Dwell time is one of those metrics you’ve probably heard of, and you might even think you understand, but you probably don’t. A ton of people make assumptions about what it is based on their knowledge of vocabulary, and it leads to false assumptions and mistakes in SEO. So what is it, really, and does it matter?
Bouncing like a Ball
Before you can know precisely what dwell time is, you need to know about bounce rates. There are actually two bounce rates for a page online, and they’re actually quite different. You have bounce rate, and then you have your actual bounce rate.
Imagine these three scenarios:
- A user clicks on your link in Google’s search results, your page loads, and the user immediately hits the back button.
- A user clicks on your link in Google’s search results, your page loads, and the user sits around for five seconds before hitting the back button.
- A user clicks on your link in Google’s search results, your page loads, and the user scrolls through your content for half an hour before hitting the back button.
In all three of those scenarios, a user clicks a link to reach your site. In all three of those scenarios, there is no second click. This is crucial! In order for an analytics program like Google Analytics to record a visitor as having done something on the page, the user has to actually do something on the page. If they don’t click a second time, there’s no usage data. There’s nothing to indicate a difference between person 1 and person 3 up above. Even though person 1 likely mis-clicked or was a bot, and person 3 was very satisfied with what they read.
All three of the scenarios above would count as a bounce. However, your site’s actual bounce rate is one third lower; that third user didn’t bounce, they were satisfied.
This brings us to dwell time:
- Dwell. Verb. “To linger over, emphasize, or ponder in thought, speech, or writing.” Alternatively, “to live or continue in a given condition or state.”
In our scenarios above, only one of the three users has any appreciable dwell time. They have spent time – half an hour in our case – sitting on your site. Longer dwell time is better, right?
The Dwell Curve
Have you ever seen a bell curve? When you chart a large number of users looking for certain statistics, you’ll notice that there are a few outliers at the top and bottom of the axis, but most fall into the middle somewhere. A very simple bell curve looks like this:
Dwell time also falls into a bell curve, when you’re thinking about value. Imagine for a moment that you’re monitoring dwell time on a single 2,000-word blog post.
At the left side of the axis, you have zero. There are people who click to your page and then leave immediately. More people than that will stick around for a few seconds, which is quite reasonable, since it can often take a page a second or two to load, and then the user another few seconds to determine if the content matches their intent. Longer dwell times are even more common: 30 seconds up to a couple of minutes can be enough to skim a piece of content and decide whether or not it’s valuable, and leave if it’s not.
At the peak of the bell curve you have the average dwell time for the average page, which will be a couple of minutes. It’s the length of time it takes the user to read or skim through you content to get the value out of it, before leaving satisfied.
As the bell curve starts to taper downwards once again, you’re seeing the people who are a little slower to use the web. These are the people who maybe opened two tabs and have come back to read yours later. These are the people who just read a little more slowly, or who were distracted part way through reading and had to come back.
Further down, where the curve tapers off again, you have the people who are spending ten, twenty, thirty minutes or more on a single blog post. No one takes that long to read a post, but there could still be some value.
Value at the long end of the bell curve diminishes because the reasons why a person might spend that long start to look worse. For example:
- The user might have opened five tabs and gotten to yours last; their dwell time is long, but they only actually spent two or three minutes reading your content before leaving.
- The user might be referencing your post for something in a discussion with a friend, or with a post they’re writing; longer dwell time doesn’t necessarily mean more interest in your post, it just means they didn’t want to close a tab with an information source before they were done. I personally do this all the time while researching and writing blog posts. I can spend hours on a page, technically, while my actual attention is only on the page for five or six minutes.
- The user might have opened the page and then left. People have to get up and go to work, or go run errands, or play a game, or anything else. Dwell time can be very long with very little value coming from it.
- The user might have opened the page, minimized it, and forgotten about it. A dwell time of 12 hours doesn’t mean your content is really good, it means the user forgot they opened it. The only pages that have continuous value in this manner are sites like Pandora, that will sit and play music for hours without interaction. Even YouTube on Autoplay still changes pages.
This leads us to the question of how value can be assigned to dwell time at all.
Time On Site Versus Dwell Time
Dwell time has a second crucial identifier that many marketers miss or simply aren’t aware of. That is, dwell time is the amount of time spent on a site before returning to the search results page. If the user closes the tab, if the user clicks to another page on your site, or if the user navigates to another site from bookmarks or the URL bar, that ends their time spent on site.
