- From: Rigo Wenning <rigo@w3.org>
- Date: Thu, 27 Jun 2013 15:18:29 +0200
- To: John Simpson <john@consumerwatchdog.org>
- Cc: Kathy Joe <kathy@esomar.org>, "public-tracking@w3.org List" <public-tracking@w3.org>
John, On Tuesday 25 June 2013 19:57:49 John Simpson wrote: > I am still having trouble understanding why this should be a permitted > use at all. I don't think that case has been adequately made. Not having a direct interest, let me tell you what I understood so far. the main argument was that audience measurement determines the pricing of advertisement. In order to have this measure, one would need a complete market overview (in terms of how large the audience is and in terms of classes of audience). If there is a significant amount of DNT:1 users, the measures aren't exact enough anymore to calibrate the (opt- in) panel. Calculation of advertisement prices would be not exact enough anymore. As we all hope for lots of DNT:1 users that switch to DNT:0 and back, the content industry (wanting to determine the conditions for revenue) and the measurement industry want to be able to do measurement despite DNT:1 as they count classes not people. Now Rob says: If there are sufficient classes in a matrix, you can single out every person by the combination of the classes a specific person is in. This is why I requested that there must be at least 812 persons in a given bucket. This one requirement is not in the text from Cathy. The rest (proportionality, purpose binding, etc) is there. So there is IMHO only one issue left. Important people to respond to this issue: Rob and Jeff Chester. --Rigo
Received on Thursday, 27 June 2013 13:19:00 UTC