tl;dr? September 28, 2010 5:49 AM Subscribe
tl;dr? MeFi readers are apparently not as engaged as Facebook or BoingBoing readers: "I would have expected MetaFilter readers to hang around [and read the article] and Boing Boing users to quickly move along; but actually the opposite is the case."
haven't read the article, don't really plan on it. Am I mentioned by name?
posted by Think_Long at 6:01 AM on September 28, 2010
posted by Think_Long at 6:01 AM on September 28, 2010
We read faster.
posted by smackfu at 6:01 AM on September 28, 2010 [34 favorites]
posted by smackfu at 6:01 AM on September 28, 2010 [34 favorites]
Meanwhile, I'm just masturbating, for pete's sake.
posted by Dumsnill at 6:04 AM on September 28, 2010
posted by Dumsnill at 6:04 AM on September 28, 2010
At least most of our engagements don't end in Klingon-style weddings or divorce notices served via Relationship Status being set to "single."
Oh, and if you're wondering why there are two bars:
Here the top 30 referrers (by volume), with the average time spent on site on the left of each site in a dark shade of that site’s color and the average number of pages visited on the right, in a lighter shade. (via)
posted by griphus at 6:05 AM on September 28, 2010 [2 favorites]
Oh, and if you're wondering why there are two bars:
Here the top 30 referrers (by volume), with the average time spent on site on the left of each site in a dark shade of that site’s color and the average number of pages visited on the right, in a lighter shade. (via)
posted by griphus at 6:05 AM on September 28, 2010 [2 favorites]
BoingBoing readers obviously have no one to talk to.
posted by 2bucksplus at 6:05 AM on September 28, 2010 [2 favorites]
posted by 2bucksplus at 6:05 AM on September 28, 2010 [2 favorites]
Actually, this seems a bit simple, because it's not measuring how much time you spent on the linked page, but how much time you spent exploring the linked site.
posted by smackfu at 6:05 AM on September 28, 2010 [2 favorites]
posted by smackfu at 6:05 AM on September 28, 2010 [2 favorites]
How do they know this?
Look, if you have SiteMeter recording the traffic to your blog, it purports to tell you the average time people read one of your pages. But this isn't accurate. It's only counting the time between when people click on two different pages of your blog. If someone goes to your blog, doesn't click anything else, reads for an hour, then closes the browser window, SiteMeter can't possibly know they read for an hour. It reports this as "zero," which of courses waters down the average figure.
posted by John Cohen at 6:06 AM on September 28, 2010 [5 favorites]
Look, if you have SiteMeter recording the traffic to your blog, it purports to tell you the average time people read one of your pages. But this isn't accurate. It's only counting the time between when people click on two different pages of your blog. If someone goes to your blog, doesn't click anything else, reads for an hour, then closes the browser window, SiteMeter can't possibly know they read for an hour. It reports this as "zero," which of courses waters down the average figure.
posted by John Cohen at 6:06 AM on September 28, 2010 [5 favorites]
It's Google Analytics, JC. He sort of addresses that as well:
A word of caution: While these stats appear here as in Google Analytics, there are likely a whole host of factors that can influence just how much they can actually be trusted, as can be seen my some of the freak stats, like Hacker News (the readers of which are exceptionally attentive!) or meneame.net (click everything in sight and get out quickly!)
posted by griphus at 6:08 AM on September 28, 2010
A word of caution: While these stats appear here as in Google Analytics, there are likely a whole host of factors that can influence just how much they can actually be trusted, as can be seen my some of the freak stats, like Hacker News (the readers of which are exceptionally attentive!) or meneame.net (click everything in sight and get out quickly!)
posted by griphus at 6:08 AM on September 28, 2010
griphus, you haven't answered my question. Saying "it's Google Analytics" is not informative.
posted by John Cohen at 6:09 AM on September 28, 2010
posted by John Cohen at 6:09 AM on September 28, 2010
Here the top 30 referrers (by volume), with the average time spent on site on the left of each site in a dark shade of that site’s color and the average number of pages visited on the right, in a lighter shade.
I would have expected MetaFilter readers tohang around open more pages [and read the article] and Boing Boing users to quickly move along open fewer pages; but actually the opposite is the case." and that is exactly what the data shows.
posted by three blind mice at 6:12 AM on September 28, 2010
I would have expected MetaFilter readers to
posted by three blind mice at 6:12 AM on September 28, 2010
Something wrong with my HTML. Sorry.
My point is that data shows users referred from Metafilter opened more pages than users referred from Boing-boing. This would seem more indicative of reading the article than dwell time.
posted by three blind mice at 6:15 AM on September 28, 2010
My point is that data shows users referred from Metafilter opened more pages than users referred from Boing-boing. This would seem more indicative of reading the article than dwell time.
posted by three blind mice at 6:15 AM on September 28, 2010
Interesting. I imagine if there is a relationship between this and
posted by Busy Old Fool at 6:15 AM on September 28, 2010 [3 favorites]
- the degree of interest in the site audience in the article topic (I took a quick look, decided my limited interest in Star Wars did not justify the article length and moved on)
- the proportion of linked words on the front page of each site represented by the article (Metafilter currently has over 60 posts on the front page, many of which link to long written articles. Reddit has 25 front page links, 12 of which link to text-based pages. I'd say it's possible to read all the text linked to by the reddit front page in a single sitting in a way that one simply can't do from Metafilter and that might explain the greater amount of skimming that goes on.
posted by Busy Old Fool at 6:15 AM on September 28, 2010 [3 favorites]
More seriously, I was one of those who followed that link from MetaFilter and read the whole thing. But after having read it, it felt like a waste of time. So maybe smarter people than me realized that it would be a waste of time before they read the whole post?
posted by Dumsnill at 6:19 AM on September 28, 2010 [5 favorites]
posted by Dumsnill at 6:19 AM on September 28, 2010 [5 favorites]
I took a quick look, decided my limited interest in Star Wars did not justify the article length and moved on
Me too. I saw the link, thought it sounded interesting, clicked it, read a bit, and realized that no, I did not want to read three miles of text about Star Wars.
