Jump to content

Fundraising 2010/Report

From Meta, a Wikimedia project coordination wiki

Overview

[edit]
Daily donations

The 2010 WMF Fundraiser was the shortest and most successful campaign to date, raising more than US$16 million.

Over a period of 50 days, more than 500,000 people from 150 countries donated directly to the Wikimedia Foundation, and nearly 200,000 people donated directly to 12 local chapters. During the 2010 Fundraiser, the Foundation received its millionth donation. Our messages were shown across Wikimedia projects localized into more than 80 languages. Wikimedia community members worked together to run an efficient campaign based on constant testing and optimization of different fundraising themes, messages, images, and designs. Donors also showed their support by sending in thousands of Wikipedia stories.

There is still plenty of room for improvement in future fundraisers: we did not provide as much support to the chapters as we wanted; we plan to work more closely with the chapters and to provide them with the resources and regular data they need to conduct quality testing. In the future, we need to provide more informative and regular campaign updates so people can easily follow the progression of the fundraiser.

Key facts

[edit]
  • All amounts are in US dollars and include donations from November 1, 2010 to January 31, 2011.

Comparison between 2009 and 2010

[edit]
Metric 2009 2010
# of donations 243,688 527,583
Total amount $8,691,995 $15,026,289
Average donation $35.67 $28.48
Number of fundraising days 67 50
Best day (# of donations; total $) 2009-12-16 (13,210; $440,455) 2010-12-31 (26,106; $778,454)
$1K+ donations (# of donations; total $; av. per donor) 280; $877,923; $3,135 429; $1,179,069; $2,748
  • The total amount raised includes neither funds to be shared from the chapters nor money raised from pre-campaign testing.
  • These metrics include donations from November 1, 2010 to January 31, 2011, and from November 1, 2009 to January 31, 2010

Top countries donating directly to WMF

[edit]
Proportion of donations from the top ten donor countries
Proportions of donation amounts from the top ten donor countries

2010 saw a dramatic increase in number of donations and the amount raised in many countries from 2009. We localized donation forms by having the local currency appear by default, adjusting the suggested giving amount based on location, and translating the form into more languages than was done in previous fundraisers. In future campaigns, we want to improve and expand these practices to optimize the forms for more countries.

Country Total donations Total amount Average donation
United States 328,939 $10,041,643 $30.53
Canada 35,401 $896,483 $25.32
Japan 15,038 $489,696 $32.56
Spain 14,011 $323,422 $23.08
India 10,476 $193,485 $18.47
United Kingdom 9,098 $207,159 $22.77
Brazil 7,561 $157,258 $20.80
Belgium 5,514 $161,927 $29.37
Poland 4,694 $60,651 $12.92
Denmark 4,497 $104,566 $23.25
Other 84,757 $2,049,152 $24.17


Top five fundraising days

[edit]
Date Total donations Total amount Average donation
2010-12-31 26,106 $778,454 $29.82
2010-11-13 18,385 $521,286 $28.35
2010-11-15 16,798 $467,131 $27.81
2010-11-16 15,470 $441,113 $28.51
2010-11-14 15,154 $426,311 $28.13




Most common gift amounts

[edit]
Breakdown of frequency of top ten donor amounts.

The average donation decreased about $7 from 2009. With more people donating this year, however, the total amount raised was much higher. We extensively tested many different donation forms with different suggested amounts and saw that changing the amounts had an effect on how much people gave.

Gift amount Number of donations
$20 158,042
$5 60,381
$10 48,135
$50 46,688
$35 31,843
$100 20,762
$1.00 16,926
$19.00 16,257
$25.00 14,165
$15.00 9,728

Daily fundraiser metrics (online donations only)

[edit]
Daily click and conversion rates for the 2010 Fundraiser. Banner conversion rates are increased by a factor of 100.


How donations work
[edit]

The donation pipeline is a multi-step process that begins with a banner "impression" viewed at the top of the page hosted on Wikimedia projects. If the banner is clicked, the visitor is directed to the landing page that includes an appeal message and a space to donate by credit card or PayPal.

