|Urchin claims 100% succes on referrers|
So can I use it for my affiliate scheme?
| 7:25 pm on Nov 3, 2004 (gmt 0)|
The Urchin tracker claims to get 100% success rate for tracing referrals, thanks to a patented js technique that beats the much-talked-about caching problem:
"when the UTM is used, all referrals (even those that are cached) are forced back to the local web server and accuracy in referrals is restored"
If this is true, I could use Urchin for my affiliate scheme instead of my current cgi tracking links, get huge google search engine placement benefit because they'd be clean links directly to my site ... and thus rule the world...
But before I bankrupt myself buying the thing, is there anyone out there who has scientifically tested Urchin's claim and willing to share their knowledge?
(PS. Please don't come back with a 'dangers of using logfiles' discussion. We know the problems with caching: the whole thrust of Urchin's claim is that they've overcome it. Thanks).
| 7:30 pm on Nov 3, 2004 (gmt 0)|
These two bits don't match:
1) 100% success rate for tracing referrals,
2) js technique
So, no way it's going to be 100%.
What they do is probably to add a random string, that's a quite normal procedure to get beyond caching, everybody does that.
Patents on stuff like that is a laugh... uhmm... i'll change that, it's quite tragic, actually: [webshop.ffii.org...]
| 8:01 pm on Nov 3, 2004 (gmt 0)|
Any other reactions on this?
| 10:26 pm on Nov 3, 2004 (gmt 0)|
I'm not 100% sure i understand what it is that you want to do.
Math.random() - that's the same thing you do when you want ad banners to get around the cache.
If, otoh, you want your own web page not to be cached, just include random content on it.
I guess what Urchin does is to put a 1x1 pixel gif on the page, which is not really a gif, but a call to a tracking script in stead. Then, in the URL to the tracking gif, they include a random string to make sure the tracking "gif" is not cached. That's the standard way of doing that kind of stuff.
| 10:55 pm on Nov 3, 2004 (gmt 0)|
What I'm wanting to do is to find a reliable way of tracing referrers so that instead of giving them a personalised url (cgi or whatever) they just place ads, reviews etc on their site that link directly to my site. Not only would I get the benefit of the ads, but I'd also I'd get the search engine benefit of having hundreds of links to my home page that the search engines wouldn't see as ads (sorry if this is sounding too machiavellian for some people's tastes).
This is impossible under most systems, because referrer details don't always get to the log files (a) because some people view the site on a proxy server and (b) because in log statistics direct hits (people writing in the site address or accessing from favorites) are over-represented & the logical corollary is that referrer visits are being under represented.
As I understand it, the Urchin system solves (a) by forcing a direct hit, even if my pages get cached on a proxy server. I'm not sure whether this indirectly solves (b) or whether (b) remains an insoluble problem arising out of users local browser settings, antivirus programs etc, and if so to what extent.
Of course overriding caches creates a performance issue, but if it takes me up the search engines (because lots of sites are linking directly to my homepage), that's a risk that's worth calculating.
| 11:27 pm on Nov 3, 2004 (gmt 0)|
Did Urchin say that they're the only ones that have this? As far as I can see, it's pretty generic.
| 11:57 pm on Nov 3, 2004 (gmt 0)|
Urchin does not have anything near a 100% success rate - and nor do any of the other programs - they'll all miss something (allthough Urchin is pretty good compared to some).
The overriding thing here is - if you were thinking of joining an affiliate program and somebody said to you: "it's ok, you don't need tracking codes, I can pick it all up through my stats program. I'll send you a report at the end of each month, don't worry about it." Would you trust them?
| 8:59 am on Nov 4, 2004 (gmt 0)|
I'm maybe giving Urchin a bad name by overstating their claims for them. What they say is
"the patent-pending Urchin Traffic Monitor (UTM), originally available in Urchin 4, was specifically designed to provide the most accurate measurements of unique website visitors, etc, etc, "
'patent-pending', to the unwary like me, certainly gives the impression of a unique service, even though I now realise it's probably quite widespread (thanks everyone, for putting me right).
However they offer good reporting facilities that you couldn't write for yourself in a month of Sundays, so their offer remains an interesting one.
I agree the confidence issue is a big problem. On the other hand, it depends on how you sell it. If someeone were to come to me and say "I'll give you $50 to put an ad on your site and I'll up your rates if the results are good" I might be tempted. You'd reserve the right to withdraw the offer from non performing sites.
Maybe this all sounds a little half-baked (it does to me too!), but on the other hand the indirect search engine benefit could be huge...
| 9:10 am on Nov 4, 2004 (gmt 0)|
We all trust adsense. If you get better money than anywhere else then you go with what you have to. I have used affiliates that just send me checks and I have no say in it. They just send really big checks. Affiliates that just send you money is quite normal. There is no way to know if they are cheating you or not.
| 9:56 am on Nov 4, 2004 (gmt 0)|
|Urchin does not have anything near a 100% success rate |
This is true, a few weeks ago I felt that urchin was not correct in the amount of visitors I was getting from google so I downloaded the log file and checked manually.
I was very shocked when it was an extra 640 visitors (a day) which urchin had not picked up, up until now I had put trust in urchin now I am not so sure.
I have used this product for 4 years starting with urchin 3 and now with urchin 5
Don't get me wrong urchin is still a very good product and I will continue to use it knowing that I am getting more traffic than its showing which is still better than the other way round.
| 10:37 am on Nov 4, 2004 (gmt 0)|
That's very interesting information, ncw164x, thank you very much. When you say '640 a day' what is that in % terms compared with the number of visitors that were recorded, so that I can get an idea of the ball park accuracy figures we're talking about.
| 5:28 pm on Nov 4, 2004 (gmt 0)|
And, did you figure out why these were missed?