So, you know, sorry for bringing you guys down.
posted by uncleozzy at 6:21 AM on September 28, 2010 [2 favorites]
Me too. I saw the link, thought it sounded interesting, clicked it, read a bit, and realized that no, I did not want to read three miles of text about Star Wars.
So, you know, sorry for bringing you guys down.
posted by uncleozzy at 6:21 AM on September 28, 2010 [2 favorites]
I think it's probably that mefites have a lower tolerance for star wars nerdery.
posted by empath at 6:22 AM on September 28, 2010 [3 favorites]
posted by empath at 6:22 AM on September 28, 2010 [3 favorites]
I think it's probably that mefites have a lower tolerance for star wars nerdery.
Ahahahahahahaa, I doubt it.
posted by Think_Long at 6:25 AM on September 28, 2010 [5 favorites]
Ahahahahahahaa, I doubt it.
posted by Think_Long at 6:25 AM on September 28, 2010 [5 favorites]
Well, bless their hearts.
posted by sanko at 6:28 AM on September 28, 2010 [4 favorites]
posted by sanko at 6:28 AM on September 28, 2010 [4 favorites]
Saying "it's Google Analytics" is not informative.
Well, no, all I was saying is that it is not SiteMeter, it's GA and that he's aware of any number of potential factors influencing the appearance of his stats.
posted by griphus at 6:30 AM on September 28, 2010
Well, no, all I was saying is that it is not SiteMeter, it's GA and that he's aware of any number of potential factors influencing the appearance of his stats.
posted by griphus at 6:30 AM on September 28, 2010
I think it's probably that mefites have a lower tolerance for star wars nerdery.
Shit just got heavy for a large portion of your readership, sir. That was like sirens and klaxons and flashing red lights for anyone itching for another Nerd Thunderdome.
posted by carsonb at 6:33 AM on September 28, 2010
Shit just got heavy for a large portion of your readership, sir. That was like sirens and klaxons and flashing red lights for anyone itching for another Nerd Thunderdome.
posted by carsonb at 6:33 AM on September 28, 2010
This confirms my biases. My general impression, gleaned after reading Mefi threads with links to long articles, is that a lot of people tend to skim the article or read what the post author highlighted from the article and make their comments based on that. Then various people who have read the entire comment about the mistaken conclusions of the first group. It doesn't seem a really huge thing, but a noticeable one, at least for me.
Also, based on a quick skim of the article, these stats are based on single post over a single week. Maybe Mefites just aren't that interested in an overgrown dog that can shoot a laser gun?
posted by nomadicink at 6:38 AM on September 28, 2010 [6 favorites]
Also, based on a quick skim of the article, these stats are based on single post over a single week. Maybe Mefites just aren't that interested in an overgrown dog that can shoot a laser gun?
posted by nomadicink at 6:38 AM on September 28, 2010 [6 favorites]
Or, y'know, maybe MeFi readers simply don't geek-out quite as much as BB or HackerNews readers over ancient Star Wars crap?
posted by Thorzdad at 6:38 AM on September 28, 2010 [1 favorite]
posted by Thorzdad at 6:38 AM on September 28, 2010 [1 favorite]
I should say that I read the entire article with interest. But it's still star wars nerdery.
posted by empath at 6:38 AM on September 28, 2010
posted by empath at 6:38 AM on September 28, 2010
Using JS you can definitely track how long the page is open: you just set up a timer that hits a given counter URL every 30 seconds or so. When the page is closed, the heartbeats stop. I looked at the page and didn't see any of that going on, but it's possible that he removed it; or if it's part of GA then I have all that crap blocked as well.
posted by Rhomboid at 6:50 AM on September 28, 2010 [1 favorite]
posted by Rhomboid at 6:50 AM on September 28, 2010 [1 favorite]
I've often wondered how Google measures that time on page stuff. Because when I read MeFi, which I do mainly via RSS, I click on a series of links that interest me, and open them in the new tab. Sometimes I also open the MeFi post that goes with them if I think the comments are likely to be interesting, sometimes I don't bother. It can be an hour before I get around to closing the last of the tabs I opened. Does that page end up looking like my time in page was an hour? Even if I actually got to it, thought 'Oh, this isn't nearly as interesting as I thought?' and closed it within seconds?
posted by jacquilynne at 6:56 AM on September 28, 2010 [1 favorite]
posted by jacquilynne at 6:56 AM on September 28, 2010 [1 favorite]
Are there as many people who follow Boing Boing purely for the discussion threads? I'm really not familiar with that community. But here on MeFi, yes, there are lots of people who clearly do what nomadicink is talking about — I'm sometimes one of them — and I get the sense that it's often because we treat the links as prompts for discussion.
If I want news, I read new sites; if I want blog posts, I use the suggestion feature in Google Reader; if I'm here, it's because I'm looking for commentary. What I'm wondering is whether there are more or fewer people who treat Boing Boing in the same way.
posted by nebulawindphone at 6:56 AM on September 28, 2010
If I want news, I read new sites; if I want blog posts, I use the suggestion feature in Google Reader; if I'm here, it's because I'm looking for commentary. What I'm wondering is whether there are more or fewer people who treat Boing Boing in the same way.
posted by nebulawindphone at 6:56 AM on September 28, 2010
I've often wondered how Google measures that time on page stuff. Because when I read MeFi, which I do mainly via RSS, I click on a series of links that interest me, and open them in the new tab.