From this process, banner click rates are the proportion of visitors who click on banners – which directs to a landing page – after viewing a project page where the banner is displayed. Donation rates indicate the proportion of banner impressions (landing-page views) that result in a donation. Trends depicting the banner click rate and donation conversion rates are plotted in the figure. The table containing the raw form of these results, including total donations for each day of the fundraiser, is here.

On December 20, 2010 we decided to turn off banners for logged-in users, because it appeared that logged in users had already donated towards the beginning of the campaign. This was the first year fundraising banners were disabled for users to minimize disruption.



Additional facts
  • Number of check donations: 6,680, up from 2,780 in 2009
  • Number of matching gift donations (as of April 25, 2011): 550
  • Amount received from matching gift donations (as of April 25, 2011): $95,000
  • OTRS tickets closed: 8,136

Chapters' fundraising

[edit]

Wikimedia chapters are independent non-profit organizations founded to support and promote the Wikimedia projects within a specified geographical region. Like the Wikimedia Foundation, they aim to "empower and engage people around the world to collect and develop educational content under a free license or in the public domain, and to disseminate it effectively and globally". Thirty chapters currently exist in nearly every continent of the world and 12 of these chapters raised funds in their countries during the 2010 fundraiser.

WMF directs potential donors (i.e. visitors who click on fundraising banners) to payment-accepting chapters' fundraising websites by using IP address geographic lookup. To accept payments, chapters agree to a standard fundraising agreement. Payment-accepting chapters observe certain practices governing fundraising, donor privacy, and reporting; and share revenues with WMF while supporting the global mission of Wikimedia.

Chapter Donors Amount received Amount retained Staff (in 2010) Tax-deductible
Austria 2,868 $97,102.55
Australia 9,088 $343,516.97 AUD$182,050.63 0 No
Switzerland 2,311 $325,504.50 Yes
Germany 73,000 $3,026,802.00 Yes
France 12,197 $650,915.19 Yes
Hungary 1,321 $27,432.94 0 Yes
Israel 2,736 $80,021.16
Italy 22,047 $425,842.59 $232,735.69 as of 2011-12-29; i.e. 150,594.17 € given; complex taxation problems) 0 Yes
Netherlands 8,913 $428,718.49 0 Yes (granted retroactively in jan 2011)
Russia 20,879 $118,053.31
Sweden 1,937 $43,717.92 No
United Kingdom 32,329 $923,559.009 All; a £290k (~$500k) grant was given to WMF 1[1] Applied for
Total 191,653 $6,491,186.62

Note: Some numbers are provisional

Chapter Feedback

[edit]

Several chapters wrote reports and lessons learned based on their experiences in the 2010 fundraiser.

Wikimedia Australia

http://www.wikimedia.org.au/wiki/File:Fundraising Report.pdf

Wikimedia Deutschland

http://meta.wikimedia.org/wiki/Fundraising 2010/WMDE Lessons Learned

Wikimedia Israel

http://meta.wikimedia.org/wiki/Fundraising 2010/Amir's thoughts

Wikimedia Nederland

http://nl.wikimedia.org/wiki/Evaluatie Fondsenwerving 2010-2011

Community Involvement

[edit]

Early in the pre-campaign preparations, the Foundation asked members of the Wikimedia community to get deeply involved in the fundraising process. Community members responded by proposing hundreds of banner messages, translating messages, and helping with weekly testing. We were searching for a message that could perform better than the personal appeal from Jimmy Wales. During months of testing, we put a lot of effort into finding a message that could beat Jimmy and tested hundreds of ideas. While we weren't able to find that message, we still gained valuable information about what kinds of messages are not effective and learned we should focus on personal appeals. Thank you to everyone who volunteered their time and energy to suggest and test messages. These weekly tests leading up to the launch of the fundraiser were really crucial for a successful campaign. Not only did early testing help us develop quality material for a strong push right at the start, but it also ensured our functional processes were in place and that we were ready for the workload during the campaign.

