Re: Accurate user-based log file analysis

Terry Myerson (
Tue, 18 Jul 1995 09:34:48 -0700

Re: Interse' market focus

>I surmise you are trying to do a type of Nielson for the Web. The intended
>target market for this product will be advertising buyers and Web site
>owners to use for justifying their "Web rates" objectively.
>If so, then without solving the problem Brian is referring to, "a lot of
>accesses from Columbus, OH and Redmond, Wa." your statistics are useless.
>The primary uses for these statistics is to quantify retail advertising
>and compare its effectiveness, if this is the right word, with ad buys on
>ESPN, Print ads, etc., as well as other Web sites.
>These numbers have no value to sites who are using their Web site to
>distribute commercial information. I.e. A catalog of left handed
>smoke shifters. A business will have other ways to measure effectiveness.
>These methods are a salesforce, email, telephone etc. When someone acceses
>this info this person already has a predisposition to purchase my product.
>I'm not going to waste my time measuring surfers. It is the same as sales
>has always been. Don't waste your time talking to people who don't have a
>need for your product.
>So...... If you can't provide this retail info, why would a retail web site
>such as Hotwired buy your product? Or, why would anyone spend $500 for your

Putting the discussion of what is a user aside for a second,

Of the domestic users, we consistently see about 15-20% from online services,
and the following characteristics of the remainder:
30% are academic - The location of these users can be determined.
30% are from small businesses - The location of these users can be determined
20% are from large businesses - The location of these users can't be assured

So you can know about 2/3 of your users, or you can know about none of them.
This is not useless.

Terry Myerson
Interse' Corporation
408 732-0932 x-230
408 732-7038 fax