As someone who spends a lot of his time staring at GA analysis screens, the answer is poorly. Especially with power users with multiple tabs.
posted by generichuman at 6:59 AM on September 28, 2010 [1 favorite]
As someone who spends a lot of his time staring at GA analysis screens, the answer is poorly. Especially with power users with multiple tabs.
posted by generichuman at 6:59 AM on September 28, 2010 [1 favorite]
All of this seems to presuppose that his article was worth reading, which is maybe a bit much.
I took a quick look, decided my limited interest in Star Wars did not justify the article length and moved on
Also this.
posted by shakespeherian at 7:01 AM on September 28, 2010
I took a quick look, decided my limited interest in Star Wars did not justify the article length and moved on
Also this.
posted by shakespeherian at 7:01 AM on September 28, 2010
It's a long article about a fucking Wookie, for fuck's sake. He should have just mailed it to Larry Groznic.
posted by pracowity at 7:03 AM on September 28, 2010 [1 favorite]
posted by pracowity at 7:03 AM on September 28, 2010 [1 favorite]
The way I look at Metafilter and the way I look at Reddit are completely different. When I look at Metafilter I am usually looking to find something that is interesting to me to spend some time thinking about. I might click on only a few links and I might back out of some links very quickly. Usually when I look at Reddit, I open every link that might interest me in a separate tab. I usually go through a pages worth of links opening tabs before I even look at one tab that I opened. Then I usually look at the ones that are just pictures first, because sometimes they don't make any sense without the Reddit headline. A lot of times I will then click through to the next page of Reddit results and do the same thing. I might open three or four pages of results before I even read part of an article. If I get distracted by something else, I might have that article open for hours before I even come back to it.
I am only one person, but the way that I use the two sites makes it really difficult to compare my level of interaction with content solely based on the amount of time that I spend with a linked website open. I'm willing to bet that a lot of Metafilter users are not opening everything in a separate tab, and a lot of Reddit users are.
posted by jefeweiss at 7:03 AM on September 28, 2010 [3 favorites]
I am only one person, but the way that I use the two sites makes it really difficult to compare my level of interaction with content solely based on the amount of time that I spend with a linked website open. I'm willing to bet that a lot of Metafilter users are not opening everything in a separate tab, and a lot of Reddit users are.
posted by jefeweiss at 7:03 AM on September 28, 2010 [3 favorites]
It might have something to do with posting style on different sites.
MetaFilter posts are short and sometimes enigmatic. MetaFilter readers frequently click on one of the links in a post and then realize they're not interested when they find out what it's actually about, but stay a reasonably long time if they decide they are interested. (Basically, what uncleozzy said.)
BoingBoing posts are short articles on a subject with link. BoingBoing readers would frequently just read the BoingBoing article, but if they're interested enough in the subject of the post to follow the links provided, they usually stay and read them.
Hacker News posts are just headlines (but descriptive, not teasers), the articles linked to are frequently technical.
posted by nangar at 7:04 AM on September 28, 2010 [3 favorites]
MetaFilter posts are short and sometimes enigmatic. MetaFilter readers frequently click on one of the links in a post and then realize they're not interested when they find out what it's actually about, but stay a reasonably long time if they decide they are interested. (Basically, what uncleozzy said.)
BoingBoing posts are short articles on a subject with link. BoingBoing readers would frequently just read the BoingBoing article, but if they're interested enough in the subject of the post to follow the links provided, they usually stay and read them.
Hacker News posts are just headlines (but descriptive, not teasers), the articles linked to are frequently technical.
posted by nangar at 7:04 AM on September 28, 2010 [3 favorites]
This confirms my biases. My general impression, gleaned after reading Mefi threads with links to long articles, is that a lot of people tend to skim the article or read what the post author highlighted from the article and make their comments based on that.
Posters are usually in the minority, most people simply read the FPP and don't post at all, so it's hard to know.
But I agree with your bias, most commenters have not read the article, or read it with passing comprehension. The MeFi site design encourages this actually, the earlier in a thread a comment is made, the more people will read it, the more attention it gets further down thread. Though this problem is not unique to MeFi, since there is no fancy thread nesting or filtering, the best way to get attention is to post often and early which sort of encourages tl;dr behavior. It dumbs down the conversation, but there are also upsides so it's a trade-off.
posted by stbalbach at 7:05 AM on September 28, 2010 [1 favorite]
Posters are usually in the minority, most people simply read the FPP and don't post at all, so it's hard to know.
But I agree with your bias, most commenters have not read the article, or read it with passing comprehension. The MeFi site design encourages this actually, the earlier in a thread a comment is made, the more people will read it, the more attention it gets further down thread. Though this problem is not unique to MeFi, since there is no fancy thread nesting or filtering, the best way to get attention is to post often and early which sort of encourages tl;dr behavior. It dumbs down the conversation, but there are also upsides so it's a trade-off.
posted by stbalbach at 7:05 AM on September 28, 2010 [1 favorite]
Well, no, all I was saying is that it is not SiteMeter, it's GA and that he's aware of any number of potential factors influencing the appearance of his stats.
I wasn't saying it's SiteMeter. I was giving SiteMeter as an example because that's what I know about. Why does it matter? Does Google Analytics have a way of knowing when you've closed your browser window?
