this blog spent significantly more time on the site than people who typed in "anna kinworthy". I can also get a feel for how many pages of the site our visitors are averaging and the percent of new visits versus returning traffic.

But what I often look at is the bounce rate on this report. It is in the final column.

Google says the bounce rate is "the percentage of single-page visits or visits in which the person left your site from the entrance (landing) page. Use this metric to measure visit quality - a high bounce rate generally indicates that site entrance pages aren't relevant to your visitors."

This is actually misleading. The first part is correct. It is the percentage of single-page visits. A bounce rate or 93% means that 93% of people looked at only the one page before leaving the site.

A lot of people, Google help included, think this is a bad thing. But what if they found what they were looking for?

Just because someone came to the site, found what they were looking for, and then left the site is not necessarily a bad thing. In fact, you could argue that in many cases this was a successful engagement!

Let's say that my friend Janet wants to know the hours of a local small business. She searches for the business in Google and then clicks through to the page with the information she wanted. She jots down the information and closes the browser. She will show up as a bounce in the report but she is happy. The business should also be happy.

Before getting back to the Keywords report, I want to look at my least favorite report for looking at bounce rates. It is titled the "Top Landing Pages" report. I get to this report by clicking on "Content" in the left hand nav and then clicking on "Top Landing Pages."

In the screenshot above it would look like there are a number of pages with very high bounce rates. And most firms would focus on fixing these pages.

But what if all these people found what they were looking for and then left happy?

What's missing is the context around the visit. If I only knew why they came to the site in the first place, I could then make a strong educated guess as to whether or not the page they visited had the information they were looking for.

This is where my favorite report comes in. Let's head on back over to the Keywords report.

On the left I can see all the search terms people typed into a search engine that brought them to the website. This is the context I was searching for! On the right, I see the average number of pages visited, the time on site and the Bounce Rate.

I now have significant insight into know why they came here (search term they used to get to the site on the left).  But did they get to the page with the information they were looking for? If they did, then a high bounce rate may not be an overly bad thing. 

This is where the standard Keywords report fails us.

Sadly, clicking on the search term on the left will not drill down and show the pages they were brought to.

This is where my custom Keywords report differs.  It lets me drill down. Other than that, it looks almost identical.

The only addition is the first item "(not set)" which means there was no keyword - it is reporing on site activity that did not originate from a search engine. So we can ignore this line for the most part.

The drilldown is the missing link. If I click on any of the search terms on the left, I can now drill down to find out what pages these people visited as a result of their search engine query.

In the graph above, I highlighted the third item ("sheryl sandberg...") because I noticed it had a bounce rate of over 90%.

If I click on this search term, I can see the pages on the site these people were sent to.

Look at that. They were searching for "sheryl sandberg" and were sent to our article on Sheryl Sandberg. They all got the information they needed and were looking for. And so 93% of them then went on about their day. 

So I would check this search term off my list and continue looking. I would look for pages where I have higher bounce rates and I am sending people to a page that does NOT answer the search terms used.

This report will identify some of the big, easy fixes that can have significant positive impacts on traffic and visitors. Someone not finding what they were looking for and leaving the site is a big time missed opportunity if I truly did offer what they wanted. I had them but they were on the wrong page.

I'm sure there are some questions so please fire away in the comments. Ot's really hard to present this in blog format.

Here is a video that shows how to build your own custom Keywords report.

(You will want to watch this full screen in hi def)

For those of you proficient with Google Analytics, here is a link to my my drilldown keywords report: http://bit.ly/fM9Dh0

In my next post I will look at real-time segmentation.  But let me know if there's another area people are interested in!

Comments

[Amrita]

Thanks Sean, looking forward to the next post.

[Sean Howard]

Me too! oh... wait... I have to write it. Damn. ;)

[Eportelance]

Loved this post, Sean. A question on bounce rates, though... I was under the impression that time spent on the page also factored into the bounce rate calculation. So, for instance, if someone visited only one page on your site but spent 3 minutes there, that may *not* be counted as a bounce. Is that the case?

