Don't Get Tricked by Misleading Analytics
Often we get clients who bring in a months worth of high-level analytics and will try to use it as a barometer for their success or failure. "My bounce rate is 45%, time on site is 3:20 minutes and pages per visit are 4.3."
Now, these aren't bad metrics, but looking at them from such a macro standpoint is a waste of time.
Why? Two main things, in my opinion, need to often be removed from the equation.
Aggregate versus specific metrics
I tend to look at bounce rate of important landing pages, such as the home page, separate from pages as a whole. On many of our sites we have pieces of content that will just tend to have a higher bounce rate. On Insivia's site we have an article in our resource center about customizing your site's facebook thumbnail which actually gets about 30 - 100 visits a day. But the truth is, many of the people who visit this article are not good prospects and are only interested in that one answer. The bounce rate for that page is 75%.
Well, looking just at our home page and not the site in total which includes pages like the above article, our home page's bounce rate is around 39%. Now we are talking. This is similar to other data such as pages per visit and time on site. By only looking at my main landing pages that typically get more targeted prospects, I can see real information on how my site is doing.
If I use bounce rate change over time as a KPI (Key Performance Indicator), I want to look at it not at a macro level, but a micro level.
Time/Day of Visit
This is only a recent change I have made to looking at analytics, but I have been really watching how I factor weekends into my interpretation of analytics. I recently did a comparison of traffic for two days during a week versus the weekend. Here are some things I found.
Visits to my main landing pages were half what they were on weekdays, but visits to articles were exactly the same. What this tells me is that most likely on the weekends people are searching for answers to accomplish things but not searching to do business. This means for two days every week, my overall site bounce rate skyrockets. So, if I factor that into my overall month from a macro level, it is really skewing the information.
The key thing to take away is that you can't just look at data from a macro level and think it tells you a story. Digging deeper into your analytics is the only way to see the real situation and always want to make decisions with the right information.
* Side note. I think critics would say that you want to convert even traffic on articles or we shouldn't create content will drive non-targeted traffic. That's true and we try to employ techniques to get people to dive deeper into the site plus our end goal is having content valuable to our target audience. But some pages will always be better for judging success than others.