We were committed to working with community members to create "the fundraiser you can edit." During the fundraiser, Wikipedia editors actually wrote their own appeals and appeared on banners. The editors worked with Foundation staff to incorporate successful elements from Jimmy’s appeal into their own stories. The involvement from community volunteers was really instrumental at this phase in the campaign and we received incredibly positive feedback from readers, donors, volunteers, chapter representatives, and WMF staff.

Wikimedia community members translated our fundraising messages into more than 80 languages, more than any previous fundraiser. 2010 was the first fundraiser when half of those languages had localized versions of the appeal letter on their projects. Volunteers worked throughout the entire campaign; translating messages from Jimmy, editors, and Sue, so that these appeals could reach people all over the world in their own language. The translators were critical in not only translating messages, but also really localizing the messages in a way that would be meaningful and relevant in each different language.


After we reached our fundraising goal on the last day of 2010, we moved onto to the the phase of the campaign aimed at recruiting new editors and publicizing the more than 400 events celebrating Wikipedia's 10th Anniversary. The Contribution Team, staff, and volunteers collaborated on ways to encourage new editors from banner messages that led to informative and instructive landing pages. The Contribution Team is still currently working to increase the number of editors and to improve overall contributions to the project. We're looking forward to working with members of our community even more on incorporating a strong contributor campaign into the next fundraiser.

Testing

[edit]

Our $16 million goal for the 2010 fundraiser was double our target from the year before. Going into the 2010 campaign, we knew that in order to raise twice as much money as the year before, we had to launch with our very best-performing material to take full advantage of the momentum and attention that our campaign would receive when we launched. Even though we didn't officially launch until about halfway through November, we started weekly testing of banners and landing pages early in August. This early testing period was crucial to optimize our designs, messages, and work processes to ensure a strong start.

The Wikimedia fundraising campaign is based on constant testing and optimization of different themes, messages, images, and donation form designs -- mostly communicated through banners placed at the top of Wikipedia pages. Only a small portion of users donate so we must always be improving our methods.

Banners

[edit]
Banner performance throughout a single day. Top banners all included an image while the bottom two banners were text only. All banners displayed the exact same text.

Banners are the first step in capturing people's attention. We had to balance capturing people's attention with ensuring that we were capturing the right kind of attention. It was really important to get people to click on our banners, but we also needed to attract people who actually have an interest in supporting Wikimedia. These "productive clicks" actually lead to donations.

During the pre-campaign testing period, we started with text-only banners but saw that graphical banners outperformed text-only by about 50%. From the 2009 fundraiser, we knew that the personal appeal from our founder, Jimmy Wales, was our best-performing material. In the months leading up to the fundraiser, we tested a wide variety of the hundreds of banner messages submitted by community members but ultimately found that launching with Jimmy would give us the big boost from the beginning that we knew we needed.

Variety of text only banners tested.

In the 2009 fundraiser, we waited until more than halfway through the campaign to introduce the Jimmy banner which reads, "Please read: A Personal Appeal from Wikipedia Founder Jimmy Wales." We were concerned about "Jimmy Fatigue". We certainly didn't expect Jimmy's message to perform at the same levels throughout the entire campaign. It was unknown how long Jimmy's appeal would serve as a compelling message, but from the pre-campaign testing we knew that we wanted to launch with Jimmy for a strong start.

In the beginning of the campaign, we tested a wide variety of different graphical treatments of the Jimmy Appeal banner with a range of colors, photos, text styles, photo placement, "Read now" button designs, etc. Ultimately, we ended up running the simple white banner with the plain button for the majority of the campaign instead of the more heavily stylized banners. During this early campaign testing period, we also learned that people need a call to action in the banners. We saw that putting a donate button on a banner hurt click-rates, but a "Read now" button (a suggestion from the community) actually increased click-rates.

Sample of Jimmy banners tested.

