Re: Demographics

>It is alleged that some advertisers want to pay content providers, not
>by the "hit", but by the "nibble" -- the number of people who actually
>click on the ad to get more info.
[...]
>What I'm looking for are comments on the privacy concerns with such an
>approach.

I had a series of discussions with the folks like clickshare who are 
trying to make money from selling demographic data.

My first approach was to push the referer field - tracking ads was 
one of the original applications I had in mind for it. Its a pity that
the concern for privacy that has reduced the impact of the referer
field was not present when cookies were thrown in. One of the problems
with concerns about security, privacy etc is that the criteria being
applied tend to shift depending on who proposed what. 

I think that any discussion about privacy needs to take account of
the following realities :-

1) Content costs money to provide. In a capitalist system there must
	be mechanisms that cover these costs or content won't exist.

2) Vanity publishing and technology research will not continue to 
	pay for the New York Times etc. indefinitely. A lot of content
	providers have been prepared to give away content for free just
	to learn the potential of the technology. If we cannot provide
	mechanisms to pay for content then sites will soon start 
	disappearing.

3) The current protocols admit any number of ad-hoc hacks that create
	linkage. Most of these mean that documents has to be
	customized for each reader which in turn means that caching will
	not work. This model enforces a communication with the host server.

4) Payment for content on a subscription model limits the audience for
	a product, it means that the rich inter-linked nature of the
	Web is lessened. The marginal cost of following a link becomes
	very substantial. If charging mechanisms are restricted to
	subscriptions alone the objective of disintermediation, removing
	the power that Murdoch, Maxwell and their cronies have over
	the movement of information will be lost. We will only be able
	to buy content that comes from large publishing corporations.
	There will not be the leavening of small independent 
	self-published works.

	Consider the reason why the software industry is so inovative.
	A large number of the key ideas come from independents who can
	inovate without the constraints of consistency of corporate
	view that constrains large companies. This keeps the large
	companies honest. They have to keep up with the pace set by the
	small companies or watch their  market disappear. I want to
	see that type of competition in news reporting so that the whole
	story comes out, not just the part that suits the politics of
	a newspaper proprietor. Remember the Time Cyberporn article?
	In that case Time were caught out and they retracted (if 
	grudgingly). Normally the would not have retracted, they publish
	the facts so the facts are defined to be what they published.

5) Current anonymous payment schemes are not practical for small 
	payments. Public key cryptography is barely practical even 
	with amortisation schemes such as payword. Empirical evidence
	suggests that the holder of key patents has an unfounded
	belief that electronic commerce cannot take place without him.

	Consequently any payment scheme for small or large payments
	is almost guaranteed to NOT provide the unlinkability which is
	technically feasible. 

	Therefore if we reject advertising as a model on the basis
	that it has privacy problems we are likely to end up with a
	model where the purchaser BOTH pays and loses privacy.


The requests that I have had for demographics tracking have had to 
take account of the following "needs".

1) Identify readership
  a) volume (exposures/hits)
  b) interest in/purchase of advertised product
  c) demographic grouping 

2) Provide audited measures of above in a manner which is
  a) consistent across content providers
  b) commensurate with other measurements from other media.

2b is a key point, the recent Nielssen report recieved criticism from
a certain quarter because it alledgedly overstated the number of
Internet "users". In fact the study had deliberately chosen a
definition of "user" which was commensurate with other media, in this
case the number of people in a household wich had internet access and
not the number of people who had used the internet. Why measure it
this way? Well consider a case where you want to buy a new car, you
know that your son can get details of prices via the internet, you
ask your son to do the search. In other words the potential audience
outreach is not necessarily the direct readership. There are similar
fudge factors for other media (pass on readership of periodicals).

One important point is that the advertisers are not really interested
in "readers", they are interested in an index that corresponds with the
measures they already use. 


		Phill
  

		Phill

Received on Wednesday, 10 July 1996 13:00:31 UTC