How Data Can Change Your Blog


Blogging is an important part of promoting your ecommerce website. It increases your page ranking by flooding your website with fresh and unique text, including those long tail search phrases that can be difficult to work into product descriptions. It can give you valuable links, which bring organic traffic. It also helps to build your brand because it sets you up as an expert in your field. However, knowing the stats behind your blog can be immensely helpful because it can drive more traffic and allow you to write content that is more relevant to your audience. Here is a short guide to changing your blog—for the better, of course—with data.

  1. Gauge engagement. This should be easy to do with any blog program. Look carefully at which blog posts generate the most traffic and also the ones that get the most comments. These are both signs of customer engagement. You should also look at signs of disengagement, such as which blog posts have a high bounce rate. This is a sign that customers were interested in the topic but that the content did not meet their needs.
  2. Look for broad trends. Are controversial or mundane posts more popular with your audience? Do they prefer “how to’s” or industry news? Do they react better to text only, or are photographs and videos the key to their interest? Your blog data can tell you the answers to these and other important questions about your market… but only if you take a step back and look at trends.
  3. Choose content based on top keywords. If you are like most ecommerce website owners, you are used to the concept of keyword research. However, content for your blog can be a little different. It is easy in a blog format to use long tail keywords and to cater to ‘how to’ questions. In fact, the editorial flexibility of the blog format is one of its chief advantages. Choose your content by the top keywords in your field at that moment, and ensure that the blog is designed with the best practices in search engine optimization.
  4. Consider planning an editorial calendar. If your field is like most, the keywords that are most popular change seasonally. You probably already know what your top keywords will be this December (or, at least, you have a good idea). Use this knowledge to plan an editorial calendar that will allow you to take advantage of that season’s unique shopping behavior.
  5. Include a call to action. Ultimately, you are not running an ecommerce website blog to edify and inform, although those are certainly nice side effects. You are writing a blog to increase traffic and to increase profits. This means that a call to action is an essential part of every blog post. Link to actual products and services in your posts and keep track of which are most effective. This will help you to hone your blog into the sales tool that it is meant to be.

Are you using bad data?

If you are like most ecommerce website owners, you keep a close eye on statistics and data from your website. You may test different strategies and also keep a close eye on customer reactions to even the tiniest changes in your website. However, studies have shown that most people—businesspeople included—don’t really understand statistics well enough to know if their data is accurate or not. Here are the most common statistical mistakes that many of your colleagues are making:

  • Testing under different conditions. Here’s a common way of testing two different versions of a website: you try out landing page A in the morning, landing page B in the afternoon, and landing page C in the evening. Can you spot the problem? These times of day bring very different types of shopper. Even if one is a clear favorite, you will still need to keep testing. A better approach is to test two different versions at the same time. This will show you how they compare with the same market.
  • Making far-reaching inferences. So you have found that your traffic over the past few mornings by far preferred version A of your landing page. Does that mean you should switch completely to the winning version? Maybe… but maybe not. Perhaps your weekday morning traffic has very important demographic differences from your weekday evening or weekend traffic. Maybe you are offering a special sale that appeals to one particular group. Maybe you were mentioned on a Facebook page and are receiving a ton of traffic that is not comprised of ‘typical’customers. See all those maybes? Keep testing before you make any conclusions.
  • Ignorance of external events. You are going to see a change in behavior over weekends and holidays. You will probably see a difference after certain news events and other external events as well. These are not good times to run a test, because the results will not be representative of your typical day-to-day operations.
  • Testing too many variables at once.  It’s fine to test the color and font of your ‘Buy Now’ button; just don’t do both at once. You’ll never know which variable was responsible for a difference in conversions and you’ll be no closer to finding the magic mix that works perfectly for your ecommerce website. Ideally, you should test one variable at a time and gradually work toward finding the right color, font, size, etc.
  • Using a too-small sample size. In order for your findings to be applicable to the huge amount of people that makes up your audience, you need to have a rather large sample size. It requires some knowledge of mathematics to determine exactly how large, but Google Analytics offers a great tool here: Google Website Optimizer calculator This will tell you how long a test needs to be run in order for the results to be statistically significant based on your own site statistics.
  • Lack of data analysis. Once you have completed an ecommerce website test according to the best statistical practices, your job is not entirely over. You need to closely examine your numbers and look for confounding factors, micro-trends, and other important information. This will bring you closer to finding out what works for your ecommerce website… and what doesn’t.