A Personal Appeal from Wikipedia Programmer Brandon Harris
(Note that this testing is ongoing; there will be further tests, so don’t be surprised to see me there in the future.)
I have a lot of words I want to say about this but really there are two subjects: the process as to how banner ads are chosen, and my personal feelings as to why you should, like, donate a fucking dollar to Wikipedia.
Today, I’m going to explain the process.
Every year, a small team of kick-ass people get together and formulate a plan to raise enough money to keep the fires at the Wikimedia Foundation burning. It’s an important task and one I most certainly do not envy: it is an extremely stressful job, composed of long-in-the-tooth days and punctuated by lots of armchair criticism.
It’s also very scientific.
Many non-profits are funded through “large gift” systems. That is, they subsist via donations and grants from wealthy philanthropists or organizations. Typically the amounts involved are in the tens of thousands of dollars but can reach into the millions (such as large trusts). Obviously, a smart charity with an agile team of large-donation fundraisers can quickly make some serious cake with this model.
The problem with the “large gift” model is that it opens the charity to being hamstrung or possibly being perceived as biased. Large grants often have strings attached: “Here is a million dollars, but you have to spend it building this feature.” Conditions like this aren’t flexible: you can’t take some of that money and spend it to keep the site running.
The more insidious problem is one of bias. If the Foundation accepts a multi-million dollar donation from a shady oil company, even if there are no strings attached, what does that say? Is there an implication that Wikipedia will be “soft” on the articles written about said shady oil company?
What happens next year, when shady oil company says, “So, we’d like to double our gift this year, but there’s some rather. . . unfortunate. . . words about us in your Wikipedia?”
Can’t do it. Won’t do it.
Enter the “small donation” model. The Foundation is fueled by thousands and thousands of “small” donations from “normal” people – the people who use it every day, me and you. The information in Wikipedia is crowd-sourced. Why not its funding, as well? In this way we avoid both the strings and the bias.
We obtain our small donations via banner ads. These banners are handled with science and undergo some rigorous A/B testing. The fundraising team creates a banner and writes (or handles) the landing page “appeal”. Multiple versions of these combinations and wordings are then run against each other in a kind of “Fundraising Thunderdome”: Two banners enter, one banner leaves.
The banner and appeal that has the highest monetary gain is the winner – until a new banner is created that tests better.
Enter the “Personal Appeal from Wikipedia Founder Jimmy Wales.” These banners – we call them “Jimmy Banners” – make an incredible amount of money compared to any other banners that are tested. So much more that it’s frankly ridiculous and irresponsible to not use them almost exclusively. There’s been a lot of exploration into exactly why this is true and various theories have been brought forth and played around with.
This year, the Fundraising team is playing around with “personal appeals” from other people associated with the Foundation. The other week, several members of the staff were interviewed (on camera) by the Foundation’s “storytellers” so that they could better write appeals in our own words. Why were we here, and why was this important.
I am, apparently, a “walking sound bite,” which is why I was first in the chamber. And that’s why my face is on Wikipedia.
(Spoiler: I made more money.)