Dwell time only applies if the user then returns to the search results page, which is an action Google can track. So in the example above, with the person who spends twelve hours with the tab open, it’s not going to be dwell time. The user isn’t likely to go back to the search results when they’re done; they’re more likely to just close their browser, having forgotten what they were doing with the tab or having long since completed their task.
Dwell time is useful in a narrow range of circumstances. Unlike the general time spent on site, dwell time is something Google can use to determine how much a query was fulfilled by the site in question.
- If a user clicks through a search results link, then returns to the search results page in five seconds, it’s an indication that the result was not a good fit for the query.
- If a user clicks through a search results link, then returns to the search results page five minutes later, it’s an indication that the user was satisfied by the content, but may want to pursue further reading, alternative sources, or simply change queries.
- If a user clicks through a search results link, then returns to the search results page an hour later, it doesn’t necessarily measure anything. The bell curve strikes again; the user could have forgotten about the tab, or any number of other things.
So you have time spent on site, which is visible in Google analytics. You have bounce rate, which is visible in Google analytics. You have dwell time, which is… not visible in Google analytics.
Does Google Use Dwell Time?
As far as the official news from Google goes, we have dead silence. Google doesn’t really talk about most of their search ranking factors, they just clarify their moves and explain their algorithm changes when they’re big enough to matter, like Panda or Penguin. Google has not acknowledged dwell time as a metric they monitor.
WordStream offers two considerations they claim indicate that Google does pay attention to dwell time. The first is the short-lived pilot program where Google had a “block this site from future searches” link next to sites in the results page. Generally, the button would only appear if you clicked through to a site and then clicked back fast enough. WordStream – and others – claim this is a measure of dwell time. Too little dwell time, and the button appears, offering the user a chance to blacklist the site.
I personally figure this is less about dwell time and more about bounce rates. That link didn’t exist for very long, and when it did, it required extremely short visits to trigger. It was generally meant for the pages where you clicked through, realized the page was dead, was spam, or was off-topic, and then left. It wasn’t meant for content that was fine, if a little too short or not robust enough, both of which can trigger short dwell times.
Their second indication of dwell time hails from the days of Authorship. When Authorship existed, after reading a post by a given author, you could return to the search results and see more recent posts from that author. WordStream claims this was a reward for dwell time; longer dwell, more prominence for the authorship links.
I personally disagree here as well. It was an attempt by Google to add value to Authorship, to try to get more people – particularly site owners and widespread bloggers like Neil Patel or Kristi Hines – to adopt it. It had nothing to do with dwell time, and a lot to do with a reward for opting in.
The way I see it, if Google was using dwell time to reward authors, you would see other relevant pieces of content in the links provided. Google would want to incentivize more traffic to that author, to keep them creating more pieces of similar content. Instead, the links tended to promote the author’s profile page, their Google+ profile, and other generic links.
Of course, Authorship is long dead at this point, so there’s no way to test these either way.
I would argue that Google does use dwell time as one exceedingly minor search ranking factor, but with caveats.
- Dwell time has little impact for pages that are not in the first page or two of Google search results. It’s well known that traffic is concentrated in the first search result, and that it drops down immensely by the 10th, and by the second page it has all but disappeared. Dwell time for sites on the second page of search results is essentially meaningless, as there’s so little traffic reaching the pages through that query already.
- Dwell time has diminishing returns. Dwell time measured in seconds is bad. Dwell time measured in a handful of minutes is good. Dwell time that reaches maybe 10 or 15 minutes is fine, but might indicate a few reasons why the user didn’t actually spend that time actively looking at the page. Dwell time beyond 20 minute indicates the user really wasn’t paying attention to the page at all, and thus becomes meaningless.
- Every single technique that goes into increasing dwell time is also beneficial for other reasons. Dwell time is improved by making better content, by pushing for more engagement, and by increasing internal linking. Everything you can do to increase dwell time is beneficial from a content marketing standpoint, and is something Google approves of organically. You don’t need to make up dwell time as an influential metric to say that “having people read your content” is good.
So that’s where I stand. I doubt dwell time specifically is impactful compared to other similar metrics, but I wouldn’t rule out Google’s ability to monitor and use it as a minor element of their results. I just wouldn’t worry about it when optimizing a page. You have more important things you can do instead.