On Tue, Dec 23, 2008 at 5:59 PM, Erik Moeller <erik(a)wikimedia.org> wrote:
Within the last 24 hours, we've raised a total of
$283,859. That's
more than 10 times as much as we made during a typical weekday in the
last few days of the fundraiser, and the single highest day on record
for community gifts. We don't know yet how steep the inevitable
drop-off will be, but it's obvious that the appeal is working beyond
everyone's expectations.
I think it's worth noting that this tenfold increase has been possible
without the use of additional pixel real estate, without scrolling
marquees, interstitials, or other serious interruptions of the
Wikipedia reader/editor experience. All it took were less than 60
characters of text on each page in a highly visible font, linking to a
personal appeal that makes our case in more detail.
We should ask ourselves why it is that based on the previous
sitenotices, 9 in 10 people who would be clearly willing to give to
us, did not do so. There seem to be at least three principal reasons
for that:
* The previous messages were below the visibility threshold for most
people: They considered them to be an unimportant part of the page
that should be ignored.
* The previous messages did not, clearly enough, make a case for
giving. They appealed to people who instantly "get" the non-profit
donation model, but not to those for whom Wikipedia is essentially the
same as any other website. The appeal directly addresses this
distinction, to the satisfaction of a great number of people.
* Because it's a personal appeal, rather than an impersonal donation
message, the letter seems more likely to resonate with people.
<snip>
I would opine that points 2 and 3 are the core characteristics, with 2
somewhat ahead of 3. Most of the banners are quite visible, and so I
think 1 is negligible factor. Or perhaps more directly, I think most
of the banners are visible to the point that people notice them, but
after reading them many fail to care about that message they offer.
For example, both the donation bar and the scales graphic starkly
standout on the page, and yet they are no where near as successful.
(I also suspect that the ability to extract gains by making the
message more visibile has already been saturated, and one could
probably reduce the height of the banner by 1/3 or so with little
marginal change in the response rate.)
So, if not visibility, then what is really going on. In my opinion,
if you want someone to read something, personalizing it is a very good
idea. I think describing it as a personal message and putting a face
to it, provides engagement and gets people to pay attention. That
Jimbo has excellent name recognition helps (if it were Sue or Michael
Snow, for example, I don't think it would do as well).
But ultimately, once one captures eyeballs, I think the biggest factor
in getting people to hit the big red button is message. We tend to
forget that among the 100s of millions of people that occasionally use
Wikipedia, a substantial fraction don't really understand our
operation or our goals. Saying "we are a non-profit" or a similar
banner-sized message doesn't capture who we are in the way the longer
text can. I suspect that simply providing the larger community with
more information about what-the-hell-Wikipedia-is goes a long way to
encouraging donations. It also suggests that the current donations
landing page could probably be improved by providing more of that
information.
If I am right that the new message captures a larger number of people
with only a casual familiarity with Wikipedia, then one might also
guess that the donations early in the drive tended to come more from
hard-core Wiki supporters who were already well acquainted with who we
are and how we work.
-Robert Rohde