Bounce Rates: Do They Matter?
“Bounce rate” and “pages per visit” are two of the most prominent statistics displayed by the Google Analytics “Dashboard.”
Don’t let that confuse you: they’re not often not relevant. Here’s why.
This week, while viewing the Google Analytics Dashboard for my LessonIndex.com web site, I noticed that during August, the bounce rate had increased by a few percent, and average pageviews per visitor had declined slightly. (“Bounce rate” refers to the percentage of visitors who view just one page at a web site.)
And those rates might seem dismal to many folks: 90% bounce rate, 1.3 pageviews per visitor. For an e-commerce retailer, or a community forum web site, these would be horrific figures — but for LessonIndex.com, they’re fine.
LessonIndex.com is a web directory site, and most visitors are referred by search engines, arriving directly to a specific “literature title” page. I’ve designed these pages to be self-contained, so that a teacher searching for a Huckleberry Finn lesson plan immediately sees all the related resources on a single page. She never needs to view any other pages at LessonIndex.com, but will click on links to the resources she needs, on other web sites.
Likewise, customers find many e-commerce retailers’ web sites through search engines, often while searching for a specific product, and the search engines often deliver the customer directly to the page describing the product the consumer wants. Of course, to complete a purchase transaction takes several extra steps (add to cart, check out, fill in form, payment, confirm), but most visitors will still exit the merchant’s web site immediately, resulting in “bounce rates” of 60% to 80%, with average pageviews-per-visitor often under 2.
And an online community forum may also draw significant traffic from search engines, with visitors shown a specific discussion thread that answers their question (or isn’t relevant); again, the bounce rate might be over 50%. However, visitors who stay and explore the community forum site often view many different pages, so even with a bounce rate over 50%, the average pageviews-per-visitor is often 3 to 5.
But the raw numbers can be highly misleading, because they combine all sources of traffic into a single metric. As I mentioned, most visitors to LessonIndex.com (90%) are referred by search engines, and arrive on a specific literature-title web page that matches their search phrase. What about the other 10%?
Some visitors (6%) are referred by other web sites which link only to the home page, and others (4%) type the URL directly into the browser window. When I “drill down” to view data only for the home page, the bounce rate drops below 40%, meaning that “more often than not,” visitors click through from the home page to a specific literature-title page at the site.
So what happened that made my “bounce rate” increase, and my “average pageviews per visitor” decrease? During August, search engine referrals increased faster than other sources of traffic (rising from 87% to 90% of total traffic).
That’s an important piece of information: I’d prefer to have a more diverse mix of traffic sources, to reduce the impact of changes in search-engine algorithms.
But when I adjust for this change, the “bounce rate” and “pageviews per visitor” haven’t really changed at all.
Finally, it’s critical to remember that nearly all of the data reported by Google Analytics are “intermediate metrics,” which can easily be manipulated in ways that don’t actually affect the site’s more meaningful measures — including “profit.” (Lies, damned lies, and statistics.)
I do think bounce rate & pages per visit are difficult metric to use as benchmarks. Sometimes they are useful, but rarely in my experience.