He might be aware of those factors. Good for him. But he's still presenting the data to us as if it's meaningful. And I'm saying I don't trust the data.
posted by John Cohen at 7:05 AM on September 28, 2010 [1 favorite]
I wasn't saying it's SiteMeter. I was giving SiteMeter as an example because that's what I know about. Why does it matter? Does Google Analytics have a way of knowing when you've closed your browser window?
He might be aware of those factors. Good for him. But he's still presenting the data to us as if it's meaningful. And I'm saying I don't trust the data.
posted by John Cohen at 7:05 AM on September 28, 2010 [1 favorite]
I think it's probably that mefites have a lower tolerance for star wars nerdery.
This is exactly the case for me. I saw the link and being a Star Wars hater assumed it was going to go to some side by side comparison of photos or some such shit that would confirm my belief that George Lucas has sucked forever, not just recently. I clicked, saw a lengthy text-piece, and realized I didn't give a shit what it was and that my hatred of Star Wars needs no more fuel anyway. I left without reading past the first sentence--and even that far a read was unintentional.
posted by dobbs at 7:06 AM on September 28, 2010
This is exactly the case for me. I saw the link and being a Star Wars hater assumed it was going to go to some side by side comparison of photos or some such shit that would confirm my belief that George Lucas has sucked forever, not just recently. I clicked, saw a lengthy text-piece, and realized I didn't give a shit what it was and that my hatred of Star Wars needs no more fuel anyway. I left without reading past the first sentence--and even that far a read was unintentional.
posted by dobbs at 7:06 AM on September 28, 2010
What dobbs said.
posted by JohnnyGunn at 7:09 AM on September 28, 2010
posted by JohnnyGunn at 7:09 AM on September 28, 2010
I read the whole damn thing, but on instapaper and very fast.
posted by dhruva at 7:13 AM on September 28, 2010
posted by dhruva at 7:13 AM on September 28, 2010
Also, based on a quick skim of the article, these stats are based on single post over a single week.
This is obviously the biggest problem with the data. It's an interesting idea but needs to be done with a decent sample to crawl out of sheer Who Knows territory. Cover a variety of topics and account for differing presentational approaches from site to site as well as from post to post on each site.
Compare as well the baseline outgoing traffic for a post on each site with the actual measured traffic for a particular post:
- if a reddit post sends (here is a made-up number) 10,000 viewers to a page from a typical popular-but-not-huge post, your "how engaged are they" analysis should take into account whether you're dealing with a typical random smattering of 10K reddit readers, in which case measuring engagement against baseline normal clickery seems reasonable;
- if the post generates 30K viewers, there's something going on there that's putting a lot more eyes than normal on the linked site (maybe the reddit post is an unusually sensational pitch or has a tantalizing hed?) and so a low engagement ratio may not be anything but an indication that when a lot more people than normal click through, the bulk of the extra traffic is made up of extra driveby readers rather than your core readership.
- similarly, if the post generates only 3K viewers, there may be self-selection in play where people are more effectively than average judging their likely interest in the material before clicking through (maybe the reddit post is very clear in its presentation of what will be on the other end of the link) and so a high level of engagement is more a measure of that accurate pre-click assessment in action than it is a measure of the larger body of core reddit readers.
It'd also be interesting to take a careful look at the presentation of the story on each site; there's obviously a lot of variety from site to site, as folks have noted, in just general presentational style, but there may also be more or less significant amounts of variation in how any given post appears on each site. Mefi certainly has both a general style to posts (mostly one-paragraph, no editorial framing or pitch) and significant variation (summary vs. pullquote vs. blind link, paragraph vs. sentence fragment length), so picking one post out of the bin doesn't really give a picture of how Metafilter presents material as it does how a given single poster presented it.
posted by cortex (staff) at 7:32 AM on September 28, 2010 [2 favorites]
This is obviously the biggest problem with the data. It's an interesting idea but needs to be done with a decent sample to crawl out of sheer Who Knows territory. Cover a variety of topics and account for differing presentational approaches from site to site as well as from post to post on each site.
Compare as well the baseline outgoing traffic for a post on each site with the actual measured traffic for a particular post:
- if a reddit post sends (here is a made-up number) 10,000 viewers to a page from a typical popular-but-not-huge post, your "how engaged are they" analysis should take into account whether you're dealing with a typical random smattering of 10K reddit readers, in which case measuring engagement against baseline normal clickery seems reasonable;
- if the post generates 30K viewers, there's something going on there that's putting a lot more eyes than normal on the linked site (maybe the reddit post is an unusually sensational pitch or has a tantalizing hed?) and so a low engagement ratio may not be anything but an indication that when a lot more people than normal click through, the bulk of the extra traffic is made up of extra driveby readers rather than your core readership.
- similarly, if the post generates only 3K viewers, there may be self-selection in play where people are more effectively than average judging their likely interest in the material before clicking through (maybe the reddit post is very clear in its presentation of what will be on the other end of the link) and so a high level of engagement is more a measure of that accurate pre-click assessment in action than it is a measure of the larger body of core reddit readers.
