Subject: Jump limits = hit conversions? From: PJacobson@aol.com Date: Wed, 20 Sep 1995 02:51:44 -0400

Bruce Siceloff and  Christopher (Kit) Spence have graciously provided some
data on hit conversions that may provide an answer to my earlier question of
how many links on a page are too many. If this is not an answer, then at
least it is a good start.
   From Inter@ctive Age, via Siceloff: 

 >>You'll see there was considerable variation in these ratios. 10:1 for
 Netscape, 3.4:1 for Pathfinder, 4.1:1 for Playboy, 7:1 for HotWired,  14.8:1
for NandO.net, 2.86:1 for Penthouse.  And were any of these ratios  accurate?
  <<

and from Christopher (Kit) Spence:

>>At canstock all users must login so we know how many people have been to
our site and we can see how many unique site visits there have been as well
as how many visits.  Today, for example, the Statbot says that there have
been 20 unique site visits and 125 total web site visits - about a 1-6 ratio.
 Overall, however, there have been 252 unique site visits and 6026 total
visits - a 1-24 ratio.  Now, the latter figure is likely to be less accurate
since there was some development going on during the time that the stats were
collected. The former ratio (1-6) may be more valid.  In fact, we have the
figures for the past six days and the ratios are 1-8, 1-4, 1-5, 1-8, 1-10,
1-9, where the 1-4 and 1-5 figures are on the weekend when we have less new
content and the markets are not open.  Adding up the numbers
the ratio is 1-7.5. <<

   If we assume that caching browsers and non caching ones will balance each
other out, then it would look like we have people making anywhere from 3 to
24 jumps off of a home page -- with the vast majority of those home pages
having readers (browsers?) who jump away and return about 7 times.   
    Was it Bill Skeet who proposed the range of 5 to 9 jumps? (Seven,
plus-or-minus two). He may have hit the bulls-eye. 
   Even if the assumption on caching browsers equalling non caching browsers
is not valid,  these numbers do help set some reasonable boundaries for the
research. 
   These numbers may also influence the number of links a person is going to
place on a given homepage, and with that, they will influence homepage
design. 
   For example, you can either have one long webpage with 100 links on it, or
1 webpage with 10 links on it, and each link leads to another page that also
has 10 links on it. The final outcome n both cases is that you link to 100
files. In the former place you do it in one big document, writing 100 links,
and in the other you do it in 11 smaller documents ( the primary, and the 10
secondaries), writing 110 links.
   If you get more user appreciation from the second method, then that will
become the "best" design for web pages. 

 Paul Jacobson




   

------------------------------

End of online-news-digest V1 #344
*********************************


From owner-online-news-digest@marketplace.com Wed Sep 20 19:00:12 1995
Received: from marketplace.com (majordom@marketplace.com [199.45.128.10]) by cnj.digex.net (8.6.12/8.6.12) with ESMTP id TAA06359 ; for ; Wed, 20 Sep 1995 19:00:08 -0400