Marketing Logic

Many people point to the bounce rate in Google Analytics and assume that those who bounced are worthless visitors. This is not the case.

Before I explain why, I want to say that I have changed the GA code so that Bounce Rate is calculated from those who spent less than 10 seconds on the site, not those who viewed one page, which is the default. The reason I did this is that many of my sites are blogs or are using blogging extensively. If a visitor enters on a category page, they can spend up to 10 minutes reading posts, yet the default GA code only sees one page viewed, and says that they bounced.

If you would like to use this bounce rate, you need setTimeout, and here the code and where to place it:

var pageTracker = _gat._getTracker(“UA-xxxxx-x”);
pageTracker._trackPageview();
setTimeout(‘pageTracker._trackEvent(\’NoBounce\’, \’NoBounce\’, \’Over 10 seconds\’)',10000);

Now, why are visitors who bounce not worthless? Simple, GA seems to be calculating bounces off of the most recent visit. This does not mean that the visitors have not spent time on your site before. Here is a screen shot of the Visitor Loyalty report, segmented by those who bounced.

(click to enlarge)

You can see that although the majority (93.5%) of visits were the 1st time to the site, around 6.5% of visitors were returning, some of them up to 9 – 14 times.

This means that these returning visitors liked what was offered on the site enough to come back, but then why did they bounce?

I believe that many of the Loyalty Bounces are looking to copy URL’s to send to other people. Yes, that may sound like wishful thinking, but when I go back to a site and leave immediately, it is usually for that very reason. Another reason may be to verify some piece of information. Every case is different, which is why Custom Segments are so important for your analytics.

To see this in action, you will need to create a custom segment by clicking on Advanced Segments in the left hand navigation of GA. Originally I used Metrics-Site Usage-Bounces with the Condition Greater than or equal to, and the Value as 1. This gave me the bounce rate of the site. Then I created another segment to combine bounce rate and returning visitors;

(click to enlarge)

Dimensions – Visitors – Visitor Type – Matches Exactly – Returning Visitors

Metrics – Site Usage – Bounces – Greater than or equal to – 1

There you have it. You created a custom segment, and you can see what I call a Loyalty Bounce.

Update: When I applied this to another of my sites, I believed it confirmed my suspicions about why people are returning to the site, then bouncing. The Top Landing Pages report segmented by Loyalty Bounces was very similar to the Top Content report for all visitors. The pages that are most popular on my site are the ones giving me the Loyalty Bounce.

So a visitor that bounces is not always a bad thing. They may be spreading links.

I know that there are a lot of people who are wondering if they should make the switch to Google Analytics, and I would say, yes. I was one of those people for a long time. I used Clicktracks (now Lyris) since the day I started analytics, and always resisted GA. It has only been within the last few months that I have really started to use GA extensively, and not only that, to like it.

Google does bother me a little bit with the amount of information that it has on the millions of people using it. Giving Google access to my analytics only increases its knowledge, and we all know that knowledge equals power. This ties in with Matt’s recent post on the failure of search engines, the bigger and more powerful the search engine (Google), the more we have to design and analyze for that engine. This can take away somewhat from the user experience and the conversion of those users.

In spite of all that, Google has given us a very powerful tool to analyze our visitors; and they gave it to us free.

Let’s take a look at some of the competition out there first.

Webtrends can cost upwards of $20,000 just for the first year. That buys you a massive amount of installation work, more graphs than you can graph, but not much actionable data. I have said that Webtrends is like a bodybuilder on steroids, it’s big and muscular, but it ain’t got no balls.

I have used Webtrends, but I found that it was too big and bulky, and I wasn’t able to really do my job and find actionable insights. I know there are people out there who will say that it is my fault, but it’s not. In my experience, Webtrends (and most other analytics programs) are implemented and run by IT or marketing people. Neither of which are web analysts.

I loved using Clicktracks. Because it uses log files, one of the major advantages is the ability to apply anything to historical data. That means I could set up a goal, and see the data for that goal for any time period I chose, even if that time period was before I set up the goal. Whereas in GA, you can only see the data from the time you implemented the goal. The only problem with log file based analytics is the slowness. A site with more than around 25,000 visitors really gets bogged down, and if you go over a couple hundred thousand, you might as well take the day off while it calculates the data.

Clicktracks is fantastic at segmentation. It is one of the basic features of CT and it makes it very easy to create custom segments, and to combine them any way you want. (Without segmentation, analytics is useless.)

GA is much easier to install, and is very easy to use. It does not look near as impressive as Webtrends, or have the ability to compute historical data like Clicktracks. However, there are a number of reasons I believe it is better.

First, it’s free. You can’t get a better price than that. Many people think that because it is free, it isn’t worth it, but they are wrong. Google wants your site to perform well. GA is their way of helping you make your site better, so that they have better quality sites in their index. Kind of selfish on their part, but it works.

Second, Avinash Kaushik is their Analytics Evangelist. Avinash helps Google to “…shape the vision, direction and features of 13 different Google tools that provide data to customers.” This is one of the biggest reasons for GA’s success, it has top web analysts working to make it usable and actionable.

Thirdly, because they give out the API, anybody can create code for Google Analytics, and there are some fantastic scripts and add-ons out there. Here are a few that I use;

Enhanced Google Analytics - This was a script for Greasemonkey, but was later made into a FireFox add-on. This puts a little button on the top of the Referral report and Keyword report that show you the “unusual” traffic. It graphs out sites/keywords that brought 50% more traffic over the last 7 days, and those that had 50% less traffic over the last 7 days.

The next is Better Google Analytics. This is a collection of scripts that give you many, many options. Here are a few of the cool toys it adds to GA;

Adds search function to each search term in the Keyword Report.

Converts percent values to absolute values in tables.

Allows you to go full screen and minimize the left hand navigation.

Links to Google Insights and allows you to check keyword terms and perform a Google Insight search.

Social Media metrics from Digg, Reddit, StumbleUpon and more.

These are only a few of the scripts that are available in the Better Google Analytics Add-on, and there are even more out there. With so many new scripts being written for GA, it will always have more to offer, and most of those offerings will be what we analysts need, not what some giant corporation thinks we need.

I am by no means a Googlephile. As I said earlier, sometimes I have to put on my tin foil hat before I start searching the web. However, Google has given us the tools to make our sites better, the code to make our own scripts for Google Analytics, and they have done it for free.

How can anyone compete with that?

If you think your analytics program is the best, tell me and let me know why.

Update:
Avinash has reminded me that by using Custom Segments in GA, you can create a goal segment and see basic historical data. I have used this, but my point was that even by doing this, there is still no historical data in the Goal Reports.

Also I have used Omniture Site Catalyst (another massive analytics program), and it does not even have segmentation abilities, and as I said above, without segmentation, analytics is useless. So I guess you can take from that what I think about Site Catalyst.