Lessons Learned From AnalyticsFebruary 28, 2017
By Joe Rstom, Interactive Director
On February 1, we launched the new Fast Horse website – with little fanfare – and over the past month I’ve been watching the usage statistics in the hopes of distilling some lessons or perhaps actionable improvements. It’s always interesting to see how a new interface guides the user experience in different ways, and how various calls-to-action determine how new visitors interact with your content compared to returning visitors, or even new visitors in the prior interface.
This particular exercise shows us quite a few areas of improvement, even if they are marginal. I’m merely comparing the month prior to launch and the month after launch: Jan. 1 to Jan. 31, and Feb. 1 to Feb. 28. Here are some observations:
- Sessions and pageviews: We saw a 13.25 percent and 15.83 percent increase in these numbers, without significantly increasing the percentage of returning visitors.
- Pages per session and bounce rate: While we increased pages per session, we also lowered the bounce rate, which is the likelihood of a user leaving without an interaction.
- Mobile vs. desktop vs. tablet usage: We saw a greater increase in mobile and tablet usage. Most notably, the pages per session in these categories went up, which indicates improved usability on those devices.
- Page speed: We’ve decreased the page-load time an average of 0.5 seconds by optimizing scripts and files.
- Key interior page traffic: Most of the interior pages traffic scaled in a linear fashion with the site-wide statistics, but after strategizing the contact and careers sections, each saw a substantial increase in traffic and interactions.
- Improved SEO and site-mapping: By optimizing the site and how it’s indexed by Google, we saw a lot of new organic search traffic to blog posts on the Idea Peepshow. Most notably, Tony wrote about his trip to Hawaii a few years ago and somehow that’s shot up on the Google search results this year!
So what are the takeaways? What do these statistics prove and how can we push it even further?
- Design and copywriting go hand in hand to optimize calls-to-action. The hard work up front clearly pays off, but you also need to design an interface that can allow you to pivot quickly. We’ve done that and can continue to work on changes that don’t require a ground-up rebuild.
- By allowing organic traffic to test the new interface without an influx of users who weren’t sent there via a public announcement, we get to measure traffic more objectively. Typically this can only be done months after a new website launches, and you have to compare statistics 12 months apart. You can miss a lot of opportunities for improvement in that time.
- Look out for confirmation bias. While we noticed quite a few signs of improvement, this did also confirm certain suspicions about content that we spent a whole lot of time designing and refining. So we need to reframe the way we look at these statistics to understand why, without making assumptions.
- Don’t go to Waikiki. The improved sitemap and Google indexing resulted in valuable blog content rising from the grave. And all jokes aside, this is the sort of unexpected success of a redesign that justifies all of the effort.