Dwell time/bounce for traffic types & content types

[Nat]

After messing up the Google tracking code a few times, I've finally got some data on actual page dwelling time (sends incremental signals but only while screen is in view mode in browser).

function timer11(){ga('send', 'event', 'TimeOnPage', '1', '11-30 seconds', { 'nonInteraction': 1 });}
function timer31(){ga('send', 'event', 'TimeOnPage', '2', '31-60 seconds', { 'nonInteraction': 1 });}
function timer61(){ga('send', 'event', 'TimeOnPage', '3', '61-180 seconds', { 'nonInteraction': 1 });}
[...]
ga('send', 'event', 'TimeOnPage', '0', '0-10 seconds', { 'nonInteraction': 1 });
setTimeout(timer11,11000);
setTimeout(timer31,31000);
setTimeout(timer61,61000);
[...]

I'm trying to optimize, refine, etc. I can Google "average bounce rates" and stuff, but its all so different and subjective.

I'm measuring dwell time in increments of:

0-10 seconds
11-30 seconds
31-60 seconds
61-180 seconds
181-600 seconds
601-1800 seconds
1801+ seconds​

For viral Reddit traffic on a news/opinion article of about 500 words (no other traffic sources) my bounce rate was like 95% with dwell times of:

0-10 seconds 38.09%
11-30 seconds 29.57%
31-60 seconds 17.45%
61-180 seconds 9.55%
181-600 seconds 3.08%
601-1800 seconds 1.59%
1801+ seconds 0.67%​

For viral Facebook traffic on a very unique large interactive resource page with lots of data (no other traffic sources):

0-10 seconds 25.02%
11-30 seconds 21.22%
31-60 seconds 18.04%
61-180 seconds 14.94%
181-600 seconds 9.81%
601-1800 seconds 6.81%
1801+ seconds 4.16%
​

For organic traffic on non-monetized "fun" list - over 1,500 words (no other traffic sources):

0-10 seconds 24.64%
11-30 seconds 20.29%
31-60 seconds 18.30%
61-180 seconds 15.76%
181-600 seconds 11.78%
601-1800 seconds 6.52%
1801+ seconds 2.72%​

For organic traffic on a detailed buyer's guide ranked #3rd-4th for term (no other traffic sources):

0-10 seconds 25.40%
11-30 seconds 23.01%
31-60 seconds 19.72%
61-180 seconds 16.43%
181-600 seconds 10.06%
601-1800 seconds 4.08%
1801+ seconds 1.29%​

I fully expected Reddit and FB viral traffic to have those dwell numbers. But, the organic traffic in the 0-30 second page dwell range was a surprise to me. Seems really high.

What does it mean if someone bounces in the 10 second range?

• Site didn't load/throws error
• They decide the info they want isn't on the page
• Can't find the info they want fast enough
• They were expecting different content style
• Judgement call that they don't trust the site
• The SERP title doesn't represent the site's content
• Found the content they wanted ridiculously fast (word definitions, "yes, no" answers, etc)

Any advice for what I can do to further increase dwell time and pinpoint problems?

 

[Nat]

Bump... anyone?

 

[Concept]

Tell us a bit more about how you are already trying to keep people engaged on your pages? eg. Site copy, images, videos, polls, etc. Why would people want to stick around on your site?

 

[Nat]

Tell us a bit more about how you are already trying to keep people engaged on your pages? eg. Site copy, images, videos, polls, etc. Why would people want to stick around on your site?

Am I wrong in thinking that a lot of engagement material (images/polls/quizzes/even copy) won't be encountered/experienced until someone's been on a page for more than 1-10 seconds? That's the segment that I'm most concerned about.

Thanks for your response, here's additional info on the pages:

Example 1 (viral Reddit traffic on a news/opinion article) had a featured image that really helped spark negative comment discussion which had a huge impact on boosting visibility & offpage engagement. It was "newsish" and the rest was just good copy + links.

Example 2 (very targeted but viral FB traffic) had two tables. The first was a "ranked" top 10 and right below it was a "full dataset table." Used WP tablepress so it was sortable by different columns, searchable, etc and had data no one knew about. There was an image & short intro before the first table and some "method" info before the 2nd.

Example 3 (organic traffic on non-monetized "fun" list) - Used fun GIFs to break up text and broke the content into specific sections with headers.

Example 4 (organic traffic on a detailed buyer's guide ranked #3rd-4th for term) - This one actually makes some money, not a ton because the item's price range is $25-$50, but I was pretty passionate and interested about the subject matter. The copy is 3,000 words and designed to help people pick between like a million "widget" options. Starts off by essentially saying "In a rush? Here's a quick summary: "

Then the article is pretty well segmented with headers organizing different sections. (Nested headers do go pretty deep, lots h4s, which I think makes it really easy to skim.) Readability should be good, there are some paragraphs with 4-5 sentences, but plenty of short ones. It may be lacking in images, but the copy is better than anything else in the serps.

---
Edit: I also just signed up for a free hotjar account. Hopefully it will also provide some helpful insights.

 

[Ryuzaki]

I'd guess there are a number of users who do what I do in the SERPs, which is to open several of the results into tabs and quickly eliminate the ones that don't match my need or expectation. That could be what you're seeing. I do that instead of pogo-sticking back and forth, which seems wildly inefficient as a user.

 

[SeoReborn]

Also there are some queries that does not need an in-depth read to make a decision.

 

[Nat]

I'd guess there are a number of users who do what I do in the SERPs, which is to open several of the results into tabs and quickly eliminate the ones that don't match my need or expectation. That could be what you're seeing. I do that instead of pogo-sticking back and forth, which seems wildly inefficient as a user.

I do the same in Google, forums, Reddit, etc. Doubt that accounts for a huge %, but definitely some.

Also there are some queries that does not need an in-depth read to make a decision.

You're right, although I don't think many of my organic traffic pages I mention above are getting quick answer search traffic, there's no way for me to read people's minds.

Still waiting on Hotjar to gather data
Edit: I set up hotjar for a homepage heat map accidenly thinking it was full domain. Refined it to the pages I want and started video recording to see behavior. Will update soon

 

[Ollie Brady]

Something I did on which dramatically decreased bounce rate (at least to half of what it was) and significantly increased dwell time was to include a table of contents on the landing page (each item a hash link to a H2, using Table of Contents Plus WP plugin), plus right under it, a colourful, smart infographic I had made.

Conversions well up too, I think to almost double.

I guess it's because it has higher perceived value and is also more visual because of these. Not sure if these methods are applicable to your situation, but maybe.