« Bonhoeffer | The 96 second window of opportunity »

Blog Statistics - Length of Stay

22 October, 2003 10:56 AM

How long does the average blog reader stay on a blog on any given visit?

I searched for the statistic on Google but couldn't find it so I decided to do some of my own investigations.

I headed over to The Truth Laid Bear: Traffic Ranking Page. It lists blogs in order of how much traffic they attract. It is limited to blogs using the Site Meter stats package that have made their statistic public.

I surveyed 350 blogs - 25% of the blogs listed (it took me a few days on my dial up connection) and found the following results.

According to Site Meter stats the average reader spends 96 seconds reading the average blog.

The blogs surveyed came from across the board in terms of their traffic levels. (ie I took results from everything from Instapundit (who reportedly has 80768 visits a day) through to The Trouble with the Baby (who has 1 visit per day).

Other findings

- The top ten blogs on the list had an average of only 37 seconds where as the bottom ten averaged 83 seconds.

- Apart from the 'top ten' there was not a huge difference between blogs receiving high and low traffic. For example - blogs receiving 60 visits per day had an average visit length of 100 seconds which was almost the same as blogs averaging 2000 visits a day (ave 97 seconds).

- Blogs with comments scored a higher average than those without. (this might partly explain the 'top ten' scoring lower as most of them do not have comments) I did not collect data on this, but it became very clear anecdotally.

Implications

96 seconds is not a very long time. It is quite disillusioning to realize that after slaving over a post for hours (or days as this one has taken me) that it is likely to be skimmed over in less than two minutes)

The average blogger would desire to lengthen the stay of their reader. This motivation might be that they are trying to create community and build relationships with their readers. It might be that they have advertising on their site (the longer the stay the more chance of a click through) or it might be that they are wanting to have some sort of a lasting impact on their reader through their writing - the longer the stay the higher chance of this.

Is interaction the key?

My study is by no means conclusive in terms of comments adding to length of stay on blogs - however it does indicate that if bloggers allow for their readers to respond and interact with the writer and each other that they will stay longer. Therefore an interactive approach might be a wise move for bloggers desiring lengthy visits.

Questions and areas for further research

- Does blog design/loading time impact the the length of stay?
- Does blog topic impact the length of stay?
- Do bloggers from certain countries (with high local readership) have different lengths of stay?
- Does posting length have an impact?
- How are News Aggregators impacting length of stay?

This brief survey is limited by the accuracy of Site Meters measuring of length of stay The way they do this is by measuring the difference in time between page views on a site. Accuracy is a problems as some readers will only view the one page on a site - thus registering a time of 0 seconds for their length of stay. Once again this may partly explain how the 'top ten' have low averages as I guess that they would have more readers surfing in throughout the day to check for updates and not surfing through links. As a result of this my 'study' is not something to base life an death blogging decisions on - but is something I'm posting more out of interest than anything else.

I would love to hear your opinions on these results - ideas for further research and ways that you try to lengthen the stay of your readers.

So grab a coffee, put your feet up click on comments below and stay a while!

UPDATE: Rachel from Cre8d Design has just posted a great post tell us Why Visit Length Statistics are Meaningless which explains my above concerns with my 'study' a lot better than I can.

Also in this series of posts How long is 96 seconds? -- Blog Tip 1 - Get to the Point -- Blog Tip 2 - Keep it Simple...Stupid.

Also out my mini study into Gender and Blogging in the 'God-blogosphere.

Comments

Page:

Very interesting. Would love to hear more of your results. Thanks for including me.

Adam » 22 October, 2003 11:38 AM

Darren, you're really on a roll posting insightful ideas. I had never really thought of this, but it's fascinating.

I don't usually visit long (usually visit when Newsgator tells me there's something new), but a good post will get me thinking long after I've left the page. I sometimes check back to see other people's comments, again not staying longer than necessary.

And now I've been here much longer than 96 seconds.

Darryl » 22 October, 2003 12:08 PM

great post - we need more research like this one.

Pamela » 22 October, 2003 1:59 PM

With regards to your questions:

Blog design & loading time definitely impacts my length of stay. For example, before Clarity Amidst Chaos changed his template, he used to have a template that kept on changing colour every few seconds, slowing down the loading time. Although I enjoyed reading his blog, I hardly ever went there because I knew it would be problematic (even on my LAN connection at the office!).

