Up until I decided to make the serious effort to increase subscribers, this site had two links to my RSS feed. One on the header that I added merely for formality, and the other on the sticky footer which I assumed would end up generating the lion’s share of clicks to my feed. My thinking was that since the RSS link on the sticky footer was persistent, larger and more visual, it would invite more clicks than a small and simple button in the top navigation. I was unbelievably wrong. Not only did the link in the header outperform the link in the sticky footer, it was outperforming it 2.5 to 1. From the data I was pulling in, I came to two conclusions: 1) I literally had no idea how people were using my site at the micro level sense my assumptions were so off and 2) the amount of clicks to my RSS feed was significantly lower than I was expecting to see. Due to the numbers (or lack thereof) that we coming in on my two RSS links, on May 18 I added a very large subscribe message just below the navigation bar. That made a significant difference in the percentage of visits to RSS clicks — in fact, I received significantly more clicks in 10 days with the promo link than I did in almost a month without it. The unfortunate side effect is that I devoted a tremendous amount of space towards promoting my RSS feed and I still was not getting the amount of clicks I was hoping for. I continued to tweak the links on the site to varied success.
Throughout these two months, there were 4 significant periods in which the site went through which saw a very different click rate on RSS links. Below is a slideshow of how this site has changed in the past month. Click through the images to see the changes throughout the month.
Below is the complete breakdown of how each link performed through each of the 4 periods. An interesting thing to observe is how the header and footer links’ click rates continually drop with the added emphasis on the promo link. This is a perfect example of how individual elements in a layout do not exist in a vacuum — the elements you place on a screen are going to affect those surrounding it. We all are undoubtedly aware of that in a theoretical sense, but rarely do we as interface designers look to the numbers to back it up. The other highly fascinating finding is how the most dramatic increase in clicks was due not to any changes in design, but by simply having the subscriber count increase.
April 22 to May 17, 24,227 visits
Link Clicks RSS Click % Click/Visit % Header 25 71.4% 0.103% Footer 10 28.6% 0.041%
May 18 to May 28, 9,309 visits
Link Clicks RSS Click % Click/Visit % Header 29 70.7% 0.311% Footer 9 21.9% 0.096% Promo 3 7.32% 0.032%
May 29 to June 13, 12,467 visits
Link Clicks RSS Click % Click/Visit % Header 46 85.2% 0.369% Footer 5 9.62% 0.040% Promo 3 5.77% 0.024%
June 15 to June 22, 8,986 visits
Link Clicks RSS Click % Click/Visit % Header 57 78.1% 0.634% Footer 11 15.1% 0.122% Promo 5 6.85% 0.056%
Aggregate Click Rates
More Questions than Answers
After two months of desperately working to get people to subscribe to this site’s RSS feed, I am still unsure how well my work has paid off. I have recorded 331 clicks to this site’s RSS feed since I have been recording, but what does that even mean? Just because someone clicks on the RSS link does not guarantee that they have decided to subscribe. Secondly, how many people are subscribing through different methods, such as using a bookmarklet or Firefox or Safari’s built-in subscribe button. Lastly, and most importantly, FeedBurner’s analytics is useless. I have had the number of subscribers vacillate by 400% (up and down) in the past 2-3 weeks. This does not mean that this process was useless — far from it.
First and foremost, my opinions on how far pure intuition can take you have been changed forever — and by that, I mean I do not trust my intuition. Your gut feeling is a great place to get started, but at some point your gut feeling must be validated with numbers. Second, I loved observing how these links clearly exist in an interconnected ecosystem. The statistics show that adding another link, no matter how effective it is, will not simply add more clicks. In my case, the percentage of RSS clicks increased dramatically after adding the large promotional link. However, as expected, the other two links’ click rate dropped considerably after the promo link’s introduction. Third, I am starting to consider that content is not king, it is king, queen, prince, princess and the entire kingdom. When you see roughly a 200% increase in click rates simply based on the subscriber count changing, it is hard to ignore that. Lastly, we need a real A/B testing plugin for WordPress that does not force us to swap out themes to simply test a different header. Everything I have presented here is open to statistical scrutiny due to the method I was forced to record it. We should not have to go to hell and back simply to test a button against another.
In the coming weeks, I will have some more on this specific experiment — dealing with these stats in a a much more analytical manner. I am also going to work on increasing the performance of an area where I can better record the actual results. There has been an interesting experiment to get Twitter followers — I think I may have a few things to add to that conversation. I am hoping to try out Genetify to run some of these future tests so I can actually test my site like all the big boys and girls do — that should be worth a post in and of itself.