Here’s why you should worry less about dwell time
It’s tempting to focus on the time your visitors spend on a page but Francesco Ciuci, Data Scientist at SessionCam, says think again…
Look at that rabbit staring out at the sea. What can you learn from knowing how long he’s been sitting on that bench? Is he dreaming of a new career as a sailor or contemplating the fate of the humble cod? Maybe he’s not thinking about the sea at all. Perhaps he’s just resting his rabbit legs before returning to his high pressure job at the Carrot Marketing Board.
Of course, the answer is that just studying the amount of time the rabbit spends in silent contemplation of the waves doesn’t tell you anything concrete about his intentions. It just tells you he was in that spot for that specific period.
The rabbit conundrum is a good way to think about dwell time. It’s a metric that doesn’t even have a definition everyone can agree on. For some, it’s about the total time spent on a page, while for others it’s about focus on a specific element or process. Dwell time is the Schrödinger’s cat of metrics — it can simultaneously be a measure of engagement and a struggle indicator.
Looking at dwell predominantly as a sign of engagement is problematic. Many of your website visitors will be aiming to get something done quickly. In that case, a long dwell time shows you’ve failed. You haven’t made it easy for them to achieve what they came to do and are less likely to convert them from a visitor into a customer.
Look at causes, not side effects
The old assumption is that people who buy things spend longer on your site than people who don’t, so targeting longer dwell time is a great idea. But let’s examine that a little more closely: Those visitors spent more time because they were already looking to buy or your content was already relevant to them. Dwell time isn’t the catalyst for engagement, it’s a side-effect of it.
We’ve analyzed hundreds of thousands of website sessions and high dwell times tend to indicate that visitors are struggling. They don’t interact with the page because they can’t understand it or don’t easily find what they’re looking for. Those sessions often feature pages with too much information or poor layout.
Conversely, our analysis of sessions with a low dwell time indicates that in those cases visitors find what they need to on the page. That allows them to more quickly complete their intended task. What you consider an acceptable dwell time will differ from page to page. You want a login process to be fast but expect completing a payment form to take a little longer.
Dwell time alone is a black box. It’s hard to work out whether a visitor is engaged, struggling or simply away from the keyboard. The distance traveled by the cursor is just one measure that can act as a proxy for duration and the visitor has to be present. Our studies suggest distance traveled is much more important in identifying struggle than duration.
Actively dwelling or just hanging around?
We’re looking at the distinction between active dwell and passive dwell. Active dwell is the period during which the visitor is still paying attention to the page. It can be detected through page interactions and tiny cursor movements within a time threshold. It indicates that a visitor is engaging with your content but may be feeling indecisive (hovering over links), interested (text tracking/scrolling) or frustrated (spending a long time within one element).
Passive dwell time is where a visitor isn’t even looking at the page. Just because someone spent 10 minutes on your page doesn’t mean they were engaged with it. They could have moved to another tab or stepped away from their device to do something else entirely. But passive dwell can be useful. On product pages, it suggests a visitor has switched to another tab to look for cheaper alternatives. At checkout, it can show when a customer has gone to find a valid promo code.
SessionCam’s machine learning algorithm considers dwell time as part of a wide set of customer behaviors. We know the time between interactions is an important factor in identifying frustration. When your visitors get confused, the delays between their interactions with a page are likely to increase. But we also see that dwell is too general a metric on its own.
We want you to have the whole picture when it comes to analyzing customer behavior. You shouldn’t be stuck guessing what your visitors are thinking as if they’re just so many rabbits, sat longing for the sea.