It'd also be interesting to take a careful look at the presentation of the story on each site; there's obviously a lot of variety from site to site, as folks have noted, in just general presentational style, but there may also be more or less significant amounts of variation in how any given post appears on each site. Mefi certainly has both a general style to posts (mostly one-paragraph, no editorial framing or pitch) and significant variation (summary vs. pullquote vs. blind link, paragraph vs. sentence fragment length), so picking one post out of the bin doesn't really give a picture of how Metafilter presents material as it does how a given single poster presented it.
posted by cortex (staff) at 7:32 AM on September 28, 2010 [2 favorites]
I did some skimming for sure.
posted by Trochanter at 7:39 AM on September 28, 2010
posted by Trochanter at 7:39 AM on September 28, 2010
Old British Music Hall Joke:
Dear Margaret,
I am writing this very slow, as I know you can't read very fast...
posted by Trochanter at 7:41 AM on September 28, 2010
Dear Margaret,
I am writing this very slow, as I know you can't read very fast...
posted by Trochanter at 7:41 AM on September 28, 2010
Metafilter user reads "George Lucas Stole Chewbacca, But It’s Okay", thinks "This Is Not Worth My Attention", gets back to running the Internet.
posted by boo_radley at 7:41 AM on September 28, 2010
posted by boo_radley at 7:41 AM on September 28, 2010
Using JS you can definitely track how long the page is open: you just set up a timer that hits a given counter URL every 30 seconds or so. When the page is closed, the heartbeats stop.
What happens when I click a link in an fpp, it opens in a new tab, and then I read the text on that page but don't close the tab for two days? How is something like that measured? I assume I'm not the only person who has multiple tabs open, some of whose pages I've read and some I have not, and will not.
posted by rtha at 7:49 AM on September 28, 2010
What happens when I click a link in an fpp, it opens in a new tab, and then I read the text on that page but don't close the tab for two days? How is something like that measured? I assume I'm not the only person who has multiple tabs open, some of whose pages I've read and some I have not, and will not.
posted by rtha at 7:49 AM on September 28, 2010
The article was posted on BoingBoing and Hacker News on the 22nd. Posted to Metafilter on the 23rd. It could be that people read it before it was posted here, which might explain some difference in the graph.
posted by infinitefloatingbrains at 7:49 AM on September 28, 2010
posted by infinitefloatingbrains at 7:49 AM on September 28, 2010
It could also be the result of BoingBoing and Hacker News posting one sad link, compared to MeFi's FOURTEEN links. A lot more interesting stuff than just the article.
Conclusion: They got Filthy Light Thieved.
posted by infinitefloatingbrains at 7:53 AM on September 28, 2010 [2 favorites]
Conclusion: They got Filthy Light Thieved.
posted by infinitefloatingbrains at 7:53 AM on September 28, 2010 [2 favorites]
MeFi users can't even
Out of a crowd of thousands of logged-in users dropping by the site on a daily basis, at least a few of them can't be arsed to read before commenting. And a larger collection don't mind having a conversation in the thread regarding in part whatever that small cadre of early birds threw in. It makes no sense to mistake the existence within a large heterogeneous community of one common kind of user for an accurate snapshot of that large community.
posted by cortex (staff) at 7:56 AM on September 28, 2010
Out of a crowd of thousands of logged-in users dropping by the site on a daily basis, at least a few of them can't be arsed to read before commenting. And a larger collection don't mind having a conversation in the thread regarding in part whatever that small cadre of early birds threw in. It makes no sense to mistake the existence within a large heterogeneous community of one common kind of user for an accurate snapshot of that large community.
posted by cortex (staff) at 7:56 AM on September 28, 2010
Who wants to read 3 miles of Star Wars when instead you could spend your time on Minecraft?
posted by allkindsoftime at 7:59 AM on September 28, 2010 [2 favorites]
posted by allkindsoftime at 7:59 AM on September 28, 2010 [2 favorites]
Maybe it's indicative of something else entirely that Hacker News readers would devote a lot of time to reading about Chewbacca, and Metafilter users would not.
I want to see how much time Hacker News folks would put into reading about Belgian domestic politics.
posted by Slap*Happy at 8:09 AM on September 28, 2010 [2 favorites]
I want to see how much time Hacker News folks would put into reading about Belgian domestic politics.
posted by Slap*Happy at 8:09 AM on September 28, 2010 [2 favorites]
i didnt read this thread but the data is probably because everyone on Metafilter already knows everything there is to know about Star Trek. QED.
posted by Potomac Avenue at 8:14 AM on September 28, 2010
posted by Potomac Avenue at 8:14 AM on September 28, 2010
In the link itself, there are multiple links in the first couple paragraphs that lead off of that site. Just going and reading it now, my own behavior would look less engaged (went, then left), except that then I hit the back button and went back to the article.
posted by klangklangston at 8:22 AM on September 28, 2010
posted by klangklangston at 8:22 AM on September 28, 2010
I was going to post about this earlier this morning, but I wasn't inclined to facilitate a double bump for the guy's hit counter.
posted by crunchland at 8:25 AM on September 28, 2010
posted by crunchland at 8:25 AM on September 28, 2010
And, for what it's worth, I have a javascript blocker on my browser, and I routinely block google analytics. Even if I read every single word on this guy's site, he'd never know it.
posted by crunchland at 8:31 AM on September 28, 2010 [1 favorite]
posted by crunchland at 8:31 AM on September 28, 2010 [1 favorite]
Ooh, is this the part where they rely on a onPageUnload() Javascript event to reliably tell them about page visit length? The same event that seems to work about 80% of the time on the (much higher volume, less technical user-base) site that I work on full-time? The same one that fails 100% of the time for users with NoScript (which is to say, technical users like us lugs on MeFi who are all web developers or some shit) running? The one that I just can't get our marketing group to let go of as a primary reporting metric?
