The Marketing Technologist.

We talk about analytics, code, data science and everything related to marketing technology. Backed by the tech geeks of Greenhouse Group.

How article size helps you understand your content performance

Last month, we shared our insights on the enhanced ecommerce for content report that we’ve implemented on GEEK. With the reports, we measure what articles are read, how often posts from a specific author are read, and how many words people read on our website. I’ve also mentioned that we would add article size to give insights on how different article sizes are performing. In this post, we’re focusing on that.

Adding the custom dimension

As described in the post, we’ve added three different types of article sizes:

  • Small: less than 500 words;
  • Medium: a word count of 500 or more and less than 1000;
  • Large: a word count of 1000 words or more.

The goal of this update was to see what article size works best for our readers. Because it gives us more info about an article, we've added it as a product dimension in Google Analytics. This way, we can use the enhanced ecommerce metrics and apply them to the article sizes.

The analyses

The first insights we can get from enhanced ecommerce is on read articles. I’ve set up a custom report with the following settings:

Google Analytics Article Size Custom Report

Set up a custom report and set the dimension to article size, and add metrics for revenue (word count) and quantity (number of reads).

You would expect to see the highest word count for large articles, and a reversed order for number of reads, as small articles will trigger a full read quicker than large ones. Let’s look at the data for October:

Google Analytics Article Size results
Article size results of October 2015

As expected, we see the highest word count for large articles, followed by medium and small articles. We also see an exact reversed order for quantity, starting with small articles at 816 full reads, and ending with large articles at 663 full reads. Now we have the data to show that our articles perform as we thought they would. It also shows that large articles are read quite well, as they almost double the word count of medium articles, with about 17% less reads.

A more interesting analyses

With enhanced ecommerce implemented, we’re also able to leverage the power of these reports metrics to give more interesting insights. For this report, I’ve added a second tab to the custom report and added the following dimensions and metrics:

Google Analytics Article Size Flow Custom Report

Set up a custom report and set the dimension to article size, and add metrics for product detail views (article opens), cart to detail rate (scroll interaction rate on an article) and buy to detail rate (full read rate).

With this report, you’ll get a better understanding on how article sizes perform. I would expect that the longer the article, the bigger the drop off during the stages of reading, e.g. a lower full read rate compared to small articles. Let’s look at our data for October:

Google Analytics Article Size Flow results
Article size reading flow results of October 2015

We see something interesting here:

  • Article opens: Small articles tend to have the most article opens (product views), the medium
    and large ones are roughly equal in numbers. This means that small articles might be good ways to get people onto our website and trigger a first read. Or they might be quick snacks for people looking for a short and interesting read.
  • Scroll interactions: The large articles have a slightly lower scroll interaction rate than the other two article size. This indicates that users find the first part of these articles less interesting, since these users start scrolling less often. Large articles have up to 8% less scrolls compared to medium articles.
  • Full reads: Looking at full read rate (buy to detail) we clearly see that medium articles have the best performance. Comparing the small articles’ rate to the medium articles’ rate, we see a performance that's 19% lower than the medium ones.

Putting it together

Combining the numbers of the two reports, we can tell that medium articles don’t generate the most read words or most article reads. But looking at quality, focusing on scroll interaction rates and full read rates, medium articles perform better than both small and medium articles. This could indicate that medium articles...

  • ... hit the sweet spot for article length; or
  • ... are better written than articles of other sizes.

As a closing comment, we can now tell that by adding the custom dimension, we know that all article sizes serve a purpose on GEEK:

  • Small articles generate the most article opens;
  • Large articles generate the highest number of read words; and
  • Medium articles are the best performing articles.