* '''In person:''' Members of the fundraising team will attend Wikimania, WikiConference North America, and other movement events for in person conversations and collaboration.
* '''In person:''' Members of the fundraising team will attend Wikimania, WikiConference North America, and other movement events for in person conversations and collaboration.
* '''Live conversations''': Virtual conversation spaces for fundraising staff and volunteers to collaborate on fundraising. Is there an existing meeting you'd like us to attend? Please let us know! To start, we've scheduled two virtual calls:
* '''Live conversations''': Virtual conversation spaces for fundraising staff and volunteers to collaborate on fundraising. Is there an existing meeting you'd like us to attend? Please let us know! To start, we've scheduled two virtual calls:
**10th of August at 1pm UTC – Meetinglink (willbeaddedsoon),
**10th of August at 1pm UTC – to attend please email Julia (jbrungs at wikimedia dot org),
**7th of September at 4:30pm UTC – Meetinglink (willbeaddedsoon).
**7th of September at 4:30pm UTC – to attend please email Julia (jbrungs at wikimedia dot org).
* '''Direct individual engagement:''' If you're interested in connecting directly, please email Julia Brungs at jbrungs at wikimedia dot org.
* '''Direct individual engagement:''' If you're interested in connecting directly, please email Julia Brungs at jbrungs at wikimedia dot org.
Collaboration for the 2023 English fundraising banner campaign is kicking off now, right from the start of the fiscal year.
This page is for en.wiki volunteers to learn about fundraising and share ideas for how we can improve the 2023 English fundraising campaign together. On this page you'll find information to increase transparency and understanding of the fundraising program, background on improvements around community collaborations that have been made since the last campaign, new spaces for collaboration, and messaging examples to invite volunteers to share ideas for how we can improve the next campaign together.
Introduction
The majority of funding for the Wikimedia Foundation comes from individual donors all around the world. These donations allow us to provide the world-class technology infrastructure that supports 20 billion monthly views to Wikipedia and its sister projects, protect free knowledge globally through legal and advocacy efforts, and support the incredible volunteer editors that have built 61 million articles across more than 300 languages. This year, the Foundation is focusing heavily on improvements to our products and technology, particularly the needs of experienced editors. You can learn more about priorities for the Wikimedia Foundation in detail in the FY 2023-2024 annual plan.
To fund these efforts, the fundraising team will run its annual Q2 English fundraising campaign (for non-logged in users) in Australia, Canada, Ireland, New Zealand, the United Kingdom, and the United States. Funds raised from these countries account for more than 50 percent of all funds per year and this is an important moment to invite readers to support Wikimedia's mission. To prepare for the campaign, the fundraising team will continue the yearly practice of running limited "pre-tests" between July and November, to ensure optimal systems and banners, in collaboration with volunteers.
As a brief recap, in December 2022, the English Wikipedia community ran a Request for Comment that underscored the importance of the Foundation's fundraising team working closely with the movement on banner messaging. The team kicked off a collaboration process that resulted in the campaign featuring more than 400 banners that came from the co-creation process with volunteers. The revenue performance of the banners declined significantly last year and resulted in a longer campaign with readers seeing more banners than previous years. The fundraising team learned a lot through the collaboration process and is eager this year to build on this work with volunteers to develop content that will successfully invite donors to support our mission. We aim to reach fundraising targets in ways that minimize the number of banners shown to limit disruption and resonate with readers and volunteers. You can read more background on last year's campaign, in the background section.
Collaborating on messaging with volunteer stakeholders is key to the fundraising team. We will use the co-created banner message that ran in December 2022 to kick off the pre-tests and work together with volunteers on new ideas for this year's campaign.
Background
Background on collaboration in December 2022
Background on collaboration in December 2022
As 2022 came to a close, a Request for Comment (RfC) on English Wikipedia proposed changes to the messaging of year-end fundraising banners. The Wikimedia Foundation accepted the guidance provided by the close of the RfC, and established a co-creation page to seek volunteer input on banner messaging from community members. Throughout the fundraising campaign, the team posted regular updates to the page.
Thank you to everyone who participated in the banner co-creation process in 2022. In the January campaign recap, we shared more information on how we ran a different kind of campaign with new banner messages that were created together with Wikimedia volunteers.
The English campaign raised significantly less than the previous year. Overall, we saw a $10 million decline in banner fundraising year over year despite showing more banners to our readers. There were tradeoffs to running lower-performing messages in 2022. Over the years, the team has worked to limit the disruption to the reading experience by making the campaigns more efficient, reducing the number of days the banner campaign ran and limiting the number of banners readers see. In December 2022, we ran a longer campaign and showed 49% more banner impressions than the previous year. Despite these changes, we still raised 30% less than the previous year. While we made gains through our work together, the donation rate was still significantly less than the previous year.