[Sean Howard]

Great question. 

My understanding is that Google Analytics uses page visits only and not time.

Some people hack the standard google code to create a second hit after a certain amount of time has passed but this will cause the bounce rate for google to no longer work as it uses page visits only.  If you visit one page and stay on it for 3 minutes, you should still show up as a bounce in GA if you then  leave the site without visiting another page.

Note that time on site is erroneous on bounces. I did not go into this in the article as this would confuse the topic. So if you only have one page with a timestamp, there is no second page to compare timestamp to. So this is why some people have hacked it as above to give them a better sense of time on site.

I am likely really confusing people now.  ;)

[Syrten]

^ Another bounce statistic leaving satisfied.

[Sean Howard]

LOL!  This is the best comment I have ever received on a post.  Thank you!

[Adam W. Warner]

This was incredibly helpful! Thank you so much for taking the time to show how to do this.

One request I would like to suggest is a "what now?" step or suggestions after identifying that a high bounce rate was actually caused by people not finding the right information.

Looking at the report I just created, there were a few high bounce rates where the page being visited didn't have the keyword info that people were seeking.

I suppose my best options would be to either write a specific article based on those keywords, or to include some reference links to other sites that have info specific to a user's search.

[Sean Howard]

Awesome suggestion and great questions, Adam!  I will make my next post on this topic a What Now?

Would you be willing to share your findings as a case study? Let me know. Id just need some screenshots of what you found as the pages in question and I could then include contextual recommendations specific to your situation.

Sean

[Fred Hahn]

Hi -

I'm a GA idiot and after reading Chris Brogan's blog decided to learn from you. But I got stumped at this. You said:

"If I click on this search term, I can see the pages on the site these people were sent to."

Well, I went to the GA page for my GA account and I cannot see the pages they went to.

Could you clarify HOW one sees the pages that people went to on your site?

Thanks!

[Sean Howard]

Hi Fred,

  The standard Keywords report will not give you this vital piece of information.

  Theres a video at the bottom of my post where I walk through how to create your own custom Keywords report. This custom Keywords report will let you click on the search terms and drill down to the pages they visited from those terms.

  Heres a link to the video directly.

  http://www.youtube.com/watch?v=QKGLVx6rATU

Sean

[Fred Hahn]

thanks Sean. It seemed from the blog that it was there. You might want to tell people that upfront cuz I spent 30 minutes feeling like a doofus cuz I couldn't figure it out! :( Thanks for the great info!

[Fred Hahn]

Did it Sean. Thanks. I'd love for you to blog on how a small business should be using GA and how often.

[Sean Howard]

Thanks Fred,

  I think there will be quite a number of stories on this series.  I had no idea the interest level would be this high! Ill gladly put your topic request into the mix!

Sean

[Volunteer]

Sean,
So, this is Heather W. from Sparked! I finally sat down to dig into this and realized we don't have Google Analytics. We have its ugly stepsister direct from the our webhost. Do you know if it's possible to add GA to a site in this circumstance? The one we have does a lot of what Google has (it seems) but won't let me do any custom reports. I'll call the company and see if they can make it happen, but they tend to be pretty useless (guess who might be looking for another web hosting service!).

Thanks for all your help! I'm already looking at our bounce rate info in new ways!

[Sean Howard]

Hi Heather,

  Here is a link to how to add Google Analytics to what I believe you site is running on.  Im making this assumption solely based on a review of the HTML of your site.

  http://support.homespunwebsites.com/faq/content/20/70/en/i-want-to-add-google-analytics-to-my-website-how-do-i-do-that.htm

Sean

[Volunteer]

You rock so much! Thank you!

[Sean Howard]

Your welcome and thanks for the kind words!

Check back in a few days and I should have part two up on this series.

[C.B. Whittemore]

Sean, this is fantastic! Thank you. I'm looking forward to part 2. Best, CB

[Sean Howard]

Thanks for the comment, C.B.!!!  Long time no chat!

I have most of part two ready to go.  Fingers crossed for today/tomorrow!

[Html Kodları]

Good comment. Thanks.