Posting length depends. I admit that when I see a long post, I am tempted to skim through. But if it manages to capture my attention at the beginning and is interesting, I will read it right to the end.

Again, blog topic depends because if it's on a topic I'm not too interested in (say, current affairs), but it's witty and well-written, I'll read it.

I don't really try to lengthen my readers' stay. I try to sometimes write an introduction that will make people want to stop and continue reading, and I try not to make my posts too long (not that it works! :-P) but that's about it.

irene » 22 October, 2003 2:15 PM

My gut feeling is that as long as the load time is reasonable (Dean and other designers tend to quote 8 seconds), loading time doesn't really influence how long people stay. I think what gets people to stay are insightful topics - subjects that make people "stop and think," as the adage goes.

It's also important to remember that it usually takes more time to create creative works than consume them. This site claims that the average reader does 250 words per minute, and yet the world's fastest average typing speed is at least 170 wpm (couldn't find anything more recent). If you're actually creating rather than copying, the speed decreases even more.

I've been here a lot longer than 96 seconds, by the way.

Jesse » 22 October, 2003 2:57 PM

I'm still trying to figure out how the sitemeter knows how long I stay at a site. Sometimes I'll open a site in a new window and leave it open in that window for hours so I can get back to it.

And how does the sitemeter know that I've actually left a site to download another? -- since I don't contact a site in order to tell it I'm leaving. Perhaps someone with more geek sense than I can explain this. Personally, I'm wondering if we're all getting snowed by this stuff and if the king isn't walking around without any clothes.

Brad Boydston » 22 October, 2003 3:20 PM

This makes me feel great. Despite getting pretty good traffic at 2-4k per day, our readers time on site totals have traditionally ranged from 2:15-3:00 (135-180+ seconds).

So I've got to ask: what was the standard deviation in your survey, measured in seconds?

Joe Katzman » 22 October, 2003 4:02 PM

Brad - my understanding is that they work it out by timing the difference between you ariving on the page and when you click to another page within the overall site. So its not completely accurate but a general indication

Darren » 22 October, 2003 5:01 PM

Yeah- in fact all visits, whether they are 10 seconds or ten minutes, will show up as 0 in Sitemeter unless there is a reload or a click to some other internal page, thus skewing the time statistic very heavily to the lower end of the scale.

Homie Bear » 22 October, 2003 5:04 PM

Great stuff, as usual, Darren.

How do you find the time for all of this?! :o)

graham » 22 October, 2003 6:10 PM

I don't have a log that measures length of time. This is terrific Darren.

I think a good heading, topic or lead line will give people pause, and I know the KISS rule applies. Shorter is better.

Load time is an issue for me, but I'll stick it out.

I'm not sure I care if my readers stay....I'm a bit more interested in directing them elsewhere.
They'll come back if they want to.

I really didn't want a comments section when I got my own blog. Now, I couldn't blog without it.
A post is a post, the good stuff is usually found when you venture into the comments. It is where I find a less 'censored' blogger, whose personality is more readable in the dialogue with others.
If I may....it's kind of where you see the fruit of their lives.

I read that the latest tech/blog idea is comment trackback.
Good stuff. Blog on!

Bene Diction » 23 October, 2003 7:19 AM

I have noticed the same thing snooping around in people's sitemeters, the one's with the most hits had the shortest visits. Good questions...

Sarge » 23 October, 2003 7:59 AM

Very, very interesting! An amazing find!

I'm sure the reason for this brevity is varied. But I can't help but think that it is a hallmark of our times. In an age of instant everything, and ever shortening attention spans, it would seem that blog reading is no different. It's the curse of our age, imho.

There is much that could be said on this topic. This is just the tip of the proverbial ice berg. People are just to hurried (and harried). Oh well, that's a topic for a different day.

Peace to you, Darren!

Rich » 23 October, 2003 8:28 AM

Hi Darren, very interesting. I used to be concerned about the seemingly short period of time people spend on blogs, but then I realized you are taking into account people who literally just click through. I doesn't take too many one or two-second visits to brong the average down. And you are right about traffic: the more you have the lower the time spent average. I am happy that we are sticking pretty close to 1 minute at the 6,000 a day average.

Eric Olsen » 23 October, 2003 8:34 AM

I think maybe we need to think about offering refreshments. It could be that people would stay longer if we gave them coffe, and a couch to sit on.

