130,000 hits per day! (2003) June 28, 2003 9:37 AM Subscribe
Say no more about the overload, in fact I rather welcome having something like 130.000 hits in one day, something to brag about in the pub when things get a bit dull. That was after my link crashed Johnny Spencer's artchive of Jamaican and American R&B on the 23rd. In disbelief, I wrote and asked for real? and got: 23rd, June = 133,130 hits or 2.2 GB bandwidth. I had no idea we had that much impact.
Metafilter stats page. In April there was an average of 388k successful page requests per day (May link is broken).
posted by eddydamascene at 10:15 AM on June 28, 2003
posted by eddydamascene at 10:15 AM on June 28, 2003
They're out there... watching...
posted by Pretty_Generic at 10:19 AM on June 28, 2003
posted by Pretty_Generic at 10:19 AM on June 28, 2003
karl , can i have your autograph ?
posted by sgt.serenity at 10:34 AM on June 28, 2003
posted by sgt.serenity at 10:34 AM on June 28, 2003
Are you sure all those hits were from this site? Did the link get passed around other blogs? Does the guy have big music or video files on his site?
posted by mathowie (staff) at 10:43 AM on June 28, 2003
posted by mathowie (staff) at 10:43 AM on June 28, 2003
Hits are often page requests, and in analog reports reflect the total number of files downloaded, including graphics...in mefi's case, not many graphics but lots of reloads...in the link you posted Karl, there were lots of graphics per page tho. So I would suspect 130,000 hits is counting each file downloaded...
Wouldn't a better indicator of the number of visitors be distinct hosts served? In the Mefi report for April that would be 326,282 for the month and 96,191 for the last 7 day period.... that would work out to more like a range of 10,000 or 12,000 distinct hosts (or computers) visiting in a day.
At least that's my understanding. Ejikate me, someone.
posted by madamjujujive at 10:46 AM on June 28, 2003
Wouldn't a better indicator of the number of visitors be distinct hosts served? In the Mefi report for April that would be 326,282 for the month and 96,191 for the last 7 day period.... that would work out to more like a range of 10,000 or 12,000 distinct hosts (or computers) visiting in a day.
At least that's my understanding. Ejikate me, someone.
posted by madamjujujive at 10:46 AM on June 28, 2003
According to my google ad stats, MetaFilter's front page and comment pages are being displayed about 2.5 million times each month. There are a lot of other pages, but they get hardly any traffic. Dividing that number by 30 gives about 85k pages a day.
For bandwidth intensive sites, two technologies to look into are bittorrent (pain the ass to implement I hear) and freecache (dead simple, but currently very beta).
posted by mathowie (staff) at 10:48 AM on June 28, 2003
For bandwidth intensive sites, two technologies to look into are bittorrent (pain the ass to implement I hear) and freecache (dead simple, but currently very beta).
posted by mathowie (staff) at 10:48 AM on June 28, 2003
m-j, you are correct in the IP=people assumption. It has its problems (people on dialup count as others every time they connect, and people on AOL or in a corporate firewall count as one IP even though there may be many users behind it), but it's a good rough estimate.
posted by mathowie (staff) at 10:51 AM on June 28, 2003
posted by mathowie (staff) at 10:51 AM on June 28, 2003
I've had parts of my site posted to the blue twice and I got about 6,000 actual mf hits a day for 2 or 3 days. I've also been "via'ed" a few times; that resulted in about 500 hits a day for a few days.
posted by iconomy at 11:03 AM on June 28, 2003
posted by iconomy at 11:03 AM on June 28, 2003
I remember on slashdot when they were talking about subscriptions, they mentioned that the vast majority of visitors don't comment, they simply visit the links referenced. An appreciable number of those people don't even sign up for an account.
Hmmm. Can't find the exact link, but here's one subscriptions story which seems to roughly imply that data (in that 82% of slashdot readers will be able to use a 1000 page subscription to view it for a year).
Anyway, if that generalizes, roughly 4/5 visitors to a weblog/community type site are more interested in just sifting through the news/information than participating.... and more than half of those only view the comments and don't post.
With MeFi, it might well be much, much more, too, because of the restricted participation.
posted by weston at 12:36 PM on June 28, 2003
Hmmm. Can't find the exact link, but here's one subscriptions story which seems to roughly imply that data (in that 82% of slashdot readers will be able to use a 1000 page subscription to view it for a year).
Anyway, if that generalizes, roughly 4/5 visitors to a weblog/community type site are more interested in just sifting through the news/information than participating.... and more than half of those only view the comments and don't post.
