Wikimedia’s fundraiser – which banners click?
November 25th, 2008
Admittedly, this is a year of growth and testing for the Wikimedia Foundation Fundraising team. We have 4 new members and the Annual Fundraiser is a new experience for all of us. In fact, I’m not sure how many fundraising projects have had this kind of reach (250+ million unique viewers in November). This provides us with an amazing opportunity to test different pitches through our site notices. With the fundraiser reaching a huge audience, we knew we had a great chance to test different messages and see what works and what doesn’t.
We started the Annual Fundraiser on November 3rd with 4 site notices (the big banners across the top of every Wiki article). Our tech team worked to track each notice and each notice had a randomized 25% chance of displaying on any given article (on every Wikipedia, in localized languages, and in other other Wikimedia projects). In theory, every notice had the same number of views. You can see the 4 site notices here: http://meta.wikimedia.org/wiki/Fundraising_2008/design_drafts.
Now, which notices do you think did the best? The results are surprising:
| Phase 1 | 11/3 to 11/17 | Display | Click | Number | Percent who |
Total | Average | |
| % | Throughs | Donations | Donated | Given | Gift | |||
| Edu1 | Wikipedia is a non-profit project: please donate today. | 25.00% | 6423 | 994 | 15.48% | $28,936.00 | $29.00 | |
| Edu2 | Wikipedia relies on your donations: please give today. | 25.00% | 44482 | 4444 | 9.99% | $126,664.00 | $28.50 | |
| Market1 | Wikipedia is there when you need it — now it needs you. | 25.00% | 29886 | 5071 | 16.97% | $140,913.00 | $27.00 | |
| Market2 | Wikipedia: Making Life Easier. | 25.00% | 56577 | 5620 | 9.93% | $155,136.00 | $27.60 | |
| No meter 1 – collapsed | n/a | 13839 | 1156 | 8.35% | $33,208.00 | $23.00 |
People love to click on the links with the thermometer…but less than 10% donated after clicking-though. However, it’s interesting to see the strength of the “Wikipedia is there when you need it — now it needs you” message. While it had significantly less clicks, nearly 17% of people donated after clicking on it.
What do you make of that? What other conclusions would you draw?
And what do you think our next test should be?
-Rand Montoya
Head of Community Giving<
November 25th, 2008 at 20:00
You should switch the thermometer to the messages that don’t have it. Then, I think it would be interesting to compare all the messages without the thermometer to each other and then all the messages with the thermometer to each other. Although you would be comparing data collected at different times, I think you might be able to gain some new clarity if you keep the thermometer variable constant.
November 26th, 2008 at 02:12
“People love to click on the links with the thermometer…but less than 10% donated after clicking-though.”
We clicked on the thermometer thinking it would lead to a page where we could get more information.
November 26th, 2008 at 12:40
I thought the templates without the meter were bugged :D
Note that Edu1 obviously is much worse than the others. Don’t you consider removing it from the rotation?
November 26th, 2008 at 14:27
Do you have a statistics for other language versions? I think it varies on each language.
November 26th, 2008 at 18:41
Anya: We are going to do a short term test of switching which banners have thermometers and see if the %donation rate also switches. It should, right?
Anthony: Agreed. We are going to work on getting the thermometer itself to be a link towards a more info page. I think the “donate now” button should take all to the donation page, but other clicks should be to a place where donors can see and read more.
lucasbfr: I thought they were bugged too, but our tech team assures me that is not the case. After we test Anya’s point above, we’ll have interesting information to go on. We did remove Edu1 from rotation…but it will be back for a short run with a thermometer so we can see how it does and what impact the thermometer has.
Yukichi: We don’t and won’t have good language statistics this year. We hope to have better tracking next year. Such is life.
Thanks for the comments.
-Rand
November 26th, 2008 at 22:22
Did the “Support Wikipedia” banners and the remixed puzzle logos have any measurable impact? Do you have any data on that? Thanks.
November 27th, 2008 at 02:15
Rand: OK. I’m looking forward to next statistics.
November 28th, 2008 at 01:36
What these data say is: Make a text “Wikipedia makes life easier – now help us make life easier for you”.
Meaning: a simple donation request will not help enough, a reciprocal request looks the most fruitful.
[http://nl.wikipedia.org/wiki/Gebruiker:Art_Unbound Art Unbound]
November 28th, 2008 at 01:40
Sorry, my response didn’t show up.
My original response was:
What these data say is: Make a text “Wikipedia makes life easier – now help us make life easier for you”.
Meaning: a simple donation request will not help enough, a reciprocal request looks the most fruitful.
December 11th, 2008 at 04:36
My donation came back. What is your address? I used 204 37th ave #330
St. Petersburg fl, 33704-1416
December 11th, 2008 at 23:13
Hi. I’m sorry to hear that your donation was returned.
Please make the check payable to ‘Wikimedia Foundation, Inc.’
Our mailing address (for checks):
Wikimedia Foundation, Inc.
P.O. Box 919227
Orlando, FL 32891-9227
USA
-Rand