On 10/25/07, Andrew Gray <shimgray(a)gmail.com> wrote:
The latter case, I presume you're thinking of an adsense model. This
makes it even worse than simply "having a hundred advertisers",
because it gets more and more diffuse. As a result, at any given time
we can't easily say which pages are and aren't "tainted" - *or* how
much money we get from any given advertiser. In effect, we have to
worry about the appearance of potential COI on a large fraction of our
pages, albeit not at a very great level.
The advertising world probably has some complex reputation metric for this...
One possibility, which hadn't occurred to me until now, is that having
ads appear with articles will create incentive for ad-buying companies
to edit pages specifically to influence the ads that appear (or don't
appear) on relevant pages--for example, adding obscure words to an
article, and buying ads connected to that word. I'm sure there would
be a number of strategies for manipulating ads based on article
content.
Regarding Erik's earlier arguments against search ads: I mostly agree.
I think big-donor COI problems are much preferably to the problems
that would come with ads, even just search ads. Of course, retaining
control over as much of the operation as possible is a good thing.
But is it really too early to start talking about this? The budget is
USD 4.6m. In about two days, donations are at USD 100,000. Of
course, as Sue pointed out, we can't expect constant donation rates,
but if past fundraisers are anything to go by, donations will trend
down from the initial rate except when matching funds are available.
So just for fun: a sustained USD 50,000 per day rate means 3m for the
whole thing, a considerable shortfall.
What level of donation will it take to hire new developers? Is that
built in to the 4.6m budget? What level of shortfall would it take
before Wikimedia would have to cut back from current staff levels?
-Sage