The Hope of a Link: Mashable Content Analysis

October 29th, 2012 - Posted by Kate Morris to Content & Blogging

The author's posts are entirely his or her own (excluding the unlikely event of hypnosis) and may not always reflect the views of SEOmoz, Inc.

The holy grail: A link from Mashable.

Don't kid yourself. We all want one. They are a content powerhouse with a mind-blowing community. Moz has that type of community, but Mashable touches on everything from kittens to major trends in the economy and technology. If it's interesting to the Internet community, they'll publish a about it. Their domain authority and homepage authority is a whopping 96. Of course we all want a link.

The first thing any reputable SEO would tell you is that you need two things before you can even begin hoping for a link from Mashable:

Great content - it has to be beyond good by this point, it must be great
A relationship with Mashable

I'm not going to claim that those two things are absolutely necessary, but I would agrue that you need to know more about Mashable before you go pitching your content to them. Well, I've done some of that for you, and today I'll teach you how to do it for other sites as well. This is Advanced Content Analysis ... starting ... now.

The Brain Child

The idea to do Advanced Content Analysis on Mashable came from a conversation Carson Ward and I had one day about getting a link from them. He made the quip that all you really need to do is write a post along the lines of "7 Ways to Do X." I laughed because it's sort of true (list posts do well), but then asked myself:

"How many of Mashable's posts are lists?"

Once I dug into how to get the titles of as many posts as possible, I realized just how much more data was available and how much deeper the analysis could really go.

Data Process

I thought I'd pull six months of Mashable posts; that is, until I started pulling the data and realized they put out almost two thousand posts a month. TWO THOUSAND. Holy content, Batman. Two thousand posts and one month was plenty for my analysis. If anyone wants to do more, I'd love to hear about six months of data.

I used three tools to pull the data necessary for analysis: ImportXML for Google Docs, SEOTools for Excel, and Microsoft Excel.

ImportXML

To gather the post titles, I used ImportXML for Google Docs. The linked guide will tell you more than I ever could about how to scrape content from a site using ImportXML. For those that are curious, I've included the formulas I used for Mashable. If you just want to see the forumlas in action, here is a document that is read-only for you to investigate.

The first important part is the page to scrape. You want to get to the blog/site's archive pages. For Mashable, you can access them by hitting "next" on their homepage. This gives us the page to scrape. A1 below is where you input the page number you want to scrape. This forumla is cell A2 for me -- you'll need to know that in a minute.

=CONCATENATE("http://mashable.com/page/",A1,"/")

Below that, you can complete the import. These three take up A3, B3, and C3. As previously mentioned, once I pulled the titles, I figured I'd just keep going. The URL is important for the things you can pull using SEOTools, so be sure to add that one. The date and comment number was important for in-depth analysis. The number of comments along with social metrics are the only real "success" metrics you can pull externally. I mean, unless Mashable wants to share the traffic numbers with me for each of the posts ... no? Dang.

Titles: =importxml(A2,"//a[@class='headline']")
Post URL: =importxml(A2,"//a[@class='headline']/@href") (muy importante)
Date, Author, Comment Count: =importxml(A2,"//p[@class='byline clearfix reviewer vcard meta']")

Side Note: Dear tech guys and gals at Mashable, sorry, I crawled tons, and others might, too. Though that is probably like .001% of your server traffic. :)

From here, you simply copy and paste into an Excel sheet, and keep going (change the page number in A1) until you get the number of posts you want.

SEOTools

This is my new favorite tools for a number of reasons, not least of which is the ability to use Regex in Excel. Are you excited, too? Well, it can also return social metrics using a URL, return the canonical URL, and so much more. If you haven't downloaded it, please do. And donate. This thing is worth it.

Alright, so I promise not to bore you, but I used SEOtools to do a few things:

Download the Facebook shares (Twitter wasn't being nice when I tried, but it's possible)
Return True/False if the post title included markers like Infographic or Video. Mashable is nice and gives us a marker in the title if the post is about something big like that.
Return True/False if the post title included a number. More on this later.

