Re: Which version in LH for?
Tim wrote:[color=blue]
> On Sun, 10 Aug 2003 05:33:02 GMT,
> Jane Withnolastname <JaneWithnolast nameNOSPAM@yaho o.com> wrote:
>[color=green][color=darkred]
> >> Well, for the number of people here who would be so interested, it's
> >> called a user survey. I ask people what they use. I know it's not as
> >> high tech as sneaking a JavaScript in and checking under the hood, but
> >> it is more honest and, as you will no doubt understand, more
> >> effective.[/color][/color]
>
>
> On Sun, 10 Aug 2003 06:17:42 GMT,
> Greg Schmidt <gregs@trawna.c om> wrote:
>[color=green]
> > This is what's known in statistical circles as a self-selecting sample.
> > While they can be useful, they are notoriously unreliable for
> > determining facts. I'm not sure whether this would be more or less
> > accurate than looking at user agent strings in your logs. Then again,
> > if the only fact you're interested in is knowing whether the "majority
> > of browsers" are using IE, I think any of us could have told you the
> > answer without resorting to a survey. Last I checked, lawsuits were
> > still ongoing because of that very fact.[/color]
>
> By way of illustrating Greg's (correct) assertion. I've had 15,000 hits
> notch up on one of my pages in the last year (most of it in the last few
> months), about 2,000 hits on the linked homepage, and I've received
> about 56 e-mails from visitors about the site (thank god they all didn't
> e-mail me). ;-\
>
> I have no idea about what the rest of the visitors thought about the
> site, nor anything else about the site and them. Whether they found the
> page accidentally, and only loaded enough of it to register as a hit,
> before going elsewhere. Whether the page was what they wanted. Whether
> they read it all. Whether they agreed with it. Whether those hit
> counters are the same people reloading the page, or different people, or
> even "people" visiting the site. How many visited without loading the
> hit counter (it's only there for my own curiosity, anyway, as I don't
> have access to the server logs). And so on.
>
> i.e. There's a plethora of "assumption s" that could be made about how
> people used the site, and that's all that they are.
>
> snip[/color]
actually you can go a little further than
that...providin g you have access to logs that is
you can look at trends month on month to see which
referrers are gaining and losing ground, which can be a
useful hint as to when different types of site promotion
are likely to be most effective
you can look at trends on the usage of individual pages
to see where there may be pages that lose visitors
etc
what you can't look at is how many people are using the
local caches at the big ISPs, at least not without
compromising the usability of the site...and you haven't
a cat in hells chance of deriving ANY usable information
about browser usage, particularly since a lot of the
locally cached access is liable to be linked with a
preponderance of a specific browser (my ISP still ships
IE5 as standard, then there are AOL and the various Web
TV providers)
so site stats are not useless...but there is no way site
logs can ever provide any useful info on browser
usage...and even if they could it isn't last month's
browser usage that is important...it is what people will
use between now and the next page redesign that matters,
and experience shows that the two can be very different
--
eric
"Hey Lord don't ask me questions
There ain't no answer in me"
Tim wrote:[color=blue]
> On Sun, 10 Aug 2003 05:33:02 GMT,
> Jane Withnolastname <JaneWithnolast nameNOSPAM@yaho o.com> wrote:
>[color=green][color=darkred]
> >> Well, for the number of people here who would be so interested, it's
> >> called a user survey. I ask people what they use. I know it's not as
> >> high tech as sneaking a JavaScript in and checking under the hood, but
> >> it is more honest and, as you will no doubt understand, more
> >> effective.[/color][/color]
>
>
> On Sun, 10 Aug 2003 06:17:42 GMT,
> Greg Schmidt <gregs@trawna.c om> wrote:
>[color=green]
> > This is what's known in statistical circles as a self-selecting sample.
> > While they can be useful, they are notoriously unreliable for
> > determining facts. I'm not sure whether this would be more or less
> > accurate than looking at user agent strings in your logs. Then again,
> > if the only fact you're interested in is knowing whether the "majority
> > of browsers" are using IE, I think any of us could have told you the
> > answer without resorting to a survey. Last I checked, lawsuits were
> > still ongoing because of that very fact.[/color]
>
> By way of illustrating Greg's (correct) assertion. I've had 15,000 hits
> notch up on one of my pages in the last year (most of it in the last few
> months), about 2,000 hits on the linked homepage, and I've received
> about 56 e-mails from visitors about the site (thank god they all didn't
> e-mail me). ;-\
>
> I have no idea about what the rest of the visitors thought about the
> site, nor anything else about the site and them. Whether they found the
> page accidentally, and only loaded enough of it to register as a hit,
> before going elsewhere. Whether the page was what they wanted. Whether
> they read it all. Whether they agreed with it. Whether those hit
> counters are the same people reloading the page, or different people, or
> even "people" visiting the site. How many visited without loading the
> hit counter (it's only there for my own curiosity, anyway, as I don't
> have access to the server logs). And so on.
>
> i.e. There's a plethora of "assumption s" that could be made about how
> people used the site, and that's all that they are.
>
> snip[/color]
actually you can go a little further than
that...providin g you have access to logs that is
you can look at trends month on month to see which
referrers are gaining and losing ground, which can be a
useful hint as to when different types of site promotion
are likely to be most effective
you can look at trends on the usage of individual pages
to see where there may be pages that lose visitors
etc
what you can't look at is how many people are using the
local caches at the big ISPs, at least not without
compromising the usability of the site...and you haven't
a cat in hells chance of deriving ANY usable information
about browser usage, particularly since a lot of the
locally cached access is liable to be linked with a
preponderance of a specific browser (my ISP still ships
IE5 as standard, then there are AOL and the various Web
TV providers)
so site stats are not useless...but there is no way site
logs can ever provide any useful info on browser
usage...and even if they could it isn't last month's
browser usage that is important...it is what people will
use between now and the next page redesign that matters,
and experience shows that the two can be very different
--
eric
"Hey Lord don't ask me questions
There ain't no answer in me"
Comment