Sent by Techwatcher on 26 June 2002 14:02
> Naturally the stats for your own site may be quite different. The
stats for
> *any* individual site will be different, as that is a much smaller
sample.
> Your individual sites are by definition much more skewed, as the
sample is
> much narrower. But, for making decisions about your own site then
your own
> stats are of course the most relevant.
Actually, if the variable of interest is browsers for specific site,
his sample is not skewed at all, since technically he is measuring the
value for the *population* rather than any sample.
> If you have good stats on *your own* site, then clearly that is great
> information. The visitors you most want to know about are your own
visitors.
> But that narrow sample from your site, or anyone else's single site,
is of
> little or no value to me in making decisions about my site.
Again, a little technical statistical/market research stuff here:
Depending on the extent to which target audience for his site
accurately represents target audience for your site, his specific
site's stats may be of *greater* value than overall stats! That's why,
for example, when I consulted for Pfizer Pharmaceuticals about 20 years
ago, we figured out which specific physician specialties would be
relevant for a new drug (or at least, for its approved indicators),
then investigated and compared costs of ads in *different* magazines
against the circulation data for those specialties. (That demographics
and cost comparison was the raw data for choosing how to place ads for
new products in a system I re-programmed.)
So if by some chance you're both writing a site for programmers or Web
developers, his site's stats about browsers are likely to be better
indicators for you than overall browser stats.
Conversely, I wrote a site for writers catering to old browsers,
because I know a larger than normal proportion of that audience has old
browsers.
> As a supplement to your own local site stats, or in the absence of it
(e.g.
> when you're planning a brand new site), orthese large stats sites are
a
> useful indication of broader trends and patterns.
Cheers
Carol Stein
[EMAIL-REMOVED]
