The Danger of Aggregate Data

Share This Post

The question I hear most often from clients and employers is “What does this data mean?” This is most frequently asked in reference to Google analytics aggregate data.  The short answer to this is aggregate data does not mean all that much.  Aggregate data summarizes or combines other data for analytical and research purposes.  It is the analytical equivalent of cliffs notes.   Data is very powerful and can help make informed decisions about content, direction, conversion and a host of other topics.  But looked at improperly it is as accurate as if you wrote down answers on scraps of paper and randomly pulled them out of a hat.  

The data you collect tells the story of your site.  The problem is that for the most part the aggregate data only tells part of the story.  Take for example the information you can see in the Google Analytics Dashboard.  When you first log into Google Analytics you see a snapshot of your site: Total visitors, unique visitors, traffic sources, etc…  This information is a good place to start but basing decisions solely off of this view will do you a disservice.  

Say for example you go to your dashboard and see that your bounce rate is 70%.  This means that 70% of all visitors came to your site saw the home page and left immediately, deciding that your site was not even worth a second click.  You would naturally assume based on this aggregate data that there was something horribly wrong with your landing page.  You decide to spend money on a costly redesign to make your home page more inviting.   After investing thousands of dollars you re-launch your landing page and sit back confident that you will now see a dramatic decrease in you bounce rate and consequently a just as dramatic an increase in conversions.  But nothing changes so you conclude that you must not have gotten the design right. Still believing that your high bounce rate is a function of the landing page design you decide that you will rotate a few different designs and see which one works best and has the lowest bounce rate and highest conversion rate.   This time you see a positive change but it is not statistically relevant.  You are at a loss, you were sure that one of these designs would surely make the impact you were looking for but all you got for your trouble is a bunch of invoices from a designer. 

Now let’s take the same situation and instead of looking at the aggregate data we segment that data.  You find that you have 3 top traffic sources: a mystery referral URL, organic search, and paid search.  The mystery URL is sending a bulk of your traffic.  You have noticed this URL before and though you were unfamiliar with it you figured any traffic is good traffic and you are happy to have it.  You start to get curious about this traffic and you drill down to the bounce rates for your top sources.   You see that organic search has a bounce rate below 40% and so does paid search but the mystery referral URL has a bounce rate close to 90%.  Not only does it have a high bounce rate the average time on site is less than 20 seconds.  What in the world is going on? How can this be?  WTF? Are all questions that you should immediately ask yourself if you see this situation. Digging deeper you find that this mystery URL is associated with a pyramid scheme paying people to surf to several web sites a day. This is all unqualified garbage traffic and all it is doing is skewing your aggregate data.  Since you took the time to segment you data you see that, while you have some more work to do driving good targeted traffic to your site the good traffic you are seeing is sticking around and checking out what you have to offer.  Your time will be best spent cleaning up your analytics as best you can and getting the story straight before you spend money or resources on redesigns or other costly fixes for problems that may not exist.

More To Explore

Baciscs of Net Neutrality
internet freedom

Net Neutrality Basics and Overview

What is Net Neutrality? The principles if Net Neutrality: Keep the Internet free for all to use Anyone can access to the Internet and all