Thanks to constructive conversations and commitment from community stakeholders, we made considerable gains in banner performance throughout the 2022 English campaign, following the initial low banner performance at the start of the campaign. Here are some highlights from co-created banners:
A new 2022 founder appeal direct from Jimmy Wales
The theme of reciprocity
More clarity around the role of the Foundation and the importance of donating
Time-sensitivity to highlight the importance of this fundraising moment
Adjustments to the design and usability of close options in our mobile banners
450+ banners tested
A closer relationship with our volunteers who are interested in fundraising messaging
The team is committed to building on the progress made throughout December to improve this year's campaign together. We're excited to kick off the collaboration much earlier, in our typical Q1 pre-testing period, so that we have time to test and optimize, along with volunteers, before the end of year push.
Background on Foundation planning and transparency in 2023
Background on Foundation planning and transparency in 2023
We recognize that there were concerns about matters outside of fundraising banners expressed in the 2022 RfC and elsewhere. Here are a few brief highlights to share on our wider approach to planning and transparency in 2023. To learn about priorities for the Wikimedia Foundation in detail, please see the FY 2023–2024 annual plan.
Sharing more detailed information about how funds are spent.
This year, the Foundation's annual plan offers detailed budget breakdowns, including across our four goals for this year, of which our technical infrastructure is the largest (48.7%). We also shared more information about how specifically the Wikimedia Foundation supports the Wikimedia projects this year, including:
23% of our budget ($39.7 million, about 26% of staff) to evolve and maintain features and functionality of the projects, such as developing new software, adapting the sites to new form functions, site security, and maintaining our servers and tech stack
26% of our budget ($46.4 million, about 30% of staff) to build analytics and machine learning services that support contributors, create and deliver analytics, and enable data driven decision making through our APIs and metadata systems.
20% of our budget ($35.1 million, about 12% of staff) to supporting affiliates and volunteer communities around the world
9% of our budget ($16.6 million, about 10% of staff) on legal and public advocacy, such as fighting censorship, defending trademarks, defending threats to free knowledge, safeguarding Wikimedia's reputation, and ensuring legal compliance
$17.9 million of our budget split between fundraising, human resources, financial, and legal support (12% of our budget combined and about 13% of staff.)
Slowing of Foundation growth.
The Foundation's annual plan budget is flattening compared to prior years. As discussed in the annual plan, we have made reductions in both non-personnel and personnel expenses to achieve this.
Engaging more closely with existing editors, and prioritizing their needs in our plan for this year.
On the 14th of April, Chief Product and Technology Officer Selena Deckelmann released her listening tour letter after her first nine months in her new role. It recognized the need for more attention to longstanding challenges like technical debt, the importance of improving the Foundation's relationship with English Wikipedia, and overall, having a clearer vision for how to responsibly use donor-funded resources to support Wikipedia and the sister projects. The annual plan for fiscal year 2023-2024 follows through on a number of these things, where work has already begun. Specifically within the Technology and Product goals, you will see a focus on working closely with editors and editors with extended rights (admins, stewards, patrollers, and moderators of all kinds, also known as functionaries) to improve their experiences. The Foundation will also work on providing decision makers from across the Wikimedia Movement access to reliable, relevant, and timely data, models, insights, and tools that can help them assess the impact (both realized and potential) of their work and the work of their communities, enabling them to make better strategic decisions. Finally, exploring strategies for expanding beyond our existing audiences of consumers and contributors, in an effort to truly reach everyone in the world.
Sharing more information about financial and operational management.
This year, we've taken the extra step of publishing salary data for our CEO Maryana Iskander and Chief of Product & Technology Selena Deckelmann in our annual plan, to supplement the executive salary data released each year in our form 990.
Building on collaboration so far in 2023
Building on collaboration so far in 2023
Since the English campaign in late 2022, the fundraising team and local volunteers have made a lot of improvements in the collaboration process for campaigns so far in 2023 in Sweden, Japan, the Czech Republic, Mexico and Brazil.
A few highlights of approaches we've tried since January to improve the fundraising collaboration process:
Partnered with 4 affiliates (affiliates were involved in the creation of the community collaboration pages, gave input on banner and email messaging, and attended community calls)
Direct individual engagement with local volunteers
A few key takeaways from the campaigns so far in 2023:
The timing of early collaboration ahead of the campaign is critical to ensuring a strong campaign launch.
Continued partnership with local affiliates is valuable to ensure quality localization and open up engagement spaces with volunteers.
Minimal messaging concerns raised from most banner campaigns.
Minimal interest/attendance in fundraising dedicated meetings, joining existing spaces is valuable to increase participation.
Messaging ideas from Mexico's community conversation were implemented in banners.
Japan campaign had 5 participants who collaborated on an improved translation of Jimmy's new December message.
Local payment methods were discussed on the Swedish collaboration page.
