In the content marketing rush, it can be hard to slow down and think about an essential element of each piece of compelling content: its title. Yet titles feed onto social media sites and play a big role in search engine optimization (SEO).
Because of this impact, titles also influence the amount of traffic a content marketing blog garners, as the Content Marketing Institute recently found out. Over the past couple of months, I’ve been working with CMI to figure out exactly what makes a good title, and how to make every single title shine.
Forget guesswork and gut feelings — we dug deep into CMI’s analytics to find out exactly which types of titles made the traffic roll in. Of course, many factors play into blog post traffic, but our research shows clear correlations between certain types of titles and higher or lower page views.
While some of CMI’s research was specific to its own keywords and audience, below I’ll share the steps you can take to perform similar research for your blog. (Be on the lookout for a future post with tips on everyday ways to improve your titles, as well as insights into some of the trends we found in CMI’s data.)
12 steps to creating a title report
1. Assess your data’s volume and availability: Do you have enough data to derive meaningful correlations? We had a wealth of data to work with because CMI publishes daily, but a study of at least 100 blog posts could probably begin to highlight trends or patterns.
2. Understand the effect of SEO: Blog posts pick up traffic over time as they’re indexed by search engines and pick up back-links. While we can offer no perfect solution to ensure consistent data, we’d suggest using blog posts that have been online for at least a month. This way, every post has the opportunity to accrue some back-links and organic search traffic.
3. Don’t bite off more data than you can chew: While it’s important to look at a significant amount of data, be forewarned: There’s a lot of manual work involved in cleaning and sorting this data. Also, analyzing more blog posts likely means using data from over a larger time spread. Overall traffic to your blog has probably changed over the past few years, so try to keep to a brief, recent time period for your analysis.
4. Download and clean your data: You’ll need to have your data in a flexible format so you can make calculations. Most analytics software should let you download your data into a spreadsheet format. Once you’ve got your spreadsheet, line up post titles with the appropriate metrics for the post. Keep in mind that this can be difficult in some circumstances — such as if you sometimes publish the same blog post on two different URLs, or if your “page titles,” (also known as “SEO titles,”) are different than your “post titles,” or headlines. For example, with Google Analytics, you’ll be able to download page titles, but not headlines. We’ve run into this issue a few times ourselves, and to ensure we were looking at actual headlines, we visited every single URL to double-check.
5. Delegate the data cleaning: If you’re like me, you’re probably freaking out thinking about how many hours you’ll have to spend staring at a spreadsheet to clean up all this data. Here’s my secret: To handle the data cleaning in an efficient manner, I leaned on Fancy Hands, my virtual assistant firm. They were a huge help on this project. You could also try Mechanical Turk, or a detail-oriented intern. Reward yourself for saving the time by writing a blog post!
6. Establish a baseline: Once you’ve compiled the data for all of the blog post titles you’ll be analyzing, take the overall averages for all posts’ performance. To do this, you’ll have to decide which metrics are most important to you (e.g., page views, social shares via Facebook, email subscription rates). While we analyzed many metrics at CMI, we focused on page views to draw our conclusions. Moving forward, you can compare every title type to this baseline.
7. Decide which title types to analyze: When I say “title type,” I mean any sort of repeated pattern or style in your title. For example, we looked at 16 title types, including the use of keywords, numbers, question marks, and even brand names. Of course, you can’t analyze the performance of a title type you don’t use, so make sure you’re basing the study on the types of titles your blog has published in the past. Scan your list of titles to see which types you use in practice. And if there’s a title type you think you should be using — well, get using it! But you may want to consider waiting a few months before starting your report if you’re not confident your titles can be separated into “types.”
8. Separate the data for each type: I’d also suggest delegating this task — man, it’s a hard slog. In this instance, I again depended on Fancy Hands to sort through the data. Types like “begins with a number” and title length were easy to separate automatically on Excel, by using the sort function and the length formula, respectively. However, title types like “uses a person or brand name” involved a human touch. For titles like this, I read every single title, highlighted the titles the type applied to, and copied and pasted them onto a separate sheet — but you should use whatever process works best for you.
9. Average the data for each type: Here’s the fun part! If you’re as much of a geek as I am, you’ll love comparing the average page views for each title type to the overall average. Your gut feeling that posts with numbers perform better? If your numbers are like the Content Marketing Institute’s, and those posts get more page views than the average, they’re confirmed! But you could also find that factors you thought made a difference actually led to fewer page views. We’ll go further into analysis in my next point, below.
10. Analyze your results: Raw numbers are fun, but take the time to break down each title type. What percent higher or lower than the post average are your average page views for each title type? I’d also suggest incorporating how many titles you have for each type in your analysis. I didn’t calculate the statistical significance of the data, because we’re dealing with correlation and many outside variables here, so this data isn’t exactly bulletproof (see my next point). Yet understanding the number of posts evaluated helped contextualize the results. You can feel more confident in data that’s attached to a significant sample size of posts.
11. Don’t take the data as gospel: Rather, you should use it as a guideline. There are a ton of factors involved in page title, and many ways the data in a report like this could be skewed. For example, perhaps your site is slowly becoming more popular, and you happened to use a certain type of title in the most recent month you’re studying. Well then, you may be misled to believe that the titles are the cause of the traffic, when really, you may have your growing email subscription base to thank. And of course, creating compelling content overall should be a much bigger priority than creating sensational titles.
12. Treat the report as a living spreadsheet: Now you’ve got a system down and have generated a lot of nice, clean data you can use on an ongoing basis. When you’re titling posts on a day-to-day basis and are wondering something like, “Should I use the word ‘research’ or ‘report’?” refer to your title report and use the data you’ve compiled to direct your decision. Go ahead and separate the data for the types relevant to your decision. You may find that you can actually go with statistics instead of your gut when you’re deciding which word to use.
So, when will you be starting on your title report? Have you done something like this for your brand already? What types of titles work for you?
Comment, and let us know!
Looking for more tips on using data to inform your content efforts? Check out CMI’s eGuide on Measuring Content Marketing Success.