Microsoft Excel

This section is all in Excel. I did one more thing and checked to see what day of the week the post was made live. That's made simple with a formula like this:

=TEXT(WEEKDAY([@Date]), "ddd")

The [@Date] references the cell in the table with the date.

Content Analysis

Now I have the data, it's time to learn some stuff about Mashable and their content. After de-duplication (they post weekly recaps of videos, etc.), I analyzed 1,159 posts. Below is a look at the content types from what I could tell. Other includes smaller types like Audio and just plain text posts.

It looks like in the last month Mashable has preferred videos over infographics, which is something to keep in mind when deciding what content to develop and pitch. But this is just in terms of the number of posts. We have answered our original first question: how many of Mashable's posts are lists? 13% in the last month (in the chart above I am referring to "# Posts"). Nothing to cough at, but not as many as Carson and I figured. But what about performance of these posts? What days are they posted? And for that matter what about video and infographic posts? On to dataland we go.

Note: the "other" category includes all other posts that are not lists, or are something different as noted in the title of the post. Mashable has a habit of marking posts with [MODIFIER] which I used to break these categories down. All other are smaller tags and any other "regular" posts.

Lists, Infographics, and Video Posts -- Oh My!

How well do they perform? I'm picking on these because they are the content types that we all harp on so much. The results for comments on the post types and Facebook metrics. The most suprising find here is that while video posts have more shares and comments, there is a higher number of list posts. I think this has more to do with traffic and brand building than actual engagement. I theorize that if we could see average unique visitors to these posts, that lists would have a much higher view rate. Reason: people are lazy and just want to skim for data. Lists work well because they are easy to skim and give people information in just a few seconds. One day I'd love to see a study in which Mashable looked at the new visitors from a list post and watched how often they came back to Mashable, and compare that to other post types. THAT would be fascinating.

Now to the more fun stuff - days of the week. This has little consequence in my book, but it's fun to look at.

Above is the average for the site overall in the last 30 days. Fun factoids:

Mashable posts less on the weekends, but the audience is more active in comments and sharing. Weekend numbers are similar to Tuesdays. Tuesday is apparently the day that Mashable audience members are bored at work. But more likely, there are just fewer posts those days. Averages run higher when you have a smaller base.
Monday is the big post day. No surprise to me there.
Wednesday and Thursday are when intereaction drops off.

What about by post type? How many are posted on each day? It turns out that the videos and lists are posted more on Mondays, and infographics on Wednesdays. Why do you think this happens? I have my theory, but I'd love to hear yours on this one.

How did these content types perform per day? Let's just look at the "lists" category for this part as we are getting pretty deep, and this post was originally about lists.

Remember my theory on boredom on Tuesdays? *points like Vanna White at the chart above* See?!?!?!? The count of total shares and comments back my theory up. Even though there are fewer posts on Tuesdays, they got more interactions on average. Huh. Maybe this is why we post on Tuesdays and Thursdays at Distilled. Will and Duncan are smart cookies. (No, really, I doubt that had a ton to do with it, but Will or Duncan can correct me later.)

Final Thoughts

Content Analysis is much easier with tools like ImportXML and SEOTools for Excel.
You should consider doing this with the content on your blog or a competitor. What's really working for them? Back up your analysis with hard core numbers of interaction with the audience, not just what you see them doing. Don't get inspiration (see I didn't say copying?) from something that isn't working.
Post awesome stuff on Tuesday.
Develop great videos for Mashable, as infographics are on the way out.
List posts get attention, but not as much interaction.

What else did I miss? This isn't fully scientific, but it is fun and gives us much more data to help make decisions than we might have had before. Take everything above with a grain of salt, pull your own numbers, and see for yourself.

http://www.seomoz.org/blog/the-hope-of-a-link-mashable-content-analysis

NSG Consulting Inc.

Thursday, November 1, 2012