[Foundation-l] Re: Our exponentially increasing costs (was Re: Re: Answers.com and Wikimedia Foundation to Form New Partnership)

[Anthony DiPierro]

>
>
>
> --- Delphine Ménard <notafishz at gmail.com> wrote:
>
> > On 10/26/05, Jtkiefer <jtkiefer at wordzen.net> wrote:
> >
> > > I thought it was generally agreed upon though that rising costs on
> > > hardware could be compensated for, since users are more than willing
> to
> > > donate for hardware costs. Correct me if I'm wrong but couldn't
> > > wikimedia just adjust their donation goals to cover that and thus
> cover
> > > the cost of new hardware/repairs/upgrades, etc...?
>
> So far, that is pretty much all we have been doing (although the
> foundation
> also wants to do much, much more than that). But buying more hardware and
> serving lots more people also increases overhead/administrative costs. So
> while
> the percentage that overhead takes from the budget has been in the 20-30%
> range, the total amount has grown in proportion to hardware and hosting
> costs.
>
> > Daniel made it clear, the curve for donations won't follow the curve
> > for costs. Since our costs are growing so fast (I won't go back to all
> > the explanations given here and elsewhere), I don't think that
> > fundraising drives will be near enough to cover those costs.
>
> Actually, I just said my gut feeling was that the donation curve would not
>
> follow the cost curve forever. What is needed is a statistical assessment
> of
> the two. However, looking *only* at two fund drive data points and two
> traffic
> data points gives some reason to be optimistic.
>
> Q3 2004 Fund Drive grand total: $44,863.95 USD
> September 2004 unique visitors to wikipedia.org <http://wikipedia.org/>:
> 3.2 million
>
> Q4 2005 Fund Drive grand total: $243,930 USD
> September 2005 unique visitors to wikipedia.org <http://wikipedia.org/>:
> 12.8 million
>
> So, there has been a 533% increase in donations for a 400% increase in
> traffic.


I know you've mentioned that this is only two data points, but let me point
out an important one that's missing. How many of the people who donated in
Q3 2004 also donated in Q4 2005? Are donations primarily a one-time thing
for new users, or are they an ongoing thing? If the former, the biggest
problems might not come until the growth slows down. Of course, at that
point the need for new servers won't be as severe, but hosting costs will
continue to steadily suck money out of the community.

I guess even without the growth slowing down, it's likely going to move
toward people with less means to contribute. That'd be another interesting
statistic, though it'd have to be approximate (we know what countries the
visitors are coming from, and could look at the changing distribution by
country factoring in the median income).

> Throw in Moore's Law and things look even better. However, the traffic
> numbers
> are only for wikipedia.org <http://wikipedia.org/> and this only tracks
> unique visitors; those visitors
> likely use Wikipedia more often today than they did last year. The other
> projects, esp Commons, Wiktionary and Wikinews are starting to become
> popular
> as well. So I'll tend to be pessimistic until I'm able to analyze some
> hard
> data for every one of our domains (data with consistent standards - esp
> for the
> non-wikipedia.org <http://non-wikipedia.org/> domains - has been difficult
> to come by).
>
> > Small donations originated by fundraising drives are something we
> > should *count* on, but not something we should *rely* on. We're not
> > the only ones out there asking for money, and I believe that it is
> > also our job to be looking elsewhere for hard cash.
>
> Well, we kinda do have to rely on small donations. Small donations, not
> big
> ones or grants, make up the vast majority of our income. Relying too much
> on
> any one or select few big donors, grants or income from partnership
> agreements
> puts us in jeopardy if any one or several of them pull their funding. But
> the
> answer to that is to have *lots* of big donors, grants, and income streams
> from
> partnerships. However, I think your point was that we should not simply
> rely on
> donations in general as our sole income stream. Yep - I agree with that
> 100%.
> We also need to diversity our income in general - we can't be too reliant
> any
> one source


Not just practically, but legally. The difference between a public charity
and a private foundation is whether the revenues come from a broad base of
people or just a select few. Creative Commons just recently started its
first fund drive, not just because they needed the money, but so that they
could meet the IRS "public support test". See
http://creativecommons.org/weblog/entry/5661 and
http://creativecommons.org/support . I'm not sure if advertising would count
as public support or not. I initially thought it would (under section
509(a)(2), if you care) but I've read some IRS rulings recently which have
made me reconsider.

> In our case, I would say that all editors, by spending the time they
> > spend providing content and making Wikimedia projects what they are,
> > are already donating. I would not blame anyone who'd argue that they
> > "don't have cash, but already give some time and knowledge". Except
> > that time and knowledge, although they are indispensable, don't buy
> > servers or cover operating costs.
>
> I get several emails in the ORTS donation queue each week from people who
> either don't have cash or who want to donate in a currency that we don't
> accept. I tell each that their donation of time in either contributing
> content
> or telling people about Wikipedia is just as important as giving money.
> That
> Wikipedia would not exist without volunteer effort.
>
> -- mav


This is the reason I think a more P2P based network would be such a great
solution for the long-term. People wouldn't need to donate cash to cover
operating "costs".

Anthony