*busily greps source of binary bonsai*
tl;dr: light shade on graph is probably right, dark shade is probably full of shit
posted by Mayor West at 8:33 AM on September 28, 2010 [2 favorites]
*busily greps source of binary bonsai*
beforeunload:{setup:function(a,b,d){if(this.setInterval)this.onbeforeunload=d;return false}
Hey, look at that!tl;dr: light shade on graph is probably right, dark shade is probably full of shit
posted by Mayor West at 8:33 AM on September 28, 2010 [2 favorites]
All these generalized conclusions based on a single referred article? Why are we wasting our time on this?
posted by rocket88 at 8:37 AM on September 28, 2010
posted by rocket88 at 8:37 AM on September 28, 2010
The alternative is doing actual work. Clearly the way to make up for a thin source is to overthink the hell out of it.
posted by cortex (staff) at 8:39 AM on September 28, 2010
posted by cortex (staff) at 8:39 AM on September 28, 2010
I read both sites. Often BoingBoing gets the links first and I read it from their link, then when it pops up again on MeFi I click quickly through to verify that it's the same thing and then bugger off.
It's easy to draw simple conclusions, much harder to look at all the factors in a pattern. Even factoring in posts that are on both sites and time of post on each site could have shown my pattern.
posted by L'Estrange Fruit at 8:42 AM on September 28, 2010
It's easy to draw simple conclusions, much harder to look at all the factors in a pattern. Even factoring in posts that are on both sites and time of post on each site could have shown my pattern.
posted by L'Estrange Fruit at 8:42 AM on September 28, 2010
Yeah, I read it on Instapaper and on a plane.
posted by These Premises Are Alarmed at 8:48 AM on September 28, 2010
posted by These Premises Are Alarmed at 8:48 AM on September 28, 2010
This is obviously the biggest problem with the data. It's an interesting idea but needs to be done with a decent sample to crawl out of sheer Who Knows territory. Cover a variety of topics and account for differing presentational approaches from site to site as well as from post to post on each site.
Yes, but how would you categorize the posts, cortex, HOW?!
posted by nomadicink at 8:57 AM on September 28, 2010
Yes, but how would you categorize the posts, cortex, HOW?!
posted by nomadicink at 8:57 AM on September 28, 2010
Rhomboid: "Using JS you can definitely track how long the page is open: you just set up a timer that hits a given counter URL every 30 seconds or so. When the page is closed, the heartbeats stop. I looked at the page and didn't see any of that going on, but it's possible that he removed it; or if it's part of GA then I have all that crap blocked as well"
My first though on seeing the data is Metafilter readers just compose a greater installed base of script blockers.
generichuman writes "As someone who spends a lot of his time staring at GA analysis screens, the answer is poorly. Especially with power users with multiple tabs."
I wonder how google handles me having 200 tabs open (that's an extreme but it's rarely less than 50) if I let analytics past NoScript. As far as Google is concerned I could get the equivalent of an entire week of 24X7 web activity in like 45 minutes.
cortex writes "It'd also be interesting to take a careful look at the presentation of the story on each site; there's obviously a lot of variety from site to site, as folks have noted, in just general presentational style, but there may also be more or less significant amounts of variation in how any given post appears on each site. Mefi certainly has both a general style to posts (mostly one-paragraph, no editorial framing or pitch) and significant variation (summary vs. pullquote vs. blind link, paragraph vs. sentence fragment length), so picking one post out of the bin doesn't really give a picture of how Metafilter presents material as it does how a given single poster presented it."
Anyone seriously doing this study would have to compensate for mystery meat posts.
posted by Mitheral at 9:07 AM on September 28, 2010
My first though on seeing the data is Metafilter readers just compose a greater installed base of script blockers.
generichuman writes "As someone who spends a lot of his time staring at GA analysis screens, the answer is poorly. Especially with power users with multiple tabs."
I wonder how google handles me having 200 tabs open (that's an extreme but it's rarely less than 50) if I let analytics past NoScript. As far as Google is concerned I could get the equivalent of an entire week of 24X7 web activity in like 45 minutes.
cortex writes "It'd also be interesting to take a careful look at the presentation of the story on each site; there's obviously a lot of variety from site to site, as folks have noted, in just general presentational style, but there may also be more or less significant amounts of variation in how any given post appears on each site. Mefi certainly has both a general style to posts (mostly one-paragraph, no editorial framing or pitch) and significant variation (summary vs. pullquote vs. blind link, paragraph vs. sentence fragment length), so picking one post out of the bin doesn't really give a picture of how Metafilter presents material as it does how a given single poster presented it."
Anyone seriously doing this study would have to compensate for mystery meat posts.
posted by Mitheral at 9:07 AM on September 28, 2010
Is this a chart posing as ... research?
posted by jessamyn (staff) at 9:14 AM on September 28, 2010 [3 favorites]
posted by jessamyn (staff) at 9:14 AM on September 28, 2010 [3 favorites]
It's actually a link to a blog post discussing a different blog post, which has more charts. Boingboing cut the chart off too, which seems particularly stupid.
posted by smackfu at 9:17 AM on September 28, 2010
posted by smackfu at 9:17 AM on September 28, 2010
I think it's probably that mefites have a lower tolerance for star wars nerdery.
Or maybe our lips don't move when we read. That can trip you up a bit, slow you down.
posted by philip-random at 9:21 AM on September 28, 2010 [1 favorite]
Or maybe our lips don't move when we read. That can trip you up a bit, slow you down.
posted by philip-random at 9:21 AM on September 28, 2010 [1 favorite]
Compare what people would have read on Boing Boing about the Chewbacca article versus what they would have read on Metafilter before deciding whether to click the respective links.
The BB article would have given them an excellent idea of what the whole thing was about, and a detailed indication of what they could expect to find if they clicked the link. Anyone that had only a casual interest in the subject would have seen all they wanted to know on BB itself, and anyone that then clicked the link would likely have had rather more than a passing interest in the material.
The MeFi post on the other hand gave me the barest indication of what kind of thing I might find at the end of the link, and so I went through the same sequence as a lot of other people... "What's this? Let's check it out... oh... too much information... I'm really not *that* deep into this stuff... close tab".
