What’s a bounce rate and why should you care?
NOTE: You can skip this section if you already understand bounce rates.
In the analytics world, a bounce rate is the ratio of visitors that leave (bounce away from) your site after only visiting a single page. The lower the number, the better for your site - because it means less people leave after only seeing just one page.
The 4 main ways that visitors can bounce from your site is if they close the browser tab, type in a new URL in their address bar, hit the back button, or click a link to an external site from your page. If your pages have links to other pages on your site, visitors following those links will not be bouncing.
The reason for this is because your analytics software will have 2 data points for that visit and it will know that the person remained on your site. See this post about 0 second visits according to Google Analytics to get a clearer picture of what I mean.
Different types of websites have different bounce rate averages, so if you’re averaging on 80% then don’t panic just yet. Landing pages can average between 70% - 90%. That’s because they usually have a single call to action and don’t provide links to other pages on your site. Content websites can average around 50%, and portals such as Yahoo! and MSN can have bounce rates as low as 10%.
Decreasing your bounce rates
Aside from making your content high quality and relevant to the right kind of visitors, there are a few other ways to decrease bounces from your website. One low hanging fruit (urg, unbearable clichés) is improving the UX of your site. Put yourself in the shoes of a person visiting your site and objectively evaluate your experience.
Pop-up adverts, surveys, auto-playing music or videos, and newsletter subscription pop-ups all contribute to the total annoyance factor of your site. People hate being annoyed and will leave once their threshholds are crossed. Ugly page designs can also contribute to bounces (there are exceptions like Craigslist). Pages that seem unorganized and confuse visitors don’t help either.
Links to external sites also increase bounce rates. The current trend is to write long-form articles because it’s better for SEO. These articles also include tons of links to external resources. You should be linking internally as well to help mitigate this. Here’s some more information about bounce rate factors.
Real world example
Let’s take a look at my Google Analytics data for August 2015 for Mobile Chat Groups. The screenshot below shows the visitor acquisition and behaviour data broken down by channel. You’ll notice that organic search has the lowest bounce rate.
This makes sense because the site is a directory for mobile chat groups and the main queries used to get there are representive of demand for these listings. People searching for chat groups will naturally spend some time browsing the site to see what groups are available. This segment also has a decent average session duration which is expected with a lower bounce rate.
One thing to note in the screenshot is that a huge number of visitors are first time visitors. That’s significant because if only 25% were new sessions then the low bounce rate would be misleading. Returning visitors would be less likely to bounce because they already understand the site, maybe they’ve signed up, or maybe they’re back to browse for new listings.
A 92% ratio of new sessions and a 37% bounce rate means that people that have never seen the site before aren’t bouncing away from it after the first impression. That’s great! I can’t really say the same for user retention though, yikes!
Exercises for the reader
Dig into your own analytics. View the breakdowns for each channel, for your top 10 or 20 pages, and for visitor context (mobile, tablet, desktop). Then play around with those elements and see how various combinations affect bounce rates for different segments.
Do desktop visitors stick around longer than mobile visitors? Do desktop visitors stick around longer than social referrals via mobile devices? If you have significant differences (with statistically significant sample sizes) then introduce multivariate testing to see if you can improve the poorer performing variants.
Don’t just look at your overall site bounce rate because that overall is made up of many, many different bounce rate segments. Spend some time looking at the granular data because improving each of them by even a percent or two will improve your overall bounce rate.
My guesses on why the BR is so low
So why does the Mobile Chat Groups site have such a low bounce rate? It could be because it’s a content site and those average around 40% - 60% bounce rates. However simply being in a certain category doesn’t guarantee you that category’s average bounce rate. So aside from the type of site that it is, it must be getting something right. I’d guess that some of the contributing factors are:
- The site is not ugly.
- The value proposition is pretty clear.
- Organic search traffic is relevant and interested (according to GA's search terms).
- The instructions for using the site are simple and visible.
- Navigating the site looks easy and unambiguous, even to new visitors.
- The search bar is prominent (and used frequently according to GA).
- There are no adverts or annoying newsletter forms on the site.
- There are no external links.
- The page load time is pretty quick.
- The site works damn well on mobile.
What are your thoughts? Think I’m wrong? Think I’m right? Tweet me.