After about three weeks, we saw the number of donations start to decrease and it was time to move to the next phase of the campaign. We took down Jimmy banners and transitioned to banners featuring Wikipedia editors. Editors from around the world wrote about their own experiences with Wikipedia and asked readers to donate. Overall, the editor banners received similar click-rates as the Jimmy banner, but about half the donation-rate. The editor appeals were a really critical part of the campaign and WMF received very positive feedback from highlighting the editors' stories. The editors also gave Jimmy a much-needed break, so that during a later phase of the campaign he could effectively come back with a more urgent ask.

Sample of Editor banners tested.

Nearly a month into the fundraiser, our Executive Director, Sue Gardner, wrote an appeal. Similar to the editors', Sue's banner had a click-rate comparable to the Jimmy banner, but the donation rate was in between the editors and Jimmy. During this time, we introduced banners with progress meters measuring how much money we had raised and how much was left to go, but they did not improve banner click-rates at this stage in the fundraiser.

Sample of Sue banners tested.

With three weeks left to go in 2010, we reintroduced the Jimmy banner with a more urgent message and saw a boost in donations. With these banners, the progress meter did improve click-rates. Near the end of the year, we tested numerous year-end messages that emphasized tax-deductibility of donations to WMF in the United States.

Throughout the campaign, our strongest banner message was, "Please Read: An Urgent Appeal from Wikipedia Founder Jimmy Wales." In the last few days of the year we were able to find banner text that performed better than Jimmy's appeal banner. We saw two messages in particular that were performing well: "Only 4 days left in 2010 to make a tax-deductible contribution to keep Wikipedia free. Please help Wikipedia pay its bills in 2011." and "If everyone reading this donated $5 our fundraiser would end today. Please donate to keep Wikipedia free."

While each of these messages was performing well on its own, we wanted to put both of these messages into one banner. We introduced an animated banner that rotated between these two messages. At the time of the test, we thought that the rotating banner performed about 20% better than the static banner. In the post-campaign analysis, we think we might have made an error during the fundraiser and it won by less. We'll have to test this again to get more data.

We learned that we could dramatically decrease the average donation by displaying the banner with the message,"If everyone reading this donated $5 our fundraiser would end today. Please donate to keep Wikipedia free." Even though the average donation amount was much lower when we showed this banner, it brought in a much higher total. This banner may have motivated people who responded to the year-end excitement and urgency.

We were able to end the fundraiser on the last day of 2010 with our most successful single day in Wikimedia fundraising, bringing in more than USD $770,000 (the previous year's record of roughly $440,000 was actually broken four times during this fundraiser).

Sample of Year-end banners tested.
Banner Lessons
  • Graphic banners perform better.
  • Banners do not need to be heavily stylized to be effective.
  • Include a call to action ("Read now"/"Please help" button).
  • Progress meters are effective at certain stages in campaign - later when closer to goal.
  • "Jimmy appeal" message was unbeatable until urgent year-end messages.
  • Keep testing until the very end. We found effective year-end banner messages in the very last days of the campaign.


  • For more detailed analysis of testing, please see the "Testing Methodology and Reports" section.

Appeals

[edit]

While always maintaining the values and integrity of the Wikimedia movement, we also used past donor research to guide our fundraising messaging. From this past research, we learned that a significant amount of people are not aware that Wikipedia is supported by a non-profit or that it's funded by user donations. People also see Wikipedia as a practical tool they use everyday and are less compelled by messages that focus primarily on mission.

