In my last article, I reviewed some of the key findings from a content analysis of Web Analytics World content views for all of 2014. If you have a website where you are generating content on a regular basis and you want to perform a similar analysis, this article will show you how to use Google Analytics to determine the worth of your content marketing.
First, set your date parameters for your analysis window. To do this, click on the date selection down arrow in the top left of the screen.
Next, use the left bar navigation to go to all pages in the Behavior, Site Content section. Select “All Pages.”
Depending on the way you’ve structured your website and your naming convention for blog posts, you can either use the default “page” view or select the “Page Title” view. WAW uses a standard date-based naming convention at the header of all of their content; using this view makes it easier for me to understand when the content was created for the sake of analysis. However, if you have a title naming convention where you explicitly call out the content format (i.e. “Infographic: Why April Wilson Is Awesome” or “POV: Three Reasons Why You Are Losing Half the Battle by Not Setting Goals in GA”), it will probably be more meaningful for you to use the “Page Title” view. This is located at the top of the detail table under the graph.
Content Analysis Methodology
You need to conduct your analysis with a vision focused on what you want people to do when they go to your website. Usually this is set up as a “Goal” in Google Analytics. When your goals are set-up, all of the reports and data you get out of Google Analytics show the page value (if you’ve assigned a monetary value to the goal).
Since WAW doesn’t have Google Analytics goals set up, I used the strategy that they want their content to a) be read and b) go viral. The logic behind this is that the virality of social sharing will lead people back to their website.
On the Pages report, the top 10 are displayed by default. I expanded the list to show the top 100 for analysis purposes, with the intent to only analyze the top 50 articles. To expand the list, scroll to the bottom and click on the “show rows” arrow to get back as many records as you want.
Then I exported out the list of 100 into Excel so I could add rows / data that isn’t easily available on the Google Analytics report, such as author, topic, format, and social shares.
To add the missing data elements, it’s time for everyone’s favorite job as a big data analyst – manual work. I went to every post in the top 50 articles and read them to classify them into formats and topics. I added the author. I also looked at the social sharing widget at the top of every page to manually add the number of shares from certain social networks (WAW has 8 social networks you can share to in their widget, I just picked three: Twitter, StumbleUpon, and Facebook).
This leads to a “raw” data set that looks like this:
All the good stuff is now in one table and it’s time to analyze. I used pivot tables to find trends and answer questions around top viewed content, etc.
As you can see, I’m a big fan of using conditional formatting in Excel to quickly spot key trends. These were my indicators for what insights I wanted to visualize (with a donut graph or bar graph, etc.). In the chart above, for example, I can quickly see that Lists get the most page views, but case studies keep people on the page longer. Infographics have the highest bounce rate, and case studies have the highest. Infographics are also most shared on Twitter while Product Reviews are the least shared.
You get the point.
The only other test you may want to run is statistical significance. For example, the chart above may make it look like a “How To” article is more likely to get tweeted than a List (24% share rate vs. 8%, respectively). It’s not a bad idea to test for statistical significance and make sure the observed difference is valid. There are many statistical significance Excel sheets you can download for free… I use the one from Visual Website Optimizer.
Given that content appetites and trends change pretty frequently, you should probably audit your content performance at least quarterly to get a read for what’s going on. For pieces that perform really well, consider updating the content or presenting that same information in a new way (example: a great “list” article could be repurposed as an infographic). For content that isn’t performing well, take a strategic step back and decide if there’s really a good reason for you to be creating it – or if you’re just vomiting out junk for SEO purposes. (Trust me, to the search engines, properly tagged junk is still junk.)