Google Analytics has one major drawback for ecommerce sites – it only gives credit to the keyword or referring source a visitor clicked on the last time they visited the site.
This “last click” approach is what makes a lot of large ecommerce sites choose multi touch attribution analysis products such as Omniture and Coremetrics over the more user friendly and easy to use Analytics interface.
The main problem with “last click analytics” is that a large number of sales are credited to direct traffic & brand keywords (organic or PPC) simply because they were the last method used to find the site before a purchase was made.
In truth users often carry out a number of long tail product searches and generic keyword type queries as they research a purchase and then visit the brand site by searching for it directly after they have made the decision to buy. By not tracking the research phase of the transaction a lot of credit is taken away from the SEO campaign when in fact it’s often SEO that drives branded search.
The Solution – User Defined Values
Luckily there is a way to track the users first click with Google Analytics which goes a long way towards addressing the problem. By default Google Analytics only allows us to set one user defined value so if you want to track your 2nd, 3rd & 4th clicks then you need a small hack to stuff multiple values.
The first step is to understand the _setVar function which allows you to add a User Defined Value to each visitor.
You can view conversion behavior for each of your custom segments from the Visitors Overview report. Click the User Defined link, then select the Goal Conversion tab.
Once you understand how to set the custom variable (it’s pretty easy) then you need to figure out a script to determine when to set it and who for. I’m not going to give code for this because all platforms are different but the basic method is as follows:
- When a user visits check if they have a cookie with the first click referrer details
- If there is no cookie set then generate a cookie with the referrer keyword
- If they didn’t arrive from a search engine store the referring site or identify them as a direct visitor
- If they already have a cookie then they are a repeat visitor and you can run the _setVar function
At this stage you might like to split your keywords into sectors using some kind of string matching algorithm. For example rather than just adding raw keywords you could group all certain keywords together as either “generic”, “brand” or “long tail”.
Viewing the Results
The data is presented within the standard search engines reporting section of Analytics by clicking on:
Traffic Sources > Search Engines > Google > Brand Keyword > Non-paid > Ecommerce tab > User defined value
What you are able to do is drill down into the
We’ve been able to credit a lot of extra sales to SEO which would normally have been allocated to direct traffic or brand keywords. Would be interested to hear your results too if you decide to give it a try.