According to a Pingdom demographic study of readers of the top 100 blogs (according to Technorati), the typical reader of the world’s top 100 blogs is 38-years old. The study analysts were surprised by that statistic, and that is not surprising.
Why? Because the data used in the study is questionable. However, that’s not to say the study is completely irrelevant and should be ignored.
Let’s take a closer look.
The Study Results
First let’s review some of the key findings. The Pingdom study found that 38 is the median age of the typical reader of the top 100 blogs (out of the hundreds of millions of blogs online), and the average reader is 41. Some specific age and gender-related results from the report include:
- The blog with the largest share of older readers is The Caucus from The New York Times with 2% of its readers identified as over the age of 64.
- On the other side of the age spectrum are Engadget, Joystiq, and Ars Technica, which have no readers over the age of 64.
- The only blog that has no readers younger than the age of 25 and no readers over the age of 54 is Ars Technica.
- The average blog has 45% female and 55% male readers.
- 59 of the 80 blogs studied were read primarily by men and just 12 were read primarily by women.
- Hollywood Life has the largest share of female readers (63%).
- On the other side of the gender spectrum is Android Authority which has the largest share of male readers (70%).
The Study Methodology
Before you start saying, “I read Ars Technica and I’m under 25,” let’s take a look at the study methodology.
The data from the Pingdom study comes from the Technorati Top 100 blogs list. I know that at this point your thinking, “Does anyone still use Technorati?” or “Is Technorati still around?” The answer to both questions is yes. Is the data from Technorati the most reliable data in 2013? The answer is no, but in the absence of a better option, Pingdom chose to use the Technorati list of the the Top 100 Blogs as the basis of its study.
Next, we need to take a look at where the data came from, and here is where we have another problem.
The data related to age and gender came from the Google Display Network Ad Planner tool. The data available from this tool is good, but it is far from perfect, particularly since Google made some changes to the tool in September of last year (follow the link to read the email sent by Google describing the changes, which was published on Search Engine Watch).
The first problem with this source is that data is only available for blogs which are part of the Google Display Network. Second, Google infers much of the demographic data about sites in its Google Display Ad Newtork, which is made available to advertisers using the Google Display Network Ad Planner tool. Thus, the Pingdom data used in this study relied on a great deal of inferred data. Here is the explanation about how Google generates demographic data from Google Support:
Google Display Network Ad Planner demographic information is generated through demographic inference algorithms that combine third-party demographic data with Google sample data. (Demographic data is currently available for some countries only.)
The September 2012 changes were a step to remove Google’s reliance on some of that third-party demographic data and begin using data it has access to through its own sites and tools.
With that said, it’s clear that the data in the study is flawed. Scroll through the comments on the Pingdom blog post reporting the study results and you’ll see many comments and tweets questioning the data. Even Ars Technica Founder and Editor-in-Chief Ken Fisher took to Twitter to respond to the study results. You can see one of his tweets (in response to a tweet by Andrew Yang of MacWorld) below.
— Ken Fisher (@kenfisher) March 2, 2013
It’s like visiting Alexa, typing in your blog or website URL and reviewing the demographic data that comes up. Much of that data will be far from accurate. Therefore, the study results are only as good as the data used to create those results.
Knowing that the data isn’t perfect, should we ignore this study?
I think there is still some value in the data, but don’t bet the farm on it. If you compare it to the same report Pingdom published last year, you can do some trend analysis, and you can use the data for guidance. After all, the data isn’t completely wrong, it’s just telling a very small part of the story and combining it with algorithmically inferred data. Use the data with extreme caution.
What do you think?
You can follow the link at the beginning of the post to see all of the Pingdom study results and charts.
Image: Svilen Milev