Luke » 23 October, 2003 10:22 AM

sarge is right! you are a machine. thanks for this research.

denise » 23 October, 2003 12:11 PM

Sarge would like to to think that
Darren Rowse is a machine. Wrong all the research is done at the Hamburg end of Melbourne.

Women go to church and male partners are kept occupied.
Blogging: Kindergarden Attention Spam Matters [World.De ]

Jozef » 23 October, 2003 1:12 PM

Scary. I don't think I've written a post on ludicrosity.com yet that can be read in ninety seconds. ;)

John Addis » 23 October, 2003 3:22 PM

I think the way site meters measure visit lengths makes it impossible to draw any conclusions from the data. I mean, when I read Andrew Sullivan or Kausfiles, I start at the stop and keep going until I reach an entry I've read before, then it's off to the next site. As you note, that registers as zero seconds even though I've spent several minutes there.

The only times I use multiple page views are when (A) I've just discovered a blog and I want to look at the archives, (B) I need the permalink so I can blog the post, or (C) if I'm using an RSS feed. And in that last case, 90 seconds sounds about right for a single post -- more than adequate if it's Instapundit sized.

Sean O'Hara » 23 October, 2003 3:35 PM

hey, that's good to know. I figured Sitemeter was probably doing something like that but at first it was pretty disheartening to find that the majority if visitors were only staying 0.00 seconds...!

Gianna » 23 October, 2003 4:00 PM

A lot of the time, though, comment sections often turn into meandering chat rooms; generally staying on topic, but often drifting into OT posts, repetitive OT links, mass agreement of someone's previous comment, or troll stomping.

Matt Kirk » 23 October, 2003 4:21 PM

Another consideration is the number of times people check a blog that hasn't posted anything new. Click to check, nothing new, hit the back button -- total time: about three seconds. That would skew the average down quite a bit. Is there any way to exclude stays of less than five seconds from the average?

mike » 24 October, 2003 1:23 AM

I find these stats interesting and wonder why anyone goes to the trouble to go to the blogs if they don't read them. Of course, I've become a compulsive reader of blogs so maybe my views are skewed. Also take into account that some blogs are not updated on a daily basis so if you click on that, nothing is new you immediately leave, that also skews your stats.

Ruth H » 24 October, 2003 1:56 AM

Great series! Very informative and practical, the type of post that the blogosphere needs more of. Best of all, you followed your own advice.

Joe » 24 October, 2003 1:57 AM

The problem: how are those times calculated? If you look, if someone merely reads your front page, their time is entered as "zero" which implies that what's being measured is the time between clicking posts.

Most readers I know don't click between posts, they merely scroll down the page. Further, even if you do click to a post, Site Meter doesn't record how long you spend reading the latter one...only how much time elapsed between when you opened the page and chose an individual post.

What I don't know is if Site Meter includes the "zero" times in its average. If it does, that's a huge miscalculation. And even if it doesn't, the averages are misleading for the reasons stated above.

Don » 24 October, 2003 3:16 AM

Were the statistics mean or median length visits?

I find on many occasions, I will regularly check for updates on a blog, see there are none and move on. This could account for alot of the zero-second posts.

In addition, I am not so sure how Sitemeter calculates time spent on a blog... on many occasions, someone has been on the site for 10 minutes (they are listed in the "who's on" section) but their visit is still recorded as a zero.

Yes, I am narcissistic and check these things!

BoiFromTroy » 24 October, 2003 3:34 AM

I think the time depends some on page views. I get a ton of visitors who never log more than 30 seconds or less in Sitemeter and they show as 1 page view.

If someone clicks on a link or to the comments, then it triggers the clock to start ticking. So while 96 seconds seems low - I think it's a misleading number that doesn't reflect how much time people are really spending reading blogs. A reader could have your page open all day long but unless they click a link somewhere, they'll log as 30 seconds or less.

My theory (truly unscientific, of course) is that most people who click in to the blog do spend more than 30 seconds reading - they read our thoughtful and thought-provoking posts. It's why I don't worry about the time in my statistics. I look at the number of comments I receive as more of a barometer of my readership.

jen » 24 October, 2003 6:15 AM

i appreciated your post. as irene comments, i got a ton of negative feedback for my chromatically evolving page...readership increased when my template was simplified. i average 2:39 on my views, but the mean isn't an accurate picture...i get stragglers that bring my average WWAAAY up.

i will soon be swapping over to Movable type and look forward to posting more creatively there, that will segment longer posts into shorter ones. we'll see how it pans out.

