At a very cursory level, bounce rate is designed to tell you if your content is meeting your audience’s expectations or not. But if there’s one thing I’ve learnt over the years, it’s that this metric should never be taken at face value.
Whilst Google describes it as the percentage of single-page sessions, it is more broadly accepted to be a visit with a single engagement or interaction hit, which may include a pageview, but can also include an e-commerce tracking, event tracking, or a social interaction hit sent to the Google Analytics server.
For the sake of clarity, there’s always Avinash’s more succinct summary: “I came, I puked, and I left”.
Yet the trap that many e-commerce sites fall into is seeing this metric and immediately making assumptions, without fully understanding the context within which it is being calculated, or the numerous external factors that may be influencing it.
To name just a few, bounce rate can be affected by:
- Page type
- Seasonality of content
- Site design
- User’s stage in the buying cycle
- Page speed
Faced with such a number of variables, the key to getting meaningful data and insights from bounce rate is ‘relativity’. Considered in isolation, bounce rate is all but meaningless. But considered relative to the likes of page speed, device or channel, this metric can offer rich, quantifiable insights into your content performance. So how exactly do you do that?
Experiment With Custom Reports
Example: Page speed vs. bounce rate
Page speed and bounce rate go hand in hand. The two are so closely interlinked that to consider one without the other would be futile. As page speed varies by device, so too does bounce rate. As such, bounce rate should form a key metric underpinning any mobile optimisation strategy, not least because it acts as a strong qualifier for optimisation in the first place – just in case you needed any further convincing.
To analyse the correlation between the two in more detail, I commonly use a great custom report recommended by Avinash Kaushik. It’s quick and easy to set up, as follows: