Infrequent updates gives increase of bounce visit rate

Bounce rate is one key metric that speculation hovers around, while interesting in itself it is far more interesting when combined with filtering of other metrics such as entry type of traffic (search engine, direct, or referral). That results in a rather interesting conclusion, but the calculation itself needs some work.

Bounce rate 101

Overall bounce rate increaseThe commonly promoted variant of bounce rate on a slew of sites is based on a simple calculation; the total number of browsers with a single page visit divided by the number of total entries. While this may seem proper keep in mind that long term analysis means that cookie inflation skews the numbers.

Bounce rate the better way

A far more optimal solution is to base the calculation on the total number of bounced visits instead and divide by the number of total entries. That way you can be sure that cookie deletion does not inflate or even make your results completely misleading.

Time to write articles is a resource that seems to not be available in the required amount for keeping this website updated on a frequent basis, and that lack of resource makes it is easy to find the impact when it comes to bounce rates in reports easily created using Digital Analytix which is the preferred tool used here.

Typically the release of a new article decreases the bounce ratio due to the visits of returning browsers that find them of interest, given the infrequent release rate this makes the numbers a bit jumpy on a weekly basis. But, as in the picture to the left, a monthly overview reduces this issue. Evidently the numbers are percentages, just didn't bother formatting them that way.

The most recent article was out in August 2011 and caused a visible drop in the bounce rate. Then as the months passed with no other new articles being released it started to grow hitting over 50% in the first two months of 2012.

Quoting Avinash Kaushik:

...really hard to get a bounce rate under 20%, anything over 35% is cause for concern, 50% (above) is worrying.

As always, it depends on website type and purpose, but also on the types of traffic the website has.

Found in the numbers

Search engines vs. all other traffic bounce rate comparison

What the filtering, of browsers into two groups, makes evident is that the amazing browser clicks from search engines (i.e. Google) consistently bounce more than direct / referred browsers. Does that mean that that traffic isn't any good?

The answer is no it can still be, however when working with bounce rates one should always split up search engine referred browsers and direct / referred browsers for the sake of clarity in the analysis.

For such a small site as this one what is sure is that the frequency of bounces has a direct connection to the release schedule of new articles. Given the results there is still no reason for neither concern or even worrying, the cause of an increased bounce rate is evident and quite easy to sort out given if more time is available.

Using total visits instead of total browsers eliminates the impact of cookie deletion, and allows for long term bounce analysis. If you know what causes the bounce rate to go up, the only reason to worry is if it has an direct impact on revenue or sheer ego.

External link:

Bounce rate on Wikipedia
Bounce Rate Demystified on Kissmetrics