Timeline
  • During the beginning Jimmy launch phase, the most successful appeal had the first line "I got a lot of funny looks ten years ago when I started talking to people about Wikipedia."
  • In the editor phase of the campaign, we tested variations on appeals written by Kartika, Goma, lilaroja, Abbas, Victoria, Sage, and Liam. Overall, the editor appeals had similar banner click-rates but about half the donation-rate as Jimmy's appeal.
  • After the editor phase, we ran Sue's appeal for about five days which performed between the editors and Jimmy.
  • With just a couple of weeks left in 2010, Jimmy returned with a more urgent ask. While on a flight to Russia, he wrote an appeal that included a phrase describing Wikipedia as a:"temple for the mind" which performed very well and received many comments from donors. In this letter, Jimmy told users that he's a volunteer and beautifully described Wikipedia as being like "a library or a public park." These elements were powerful messages to motivate readers to give.
  • On the very last day of 2010, the "This is it. 6 hours left." appeal proved to be a strong message to end the campaign.
Appeals lessons
  • Opening line is key. It must be straightforward and compelling.
  • People really read the appeals. Text changes in the 5th paragraph affect performance.
  • Include a clear and direct ask for donations.
  • Tell people where the money goes, be specific and concrete.
  • No nonsense, no gimmicks.
  • Tell a story to clarify people's understanding of Wikipedia and that it needs them to keep the story going.
  • No ads. No agenda. No strings attached. - People respond to messages that emphasize that Wikipedia is free to use and ad-free.
  • Year-end appeals that highlighted the urgency of time and how close we were to reaching our goal were effective messages in the last days of the year.
  • Jimmy's unique voice really resonates with readers.
  • For more detailed analysis of testing, please see the "Testing Methodology and Reports" section.

Landing Pages

[edit]

There were many changes to the donation form from 2009 to 2010 based on results from weekly pre-campaign testing. The main goal was to simplify the giving process in order to remove barriers to conversion.

Landing page lessons
  • Minimize checkboxes and links.
  • Move the appeal text higher up in the form.
  • Include an animated progress meter when more than halfway to goal.
  • Use ascending ask-string radio buttons.
  • 7 radio button options performed better than 4 (in the US).
  • Include a box highlighting "Where your donation goes."
  • Smaller grey donate buttons performed better than larger blue buttons.
2009 Donation Form
2010 Donation Form
  • The form without the field for donors to leave a comment performed better than the form asking for donor comments. We'd like to find a way to bring the comments section back, possibly at a later stage in the donation process, without significantly lowering donations.
  • Please see the "Testing Methodology and Reports" section for more detailed analysis on specific tests.

Testing Methodology and Reports

[edit]

The campaign is based on constant testing and optimization of different themes, messages, images, and donation form designs. Please see the testing methodology and reports pages for an explanation of the testing and analysis process.

Throughout the fundraiser, we ran several banner and landing page tests a day. Below is the analysis on a very small sample of the hundreds of tests we conducted.

Banner Test: Hands VS Jimmy
Banner Test: Jimmy Face VS Jimmy Arms Crossed
Landing Page Test: "Where Your Donation Goes Info" Info Box
Banner Test: Static VS Fader
Landing Page Test: "Where Your Donation Goes" Pie Chart
Landing Page Test: Ask String Compare
Banner Test: Banner Message Compare
Banner Test: Animated Banner
Banner Test: Progress Meter
Banner Test: Border / No Border Compare
Banner Test: Jimmy VS Sue

Email

[edit]

Over the campaign, WMF sent two different email messages to contacts in the donor database. The first email was a message from Jimmy to kick off the fundraiser. This message was intended to gather support from past donors as well as to invite our supporters to give us their feedback on our fundraising message. The email was sent November 22-23, 2010 to all contacts who had given in past years, but who had not yet given in the 2010 fundraiser.

The second email was a message offering the option to make a monthly recurring donation. This email was sent December 15-16, 2010 to both past donors and 2010 donors. Emails sent to 2010 donors included a thank-you message and minor copy changes to reflect that they had already given this year.

Email Emails sent Emails Delivered % Bounce Unsubscribe Donations Amount Average Amount Donations per email $ per email $ per open $ per click
Email 1 213,224 194,642 8.71% 919 10,693 $303,976.89 $28.43 .055 $1.56 $5.86 $27.66
Email 2 472,147 454,024 3.84% 6,621 7,109 $599,534.39 $84.33 .016 $1.32 $4.98 $24.17

Note: Amounts are in USD

Recurring

[edit]

In response to many requests from the community, we introduced recurring giving as a donation option from December 11, 2010 through the end of the campaign. We sent out an email to past donors, and supporters who had asked for a recurring option in the past, and announced it on the Foundation and Fundraising mailing lists.

