However, this approach is too simplistic. Website bounce rate can be misleading most of the time. Taking an average bounce rate of thousands of pages is too generic to measure the health of your website. A better approach is to evaluate pages with a bounce rate above a certain mark. We evaluate pages with bounce rates above 20% to 25%. This is because you can assume that around 20% of your visitors land on your website by mistake. They typed in the wrong keyword or clicked on the wrong advertising. These are the visitors who leave right away. And these are the ones who should be responsible for a 20% bounce rate. Bounce rates above 25% indicate that visitors who are interested in your product and service are not finding what they are looking for.

Finally, optimizing pages with high bounce rates will only make sense if these pages receive large numbers of visitors. Optimizing a landing page that has a bounce rate of 95% would only make sense if that page gets thousands of visitors in a month. The minimum number of entrances a page should receive before evaluating its bounce rate will vary based on the size of the site and its average order value. We generally recommend the following guidelines:

These guidelines change from one client to the next, so you should develop your own minimums based on your data.

Most marketers rely on website average bounce rate as a general metric against which to measure the bounce rate of different pages of the website. If you decide to follow this approach, we recommend that you rethink how you should calculate your website bounce rate.

Website bounce rate is usually calculated by averaging the bounce rate of the different pages of the website. The theory is that with many pages on the site, data abnormalities will disappear. This theory, however, fails the real-life test. Figure 2-3 shows how bounce rate is calculated for a website that has 20 pages.

Calculating website bounce rate
Figure 2-3. Calculating website bounce rate

We recommend evaluating the top 10 or 20 entry pages to the site. These pages are usually also responsible for the most bounces. To determine whether your website average bounce rate might have some abnormality in it, you should calculate:

  • The percentage of entrances these pages are generating to your website

  • The percentage of bounces these pages are responsible for

These two percentages should be close to each another (+/– 5% deviation). If they are not, the top landing page might be skewing the overall website bounce rate. As a result, you should calculate the site bounce rate after excluding these pages.

For example, one of our client’s top 10 entry pages was responsible for 32% of their website entrances. At the same time, these top 10 entry pages were generating 38% of the bounces. The website average bounce rate was 47%:

The difference between the two percentages = (38% – 32%) / 32% = 18.7%

Since the two percentages were more than 10% apart, we recalculated the website bounce rate, excluding the top landing pages. We did this by taking the average bounce rate for all the pages on the site, except for the top landing pages; the real bounce rate for the site dropped from 47% to 41%.