Here at Pure Visibility, analysts recently used competitor data to find a decrease in Time on Site for a client competitor’s site. Is this evidence of an improved site design, perhaps a redesign? Happy visitors, quickly finding what they need? But isn’t Time on Site also sometimes used as an indicator of successful engagement, in the absence of Conversion Rates, to determine whether a page is successful? (albeit with a bit fuzziness due to the way time on site is calculated online). Should we celebrate a trend toward long sessions, or cringe in horror?

Time on Site

There are quite a few factors influencing the length of sessions – the site objectives, the presence of detailed informational resources on the site, the visitor’s personal tendency to linger. There are the goals of the visitor, and the goals of the site creator, which may be at odds in some situations. Within social media environments, for example, the goal is often to support engagement. On a government site for renewing license plates, a shorter session is preferable – who wants to hang around once they’ve completed the form?

First, the objective. For some sites, the length of time a visitor spend may not be important as long as the conversion happens in the end. For example, many well-designed sites provide quick access to forms (meaning conversion), while at the same time containing detailed pages on the company’s products/services/history, etc. Sure, it’s great if some visitors spend some time reading these – it will likely increase the probability of conversion, provided they like what they see.

In finance, there is a way of analyzing data like this in terms of the distribution of the probability that an action (conversion, here) will happen in the end, plus the probability distribution that a session will last over a given amount of time without the conversion happening. A technique like this is ideal for engagement on sites where time on site is only important so far as it affects conversion probability. In situations where long times are shown to have a negative effect on conversion rates, we could make a reasonable assumption that the cause might be time spent searching for nonexistent information/links, or time spent recovering from errors in navigation, text input, etc.

Finally, the most accurate insight on the true meaning of time on site is best achieved with a good estimate for the amount of variance in the individual visitors’ times on site. Discussions of credibility in web design like this one refer to different types of decision makers – slow vs fast. Some variation in time on site will most likely be seen even in cases where all visits lead to conversions.