General
  • With monthly recurring giving, a donor selects the amount they wish to give, and the payment is made automatically each month, for 12 months.
  • Recurring donations are processed through PayPal, and therefore are limited to currencies PayPal accepts.
  • Donors can use either their PayPal account or a credit card to sign up for recurring.
  • 2010 Recurring Announcement email: http://lists.wikimedia.org/pipermail/foundation-l/2010-December/063131.html
Results
  • Recurring giving brought in about USD 700K in the 20 days during which it was offered (this assumes an average of ten monthly payments).
  • We emailed 454,024 people with the option to donate monthly, raising about USD 400,000 from 1,645 donors.
Recurring Testing

We tested the recurring option on several different recurring forms, including radio buttons, gray payment buttons, as well as including it on our "ways to give" page. Testing showed that including the option on the donation form actually resulted in fewer transactions. We removed the recurring giving button from the donation form but kept it as an option under "ways to give."

PayPal Recurring Information

Fundraising Updates

[edit]

WMF reported test findings and other fundraiser highlights through “Updates” on the fundraising meta pages. The updates were designed to explain what we were testing and why, the significance of the findings, and to share numbers from our data-driven campaign. In addition to the updates on meta, we also distributed several updates to mailing lists and posts on the Foundation's main blog.

Best practices and content for Meta updates
  • We aimed for a minimum of two updates per week along with a weekly review post on Fridays. At the least, two of these weekly updates were dedicated to the testing process.
  • The 'week in review' post each Friday served as a recap of the week's banner and landing page testing – and a preview for what to expect in the following week.
  • In addition to sharing what we had learned through testing, we also reported on the progress of the campaign in general, chapter updates, media coverage, donor stories and feedback.
  • We tried to ensure that the testing page was updated along with the written updates so that there were numbers to illustrate what we were reporting.
Suggestions for next year
  • In future fundraisers, we want to expand our dialogue with a larger audience by reporting updates on a daily blog.
  • We want to incorporate more regular updates from every chapter involved with the fundraiser.

Tech

[edit]

The 2010 Fundraiser brought all new requirements and challenges to the WMF Engineering team. With a large and ambitious goal set forth we decided to shift the existing fundraising team of one in house developer (Tomasz Finc) plus a contracting house to an EPM (Tomasz) and two dedicated developers (Arthur Richards and Ryan Kaldari). This was not only due to the large amount of projects requested but also because of the recommendations made after the 2009 fundraiser.

Team

[edit]

Tomasz acted as the manager, scrum master, operations support, tech lead, and deployment engineer. Arthur Richards and Ryan Kaldari joined us in the summer of 2009, bringing the team total to three. Arthur took responsibility for back end development and Kaldari worked on the front end. Through their early projects of revamping central notice, fraud fighting, and cleaning up donor databases, everyone quickly came up to speed on the existing fundraising platform and how to best work together. As the fundraiser development accelerated, the engineering group began including Zack, Philippe, and Megan from the Community Department into the daily fundraising development scrums. By working directly with our customers, we were able to finish more projects in a shorter amount of time than ever before.

Projects Completed

[edit]

Here is a small amount of the projects that we completed:

  • Recurring giving
  • Advanced fraud filtering beyond what PayPal offered
  • Dynamic donation forms
  • Single step donation forms
  • Optimized our payment pipelines
  • Newly re-architected donation cluster allowing us to scale horizontally
  • Full redesign of central notice banner system
    • Geo targeting of banners
    • Simpler interface
    • Faster banner creation
    • Streamlined back end
  • Chapter control over landing pages
  • CiviCRM upgrades
  • Fully re-written donation queue system
  • Remote log analysis for identifying fraud
  • Streamlined analytics pipelines
  • Helped to support training of non engineering developers
  • ... and many more

Agile Project Management

[edit]

We wouldn't have been able to keep up with all of these feature requests if it wasn't for adjusting the project management structure away from waterfall-like development seen at the WMF to an agile approach. The fundraiser brings all sorts of uncertainties and in order to respond to these uncertainties, we had to approach it in a very agile and iterative fashion. Through daily SCRUM's and keeping our clients close, we worked together to respond to changing requirements while maintaining a sane work load. We kept an up to date list of our backlog on meta and used this on a daily basis to organize our work, prioritize incoming requests, and keep our clients informed. We did timely code review, qa, and deployment while balancing responsiveness of project delivery. This was one of the first projects at the WMF that adopted a lot of the agile project practices and got to reap the benefits.