I'm a newbie here, but that doesn't seem to be untypical of how it goes here. It's not necessarily a matter of "too long, didn't read" but more "not quite what I had imagined it would be, so skipped".
posted by philipy at 9:31 AM on September 28, 2010 [1 favorite]
The BB article would have given them an excellent idea of what the whole thing was about, and a detailed indication of what they could expect to find if they clicked the link. Anyone that had only a casual interest in the subject would have seen all they wanted to know on BB itself, and anyone that then clicked the link would likely have had rather more than a passing interest in the material.
The MeFi post on the other hand gave me the barest indication of what kind of thing I might find at the end of the link, and so I went through the same sequence as a lot of other people... "What's this? Let's check it out... oh... too much information... I'm really not *that* deep into this stuff... close tab".
I'm a newbie here, but that doesn't seem to be untypical of how it goes here. It's not necessarily a matter of "too long, didn't read" but more "not quite what I had imagined it would be, so skipped".
posted by philipy at 9:31 AM on September 28, 2010 [1 favorite]
What I thought this was going to be before I read it: a numerical analysis of how many times people in threads on all of those sites said tl;dr indicating, at some level, their level of detachment from the subject and need to proclaim same in thread. That's what I'd like to know about. Average lenght of thread. Average length of linked content [in words or minutes]. Average number of people saying tl;dr based on and/or correlated to the length of thread/content.
posted by jessamyn (staff) at 9:34 AM on September 28, 2010
posted by jessamyn (staff) at 9:34 AM on September 28, 2010
Too bad we don't know anyone who is willing to compile information at a moment's notice like that like some kind of a trick monkey.
posted by crunchland at 9:39 AM on September 28, 2010
posted by crunchland at 9:39 AM on September 28, 2010
Shiite article with an intriguing headline - I thought there was going to be a smoking gun in there about Lucas actually stealing the idea for Chewie. But when it turned out that, shocker, Chewbacca was probably the result of a bunch of different people's different ideas acting upon one another, and that he did not just spring fully formed from Lucas' (or anyone's) mind, well, rather than read the pporly written article I just went back to watching Gimme Pizza, over and over again.
posted by dirtdirt at 9:50 AM on September 28, 2010 [1 favorite]
posted by dirtdirt at 9:50 AM on September 28, 2010 [1 favorite]
like some kind of a trick monkey
I think around here we call him the CIO.
posted by jessamyn (staff) at 9:52 AM on September 28, 2010
I think around here we call him the CIO.
posted by jessamyn (staff) at 9:52 AM on September 28, 2010
Hail Mefites!
The best explanation for BB's epic engagement rating here is that we could run a picture of Chewbacca with six tits.
posted by beschizza at 10:11 AM on September 28, 2010
The best explanation for BB's epic engagement rating here is that we could run a picture of Chewbacca with six tits.
posted by beschizza at 10:11 AM on September 28, 2010
Has anyone considered that boing boing readers might just read more slowly than the kids not, you know, getting their links from boing boing?
posted by DarlingBri at 10:58 AM on September 28, 2010 [1 favorite]
posted by DarlingBri at 10:58 AM on September 28, 2010 [1 favorite]
Interesting question, but terrible methodology/execution. Skewed results; meaningless interpretation. This type of pseudoscientific bull shit is not to be encouraged.
posted by polymodus at 11:07 AM on September 28, 2010
posted by polymodus at 11:07 AM on September 28, 2010
Also: how much of this was time spent reading the article versus time spent masturbating to pictures of Chewbacca?
posted by shakespeherian at 11:10 AM on September 28, 2010
posted by shakespeherian at 11:10 AM on September 28, 2010
I wonder if it might be a matter of demographics, too. Maybe boingboing skews a little older, so that the readership there might be more inclined to be interested in Star Wars stuff. Hacker News is apparently loaded with the perfect group of people specifically interested in the origins of a fictional dog-man.
Or maybe we should just face facts: Boingboing is better than us when it comes to loving dog-men.
posted by crunchland at 11:12 AM on September 28, 2010
Or maybe we should just face facts: Boingboing is better than us when it comes to loving dog-men.
posted by crunchland at 11:12 AM on September 28, 2010
MetaFilter: The alternative is doing actual work.
posted by nickmark at 11:48 AM on September 28, 2010
posted by nickmark at 11:48 AM on September 28, 2010
I think cortex and philipy explained this pretty well. The BoingBoing post is like a magazine article, designed to funnel you towards that single link - while the Metafilter post is a small essay with links that fill in background details and give you context. The BoingBoing people decided if they wanted to read that one link and clicked it or moved on, while the Metafilter people lingered over the different links, spreading their attention more widely.
posted by Kevin Street at 12:12 PM on September 28, 2010 [1 favorite]
posted by Kevin Street at 12:12 PM on September 28, 2010 [1 favorite]
Boingboing is better than us when it comes to loving dog-men.
Yeah, until the inevitable break-up. Then it's all tears and *cc*s*ti*ns.
posted by octobersurprise at 12:36 PM on September 28, 2010
Yeah, until the inevitable break-up. Then it's all tears and *cc*s*ti*ns.
posted by octobersurprise at 12:36 PM on September 28, 2010
The BB/DF/FB(?!) differential is interesting, but at fewer pages and way more time, I'm a little concerned about Hacker News. Did somebody drop it on its head as a child?
i didnt read this thread but the data is probably because everyone on Metafilter already knows everything there is to know about Star Trek. QED.