With MeFi, it might well be much, much more, too, because of the restricted participation.
posted by weston at 12:36 PM on June 28, 2003
when I first discovered mefi, I didn't even notice that there was a comments option; for the first ?couple months maybe, I just stopped by every now and then for interesting links.
After discovering the comments, I was frustrated for a while (another few months?) that I couldn't get an account. Finally membership opened again, and now I'm likely to click on the conversation before the link (just to see how interesting other people found the link) and I don't ever click on the link without eventually clicking on the conversation.
Still, 130,000 sounds like it was due to how many files each visitor referenced, not how many visitors there were... How many hits does he register on a normal day?
posted by mdn at 1:27 PM on June 28, 2003
After discovering the comments, I was frustrated for a while (another few months?) that I couldn't get an account. Finally membership opened again, and now I'm likely to click on the conversation before the link (just to see how interesting other people found the link) and I don't ever click on the link without eventually clicking on the conversation.
Still, 130,000 sounds like it was due to how many files each visitor referenced, not how many visitors there were... How many hits does he register on a normal day?
posted by mdn at 1:27 PM on June 28, 2003
So. y2karl - was the site's owner actually invoiced for bandwidth overage?
posted by crunchburger at 4:43 PM on June 28, 2003
posted by crunchburger at 4:43 PM on June 28, 2003
On June 2nd, astruc.com got 3,877 hits. On June 3rd, the day of this Mefi thread, I got 12,363 hits. Our UBB slowed down to a crawl for 12 hours. It was pretty impressive.
posted by astruc at 4:46 PM on June 28, 2003
posted by astruc at 4:46 PM on June 28, 2003
My site's been front-paged a couple of times too, and my meta-traffic was more along the lines of what iconomy mentions. I'm far too lazy to actually go try and check the details, though.
posted by stavrosthewonderchicken at 10:03 PM on June 28, 2003
posted by stavrosthewonderchicken at 10:03 PM on June 28, 2003
I'm far too lazy to actually go try and check the details, though.
Hey, man, don't be bitin' my mission statement.
posted by y2karl at 10:41 PM on June 28, 2003
Hey, man, don't be bitin' my mission statement.
posted by y2karl at 10:41 PM on June 28, 2003
When I was recommended on the front page of Webmonkey in April, I got about 800 unique visitors per day for about 4 days.
posted by planetkyoto at 1:57 AM on June 29, 2003
posted by planetkyoto at 1:57 AM on June 29, 2003
I had a page linked on MeTa once and that got about 1,000 visitors a day for the next week or so. A 6:1 ratio of users visiting MeFi as opposed to MeTa sounds about right, I guess (based on iconomy's and stavrosthewonderchicken's figures).
posted by dg at 9:16 PM on June 29, 2003
posted by dg at 9:16 PM on June 29, 2003
My site (sesameseventies.com) was frontpaged a few days ago, and so far my stats program says I got a grand total of 757 visitors direct from MeFi. It's possible, though, that some referrals weren't being counted correctly, because the total visitor numbers were up significantly, at one point up to 2,400 in a day, up from an average of about 200. So probably 1000-1400 were attributable to MeFi.
But Johnny's talking about hits, not visitors, because it's very easy (especially on a multimedia-intense site like his, which has tons of graphics and sound clips) to have a small bump in the number of visitors quadruple (or more) your hit count. 130,000 hits in a month is not inconceivable at all: I'm up to 176,943 this month so far, maybe 1/6 of that due to that MeFi link. That's way above normal.
But not, I might add, unwelcome.
posted by Asparagirl at 2:29 PM on June 30, 2003
But Johnny's talking about hits, not visitors, because it's very easy (especially on a multimedia-intense site like his, which has tons of graphics and sound clips) to have a small bump in the number of visitors quadruple (or more) your hit count. 130,000 hits in a month is not inconceivable at all: I'm up to 176,943 this month so far, maybe 1/6 of that due to that MeFi link. That's way above normal.
But not, I might add, unwelcome.
posted by Asparagirl at 2:29 PM on June 30, 2003
You are not logged in, either login or create an account to post comments
Johnny's OK with the outage--he gave permission when I asked--and now wants to know more about Pay-Pal, hint hint. As in if anyone wants to explain it coherently to someone fairly clueless like him--and CC me!--I'd appreciate it.
No, what's got my attention are all the questions this raises--bandwidth and site crashing being one realm and, um, our personal comportment being another--as in our being MetaFilter's ambassadors to the world right here. Several recent dust ups come to mind on the latter. We're humiliating ourselves in front of a larger arena than I had thought--this is certainly going to affect my writing from now on. And there are oh so many more implications, implications. ..
So, um, by the way... What is our current daily hit rate, anyway?
posted by y2karl at 9:44 AM on June 28, 2003