Bryan » 24 October, 2003 1:14 PM

Good reasearh - good methodology and thanks for sharing it with us.

Trevor Woodward » 28 October, 2003 4:28 PM

Very interesting. Thanks for your hard wor-- oop! 96 seconds is up!

beerzie boy » 31 October, 2003 6:44 AM

Great points....my weblog is simply an article page on sports performance and only has about 1200 unique IPs a week but many read for minutes due to the fact that all the content is specific. If I expanded the audience then I would make them more like a bizstone.com bit and keep the articles in archives.

Carl
www.regenerationlab.com

Carl » 11 November, 2003 5:29 AM

thanks for this research. for the record it took me 5 mins. to read through your excellent blog. (i'm a slow reader.)

it's a litte disheartening because i'm one of those bloggers who don't know when to shut up. checkout my two blog projects and you'll know what i mean:

http://www.coolmel.com
http://www.projecttrinity.net

but anyway, regardless of the ADD syndrome of most readers, i will continue to write as i feel.

thanks again for this well-researched post.

- Rommel (aka coolmel (aka iceman))

coolmel » 13 November, 2003 6:11 AM

One other things that I didn't see was how search engine traffic can affect the length of stay. We've all looked for something, had a page load, and then said, "Heck that's not it!" and you hit the back button and go.

Those visits would get rolled into any average as well.

pops » 13 November, 2003 11:45 AM

It is important to note that sitemeter measures the time of visit by gauging the length between two clicks of the same site.
So, for instance, if a visitor visits a site by one click from somewhere to a site and then shut off the browser, that will be 0 seconds.
Therefore, your statistics might probably off by a huge margin.

__earth » 15 November, 2003 7:27 PM

irst, I can'tbelieve I didn't find this post before now. My apologies.

Something that might also affect visit times is whether a particular blog codes all of their links on the site to open in a new window (which I hate). Unless the user closes the original window the visit time could be lengthened considerably.

Matt » 21 November, 2003 10:09 AM

How can you measure length of stay? HTTP is stateless. You can only make a hunch if the same person visits two pages of the same site, but it will be an upper bound anyways.

JJ » 28 November, 2003 11:16 PM

I want to know ether!
How long does the average blog reader stay on a blog on any given visit?

google » 19 June, 2004 12:32 AM

Yes, 90-120 sec

andry » 2 July, 2004 9:19 AM

Thanks for great info

Refinance » 5 July, 2004 6:51 AM

I've just done a survery of site stastics after the first week (7 days total) of my new blog.... here's what I've found :

http://www.blogweblogs.com

Total Hits : ------------------------
4,352

Duration : ------------------------
over 40% of visitors are staying longer than 2 minutes

Spiders : ------------------------
Inktomi has spidered 152 times
Googlebot has spidered 112 times

I was most impressed by the frequency of spidering! After 5 days I was already ranking at google in the search results for blogging keywords :)

Gblogs » 30 July, 2004 2:23 AM

Thank you for great information!

Ovarian cancer center » 15 September, 2004 3:27 AM

Hi Darren,

I was surfing the web for some ideas of how do people use blog nowadays, and I found your column. It was a very interesting reading, but I realised that it might be a bit out of date now, and I was wondering if you had made any progress on the further research you mentioned or if you could recommend any other site that would increase my knowledge.
Thanks again for the research.
Philippe

Philippe » 11 November, 2004 1:19 PM

I'd heard a little of this before, but some of it was news to me. About how Site Meter measures that length of stay, I've found that the "no links" habit can't be broken, either with friends or by hinting at nude photos to look at or any other bizarre titles for the links. I've decided many new visitors don't know what Links are and many familiar visitors have already had their fun in that direction. I have personally found that I can't kiss enough ass to make any real difference, not with the Free version of Site Meter. Good luck!

Ron » 2 December, 2004 5:35 PM

farg

xaf » 3 December, 2004 4:04 AM

I can not open http://www.livingroom.org.au/blog/archives/gender_blogging.php ?

casino » 25 December, 2004 2:14 AM

Email this entry to a friend:

Friend's email:


Your email:


Message (optional):