I see what you did there.
posted by Durn Bronzefist at 12:38 PM on September 28, 2010
i didnt read this thread but the data is probably because everyone on Metafilter already knows everything there is to know about Star Trek. QED.
I see what you did there.
posted by Durn Bronzefist at 12:38 PM on September 28, 2010
It's "Wookiee" dammit. Two Es.
posted by deborah at 3:00 PM on September 28, 2010 [1 favorite]
posted by deborah at 3:00 PM on September 28, 2010 [1 favorite]
I thought the timeline with the overall traffic spike l-r was a much more interesting graph for a single article. You can see where it got picked up initially, where it gained steam, and who the tag-alongs were. How articles go viral is an interesting thing, and that was a cool snapshot. The "average time" viewing thing doesn't reveal much about the "average" site user, really, since there's nothing to average against, like a larger sample. Let's see that with a year's worth of data, then it'll be worth a glance.
posted by Devils Rancher at 8:27 PM on September 28, 2010
posted by Devils Rancher at 8:27 PM on September 28, 2010
Yeah, I bet MetaFilter members read faster.
posted by limeonaire at 5:27 AM on September 29, 2010
posted by limeonaire at 5:27 AM on September 29, 2010
We're all just so erudite that we mostly can't be bothered with that piffle.
posted by Devils Rancher at 7:34 AM on September 29, 2010
posted by Devils Rancher at 7:34 AM on September 29, 2010
"We read faster."
"We're all just so erudite that we mostly can't be bothered with that piffle."
My thoughts exactly, we're all really smart!
posted by Blasdelb at 8:37 AM on September 29, 2010
"We're all just so erudite that we mostly can't be bothered with that piffle."
My thoughts exactly, we're all really smart!
posted by Blasdelb at 8:37 AM on September 29, 2010
Ever since Tubgirl, I go to links first by opening and closing them in under a tenth of a second. If my mind isn't screaming by what it glimpsed in a flash, I will go back. But that's the sort of thing that messes up stats. Sorry.
Blame Tubgirl.
posted by Astro Zombie at 6:45 PM on September 29, 2010
Blame Tubgirl.
posted by Astro Zombie at 6:45 PM on September 29, 2010
It could also be the result of BoingBoing and Hacker News posting one sad link, compared to MeFi's FOURTEEN links. A lot more interesting stuff than just the article.
To be fair, three links were to Ralph McQuarrie (two to previous threads), someone who is probably known to fans of Star Wars artwork and factoids. Then I duplicated the main link to make it clear who had done the research, and four links to YouTube sources (two to the same vid: one link was directed to the moment where Lucas making the claim about his dog, the second to the beginning of the clip).
And I found the Binary Bonsai piece via the io9 post, and I took a bit of time crafting the piece, so it was quite possible that some folks had read this elsewhere before seeing my post on the blue. Below the break, I told most of the story, and was going to go further but realized it would further limit the interest in reading the (rather long) article, so I stopped.
I'm a bit surprised that Binary Bonsai simplified the story of readership to assumptions about site users, but didn't look at how the story was framed, and the larger readership of the site. The story was "inside baseball" for Star Wars (design/art) fans, so if the question "who designed Chewbacca?" doesn't really interest you, you may not have gone into the linked site for very long, or maybe you read below the break and got enough to satiate your interest, and you glanced at the story to see if there was more.
I'm sorry I tarnished the reputation of MetaFilter, and I'll strive to do better next time.
posted by filthy light thief at 9:47 AM on September 30, 2010
To be fair, three links were to Ralph McQuarrie (two to previous threads), someone who is probably known to fans of Star Wars artwork and factoids. Then I duplicated the main link to make it clear who had done the research, and four links to YouTube sources (two to the same vid: one link was directed to the moment where Lucas making the claim about his dog, the second to the beginning of the clip).
And I found the Binary Bonsai piece via the io9 post, and I took a bit of time crafting the piece, so it was quite possible that some folks had read this elsewhere before seeing my post on the blue. Below the break, I told most of the story, and was going to go further but realized it would further limit the interest in reading the (rather long) article, so I stopped.
I'm a bit surprised that Binary Bonsai simplified the story of readership to assumptions about site users, but didn't look at how the story was framed, and the larger readership of the site. The story was "inside baseball" for Star Wars (design/art) fans, so if the question "who designed Chewbacca?" doesn't really interest you, you may not have gone into the linked site for very long, or maybe you read below the break and got enough to satiate your interest, and you glanced at the story to see if there was more.
I'm sorry I tarnished the reputation of MetaFilter, and I'll strive to do better next time.
posted by filthy light thief at 9:47 AM on September 30, 2010
For one thing, I admit that I often skim long articles and then come back and read Metafilter for context. Yeah, it's bad, but it's also why I stick around.
Secondly, with that particular article, I remember leaving it open in its own tab for hours (I opened it in a separate tab, read the first half, had to go to class, then read the rest in pieces throughout the rest of the day). This doesn't seem like the best indicator of how long a user from a particular site spends reading a particular link or article.
Lastly, it didn't even occur to me that I, like crunchland, block Google Analytics entirely.
posted by Demogorgon at 7:10 PM on September 30, 2010
Secondly, with that particular article, I remember leaving it open in its own tab for hours (I opened it in a separate tab, read the first half, had to go to class, then read the rest in pieces throughout the rest of the day). This doesn't seem like the best indicator of how long a user from a particular site spends reading a particular link or article.
Lastly, it didn't even occur to me that I, like crunchland, block Google Analytics entirely.
posted by Demogorgon at 7:10 PM on September 30, 2010
You are not logged in, either login or create an account to post comments
Post snark
posted by fire&wings at 6:01 AM on September 28, 2010 [6 favorites]