Analytics Speak

Being accurate about counting PDF downloads involve two challenges

1. Clicks to the "download" button initiates the PDF download, but you have no way of telling if the download was completed successfully using GA. Log files are typically used to count partial vs. completed downloads.

2. GA allows only one conversion per session. ie If you went to a website and downloaded the same pdf 5 times. GA would register your session as having completed 1 conversion.

There are 2 ways to track actual number of PDF's downloaded in GA. Remember these are counts of the start of the PDF download, does not guarantee the PDF was fully downloaded

1. Go to the "Content" tab -> Top Content
Use the filter at the bottom to filter pages containing ".pdf"

                                                              




This will list all the people who saw the pages that ended in .pdf. Specify file name i.e filename.pdf in the filter if you have multiple pdf's.

2. If you have set up conversion goals for your PDF files then, Go to Goals -> Goals Verification Report to see total PDF downloads and which files they were.

                                                                                                                                                                                            

Log-files solution:

Log files record HTTP 206 for partial downloads and HTTP 200 for completed downloads. If you query out the number of HTTP 200's that should give you a record of all completed downloads.

I'd love to hear from you on comments/suggestions or better ways to count PDF downloads. Do leave your comments below.

Thanks!

This post goes into detail on a question I was asked in an interview. The question wasn't as straightforward as my writing, for obvious reasons.

" If you had aggregate non PII information - how would you go about making deductions on the value of each new visitor "

Assumptions
The visitor is a new visitor
You have no information beyond cookie information on the visitor
They don't login or sign up

Information on visitor
IP address (geographical area)
Referral source

My answer: The only way to tag a value to a new visitor with little or no information is by analyzing their clickstream on your website.

There should be a score assigned to each link on the website, or to a combination of links.

For example:
User 1
LP -> Products (click1) -> Most Expensive product name (click2) -> Features (click3) -> Exit/Abandon page

User 2
LP-> Products (click1) -> Productx (click2) -> Help center(click3) -> Q&A troubleshoot (click4) Exit/Abandon

User 1 would/should accrue a better value score than User 2 simply because he seems more likely to be a potential customer than User 2.

Now this scoring method is a common sense method - but put into a numerical format makes it easier to compute scores for all the traffic to your website.

The ultimate goal would be: Every user accruing a score > X, starts to see stronger call to actions, offers etc. The information is recorded in the persistent cookie, and the system every subsequent visit with a higher priority/more targeted ads etc.

I am an analytics junkie.

I have a keen interest in analytics, have completed the Google Analytics Individual Qualification, and have in the past managed and optimized SEM accounts.

I have worked on testing and optimization of offers/creatives and landing pages with Google Website Optimizer. I would like to delve more into the world of testing and optimization.

Any feedback on the blog, or my posts is greatly appreciated.

If you want to contact me regarding my work, analytics questions or work related discussions, I can be contacted at sri03@gwu.edu.

This blog will outline my journey to being an analysis ninja fighting through the hippo's to make it in the WA world.

Almost sounds like navigating a video game. :)

Cheers!