We've subsequently passed on a lot of our stories to other teams from our lesson learned.

One of the toughest challenges that we faced was to balance ease of changes for quick testing while maintaining strict security requirements to safeguard our donors' privacy. In order to maintain our security, we paid close attention to how our servers were configured and who had access to which elements. We limited access where necessary and imposed stringent fraud detecting features to lessen our exposure to bad transactions. This ensured that our donor data was protected from unauthorized access. We also developed new methods for quick content creation, reliability and deliverability.

In order to assess our progress, we made use of a check-in mid-fundraiser and another at the end of the fundraiser. This allowed us to connect directly with members of the fundraising team to asses our progress and course correct as necessary.

Media Coverage

[edit]

The 2010 Fundraiser received various forms of media attention throughout the different stages of the campaign ranging from news articles, blogs, and video.

November
December
January

Feedback

[edit]

Editor Feedback

[edit]

From Mayo Fuster (lilaroja):

From Sage Ross:

OTRS Responses

[edit]

This section is based on messages sent to problemsdonating(_AT_)wikimedia.org (which was an email address added to the donation form that requested feedback on the donations process) as well as general trends seen in OTRS during and after the fundraiser.

What do donors want? Top requests/complaints
  • More local bank options; especially for Brazil, India, Indonesia, China, Venezuela, and many African countries.
  • SMS donation option for non-US donors.
  • Japanese JCB credit cards donation option.
  • More options for recurring donations that don't use PayPal.
  • Donors want merchandise like mugs, globes, stickers, bags and shirts.
  • People were frustrated that their local currencies weren’t an option, particularly Brazilian Reis and Indian Rupees.
  • In several countries, donors requested that we register for charitable status for tax purposes.
  • There is strong interest in WMF participating in the Combined Federal Campaign.
  • Donors requested the option to make donations in honor or memory of someone.
  • Donations as part of a Christmas gift or birthday present.
Frustrations with technology/bug issues
  • Hundreds of people had the page time out and as a result had a duplicate donation processed.
  • Some donors wanted banners to be turned off after making a donation. Others wrote in saying they enjoyed watching the progress of the campaign, even after donating.
  • Banner loading occurred after page loaded. This caused “ghost clicking” which frustrated some users who did not intend to click on the banner.
  • Donors wanted the credit card security number not to be shown when typed and instead suggested that **** replace numbers when typed.

Donor Stories

[edit]

Below are just a few of the (roughly 100,000!) amazing stories that our users shared with us. We thank them for their willingness to tell their story about Wikipedia. You can find more stories here.

Community Feedback

[edit]

We made a public request for feedback about what we did well this year, and what we should strive to improve on for next year. Here is a summary of comments from talk pages, mailing lists, and many conversations. It is by no means conclusive, and should be evolving. Please add your feedback about the campaign so we can improve future Wikimedia fundraisers.

