Looks like Patrick was right, and I was wrong, about the SiteMeter visitor count. Not about everything, mind you. But definitely about the major issue.
The things he was right about: If you stay on a high-traffic site while the visitor count refreshes, it counts you as a new visitor. The 30-minute rule is also in effect. If you drop off the visitor count, you do go back on as a new visitor. (Not unique, though; see below.)
The things he was wrong about: Unique visitors. Sorry, Patrick, but nope, they’re just visitors. SiteMeter doesn’t track uniques. As for having the ability to log more than 100 (or 4,000) visitors at a pop: Here, too, you are wrong. I have been informed by SiteMeter staff that arrangements could be made for higher-traffic sites. Which means they do collect the data, and it does not disappear into the ether.
The things we were both wrong about: SiteMeter does not use cookies to track visitors.
The things I was right about: Sorry, Patrick, but unless you have some programming background that I’m not aware of, you are a technical neophyte. That doesn’t mean you can’t catch things like the SiteMeter double count, and that wasn’t what I was implying. I was implying you were lacking a certain strategic depth to your knowledge due to a lack of understanding of certain technologies. But I was wrong about the concept you were getting across. I simply couldn’t fathom the idea that someone would write a visitor-tracking program that would allow a visitor to become brand-new if someone were still on the site, or clicking to another page. It seemed utterly counter-intuitive to me. But hey, there’s a lot of programming out there, and site statistics are utterly impossible to track with 100% accuracy. I have three different site trackers and three different traffic counts. I was also pretty much on the money about clickthroughs. You really can’t judge a DKos clickthrough rate as indicative of traffic rate. Clickthroughs from referring sites depend on so many reasons that they’re not indicators unless the site is specifically geared to drive traffic (a la Instapundit).
Things I can’t say for sure: I don’t know if his 60% figure is accurate. That one I think is not easily charted. Testing doesn’t tell you how many people are doing the thing you are testing for—only whether it works or not.
I could go point-by-point on the rest of the things, but, well, the argument is done.
“I have been informed by SiteMeter staff that arrangements could be made for higher-traffic sites. Which means they do collect the data, and it does not disappear into the ether.”
I didn’t see the email you got from Site Meter, but to me, “arrangements could be made” isn’t the same as “data is currently being collected”. I still don’t think Site Meter is saving any data at all on visitors once they drop off the top 100 list. I think the only permanent record of a visitor is that the aggregate counts are incremented (daily, monthly, annual, total, etc.). From the way you described your email from Site Meter, I think they are just saying that 4000 is not a hard limit, and they could store even more if you asked them to.
The most important point, though, is that only the visitors who did the last 100 page views are currently being tracked for users with the free account, and that is greatly inflating traffic numbers on several high volume sites. Kos in particular.
And if he were to upgrade to the premium account, where the details on 4000 visitors were tracked instead of 100, then his visitor counts would drop dramatically.
Matthew… it’s over. You’ll have to ask SiteMeter yourself if you want answers. I tend to lose interest in a question when it’s, you know, settled.
Hey, I was just sayin’…
And if you really don’t want anyone commenting – disable comments on your post. You know how to do that, right? Or are you lacking “a certain strategic depth to your knowledge”?
Matthew, now you’re being borderline annoying.
Show me a programming degree from either you or Patrick, and I’ll be happy to not mention that you have no technical backgrounds to speak of, and yet, you are spouting off on technological issues as if you do. If you can even tell me what the http in a web address stands for without Googling it, I’ll admit you’re at least semi-knowledgeable about tech.
You’re on the honor system there, bud.
But the last thing–the very last thing–you want to do, is snark with me in my backyard.
No sense of humor today, gotcha.
Yes, I do know what HTTP stands for, and I was using it back at Rice in ’94 when the only browser around was NCSA Mosaic. I was once quoted as saying there was no way this HyperText crap would ever replace Gopher.
That being said, I won’t try to win a pissing contest with you – I write client/server business software for a living and am just a hobbyist when it comes to web pages. I’m really not that good at it, and if we got into some kind of IIS or Apache trivia match, you would probably obliterate me.
I just want to make sure it gets out there for the permanent record that a lot less people read DailyKos each day than are currently advertised and there is an easy way for Kos to fix this. This is important, because his readers cite the size of Kos’s readership as a key component of their political muscle. Being politically opposed to them, I want that muscle to shrink, and outing their numbers as highly inflated is a good way to do that. Successfully painting Kos as someone knowingly keeping those numbers inflated is even better. Thus, I want to make sure HE knows his numbers are too high, so that Kos can not claim ignorance as a defense.
So that is my motivation to defend Patrick’s post. Yeah, I got snarky, but I’m having a bad day at work and figured it was safer to take it out on you than my boss.
OK, enough of this. You can have the last word and I’ll shut up for today.
I think, Matthew, we were both taking it out on each other. Although I do have a fairly quick temper. Online, anyway. Let’s shake hands (virtual shake) and forget about it.
I really don’t think Patrick is right on the 60% inflation rate, though. Something to work on next week, perhaps.
Hi Meryl,
GoStats is web stats service with accurate visitor counts. (since the entire day of traffic is considered) Doesn’t hurt to try out GoStats. :)