Although I would listen and argue the point that this is a useful metric which indicatively demonstrates how long users are spending on the page, I think that we have unquestionably led into believing this as a meaningful data point or metric. We accept it on face value without questioning it or perhaps because it is more commonly miss-interpreted. In the interests of understanding usability and it’s impact on user behaviour I think that we should steer well clear of using this metric singularly.
How can you prove categorically that time by itself provides an indication of quality or success? It is easy enough to mistakenly assume that one figure is ‘good’ or that another is ‘bad’. We sometimes assume that more time equals a successful visit, whereas the opposite might have been true if we might had presented the user with what they needed easily and efficiently. In contrast, a user may be being forced to spend a lot of time on the site because the navigation system is poor.
Having conducted many user tests I have witnessed different levels and speed through which people use the web (more specifically on desktops rather than mobiles or tablets). Some users are more reactive and just click around sites until they see what they need, potentially clicking up and down in the hierarchy many times; others are much more considered and need to validate their actions or navigation choices. The former behaviour is much more exploratory and expends more effort where as the polar opposite was like ‘watching my Dad use Ceefax’.
So average time is relative to the make up of audiences and their needs. Would you potentially be more considered about your banking activities online than you would for finding an answer to a question by exploring content. The point is that different audiences behave different in certain circumstances in addressing their needs and react differently to different circumstances, content and activities.
In addition to this fundamental point about sites or apps targeting different users and different needs, some pages are designed so that the user doesn’t need to linger. For example, there are plenty of pages which are designed to offer users contextual signposts. If time (by itself as a metric) is used to demonstrate the effectiveness of the usability of this page, how do you categorically prove that the page and the content is useful? What is the difference between 3.75 seconds on average as apposed to 45 seconds. You could inadvertently be putting barriers in the way of motivated users and interpreting this figure as an extended dwell time but considering it a success.
Similarly the page could be really effective at driving traffic to other areas of the site but you think that you need to increase dwell time by adding more additional content therefore obscuring the core message or activity. Even tabbed browsing can skew this number depending on the activities; comparative browsing habits are a good example of this.To understand these behaviours, we need to evaluate other metrics and instead use time as an indication not as evidence.
Time is only useful in contributing to the story of perceived user behaviour. If the page has a high bounce rate and very low time spent on it, it is obvious that this page hasn’t met users expectations or engaged them – that is a given (consider that the time recorded is only a much smaller percentage of the traffic that didn’t bounce). However, if the opposite is true how can we prove relatively and protractedly that a user’s behaviour is matched to our expectations.
In short ‘Time spent’ is only an indicator of user behaviour and is too ambiguous to provide a valuable and useful context on which to base a meaningful, compelling and revealing KPI. It would be more appropriate to rely more on ‘active’ rather than ‘passive’ metrics to reveal the story and user behaviour. Alternatively, there are much better ways to measure the effectiveness of usability. My advice instead look in to usability issues concerning context, content and interaction by user testing to find out what and why users are thinking the way they do.
Frazer Cooper 