Banners & Jimmy

[edit]
  • Mixing up the banners between the 'hard money' Jimbo asks and the softer/brighter/warmer editor appeals was a great way to stagger the intensity.
  • Avoid Jimmy, if possible. If we are able to do it without him, that would be much better. Consider self-reference to mitigate backlash ("me again", "make me go away", "I don't want to be here either"...)
  • People will tolerate banners, but they get tired of them over time. We might want to publicize that registering an account allows getting rid of them permanently, or even give an in-between option to IPs.
  • I think that the underswell of irritation about the banners should caution against significantly expanding either the duration or the invasiveness. It's one thing to use banners, then banners with images, then bigger banners, then banners with moving graphics, then banners with moving graphics that fade into other banners... just be cautious that the increased effectiveness is not always free, and the more intense the ads eventually the bigger the backlash will be.
  • We tend to take kindly to the 'great man' theory of history. It rewards our feelings of individualism, independence, the ability to craft our own destinies, the strength of fathers, the triumph of will over the bubbling masses -- but Wikipedia is a gallant rebuttal to that whole meme. Wikipedia is the 'great community' theory of history, and yet, our product is still represented by the single face of a single person. There's some inherent positive effect from having an identifiable face, a pseudo-celebrity represent the foundation, but it also misses out on the big picture of what this community is about. I founded Us. Does it work?
  • Advertising is Advertising: Jimmy Wales should not appear in the banner.
  • Avoid putting the Wikipedia logo on other sites.

Communication & Media

[edit]
  • We need a quick and I think more prominent answer to two questions: 1) Where is this money going? (servers, staff, expenses, expansion, outreach, upgrades, innovation); and 2) Why not ad-sense? (independence, community, reader experience). These answers are out there, but I don't think readers should have to dig to get them.
  • The 5-year plan, strategic planning documents, global south, new interface, new hires... all of that stuff needs communication. Average readers see English Wikipedia, and only English Wikipedia, and they just know that they don't want it to disappear. But they don't know anything else about the scope, scale, or future of the project. That kind of simplicity is sometimes OK, but it can also create a disconnect between the perceived costs (moderate, and stable) and the actual costs (substantial, and growing). I feel like we're missing a small but important opportunity to educate people about the bigger picture in which Wikipedia is a part. Most readers don't want to know 'too much', but we could make it a little easier to peek through that crack.
  • More frequent blog posts in the early phases, followed by announcements on all mailing lists. The testing stat page was public, which was nice, but too much reliance on it might not be helpful. The issue should be consolidating and simplifying the statistics to give trends; results instead of referring people to cold hard stats.
  • IRC was done very well this year: it was a place where random people, even board members, would show up about issues, suggestions etc. Giving them an outlet and letting them feel like they have been heard really helped a lot but it can also get distracting. The structure we had towards the end with a dedicated team person overseeing the IRC all day who would ping in other team members on IM about questions and issues, and provide answers, was really efficient.
  • Regular updates were key. Better link to main stats page helpful.
  • Consider a volunteer response team who could be prepared to respond to blog/news posts, especially critical comments. This would have to be smartly done, to avoid tribe-response appearance, but could involve: answers to common questions, responses to common criticisms, and links to important pages (e.g. wikipedia:about, strategic planning doc, FAQ).

Testing & Messaging

[edit]
  • The lesson of this fundraiser, for me, was testing, as much as community. Decide if a massive crowd-sourced slogan effort is helpful, or if it creates more noise than signal... and I really liked the slogans, but to no avail.
  • Highlight new features (from the past year... vector, commons uploader, video support, WYSIWYG???)
  • If text messages aren't performing well, recognize this early on and have the community use their considerable talents elsewhere.

Team

[edit]

The fundraiser is a monstrous effort, requiring massive support from hundreds of volunteers who are supported by a very few staff members. For 2010, the staff team included:

  • Zack Exley, Chief Community Officer
  • Philippe Beaudette, Head of Fundraiser
  • Tomasz Finc, Engineering Project Manager
  • Megan Hernandez, Community Officer (responsible for massive amounts of the operation, including all testing)
  • Molly Connelly, James Alexander, Alex Zariv, Dan Rosenthal, Kelly Lyman, Salmaan Haroon, Sam Chapman, Deniz Gultekin, Josh Vandavier, Moushira Elamrawy, Keegan Peterzell, Peter Gehres, Ryan Faulkner (all Associate Community Officers, responsible for outreach, testing, email, coordinating with chapters, and execution of design)
  • Arthur Richards, Ryan Kaldari (fundraiser tasked developers)

Additional Information

[edit]