Czech Republic volunteers asked questions to understand the campaign (It was the first time we ran a fundraising campaign in the Czech Republic).
Collaboration spaces
To increase information sharing and collaboration opportunities for the next campaign, the team is eager to engage in multiple spaces. Here are a few initial ideas. We welcome more input on channels for collaboration!
On wiki: Right here on this en.wiki collaboration page, fundraising Meta page, English Wikipedia Miscellaneous Village Pump. The team will share campaign insights, plans, and updates on this collaboration page. Updates will include message ideas for input, summaries on banner testing, changes to messaging over time, and space for new ideas and questions from volunteers. While we won't be able to test every single message idea shared here, we will build from the process last year to continue to try ideas shared in this collaboration space as well as other new spaces we're setting up this year.
In person: Members of the fundraising team will attend Wikimania, WikiConference North America, and other movement events for in person conversations and collaboration.
Live conversations: Virtual conversation spaces for fundraising staff and volunteers to collaborate on fundraising. Is there an existing meeting you'd like us to attend? Please let us know! To start, we've scheduled two virtual calls:
10th of August at 1pm UTC – to attend please email Julia (jbrungs at wikimedia dot org),
7th of September at 4:30pm UTC – to attend please email Julia (jbrungs at wikimedia dot org).
Direct individual engagement: If you're interested in connecting directly, please email Julia Brungs at jbrungs at wikimedia dot org.
Many ideas shared by volunteers on the English campaign co-creation page last year were incorporated into banners, such as messages around the theme of reciprocity, clarity on the role of the Foundation, and a new appeal from Wikipedia Founder, Jimmy Wales. We're eager to build from that process to work together to find ways to improve the banners this year. Throughout the first quarter of the fiscal year (July-September), the team runs limited pre-campaign tests ahead of the fundraising campaign. Similar to last year, it won't be feasible to test every message idea shared on this page as there may be more iterations than space in the "pre-test" period, but we will continue to try ideas shared in this collaboration space as well as other new spaces we're setting up this year.
As is regular practice for the fundraising team, the first tests of the new fiscal year in July will be technical systems and payments tests. For these tests, we'll use the same banner message that was co-created in December. No new language will be introduced in the first tests.
Starting in August, we'll kick off messaging testing. We welcome your ideas! There were many messages tested in December and we'd like to revisit some of these to try different combinations of the variety of themes explored last year to reach our fundraising targets in ways that limit the disruption of campaigns and resonate with our community of volunteers and readers.
Co-created banner message from 2022 (with light edits to remove the "year-end" message):
Wikipedia is not for sale.
A personal appeal from Jimmy Wales
Please don't scroll past this 1-minute read. This Wednesday, June 14, I humbly ask you to reflect on the number of times you visited Wikipedia in the past year, the value you got from it, and whether you're able to give $3 back. If you can, please join the 2% of readers who give. If everyone reading this right now gave just $3, we'd hit our goal in a couple of hours. $3 is all I ask.
When I set up the Wikimedia Foundation as a nonprofit to host Wikipedia and 12 other free knowledge projects, it meant that we could preserve our core values: neutral, high quality information, not outrage and clickbait. Being a nonprofit means there is no danger that someone will buy Wikipedia and turn it into their personal playground.
If Wikipedia has given you $3 worth of knowledge this year, please donate now, it really matters. Thank you for your generosity!
There are also many powerful lines that emerged from the co-creation page in December, and we hope to iterate and repeat some of these messages. Here is a sample and we'd love your input into what you like or would want improved.
We ask for much less than many other nonprofits: just $3, or whatever is available to you. The reason we ask for these small donations is because Wikipedia and its sister sites are owned and built by everyone. There are no small contributions: every edit counts, every donation counts for our nonprofit. We're proud of the work we do.
Wikipedia is different. No advertising, no subscription fees, no paywalls. Those don't belong here. Instead, the Wikimedia Foundation relies on readers to support the technology that makes Wikipedia and our other projects possible.
Now is the time we ask: If you donate just $3, or whatever seems right, we could expand the reach of free knowledge and keep improving the technology behind Wikipedia. There have never been ads or subscription fees on Wikipedia. It welcomes everyone, like a library or a public park.
We don't run ads and we don't sell your data, because you're a community to us, not a commodity. Instead, our nonprofit relies on readers for support. Wikipedia is part of the infrastructure of the world today, like a library or a public park where we can all go to learn.
Give only what you can comfortably give: what matters is your support, not the size of your gift. Together, let's support this special space on the internet, with no advertising, no subscription fees, and no paywalls. Wikipedia welcomes everyone. And that's priceless.
One donation may seem small, but when millions of readers each give, we can do great things.
Add your ideas here!
Please share your ideas here! These can be iterations on the message above, new sentences, inspiring words, themes, or new concepts to try. We'd love to use this space to plan out the first message tests of the year together. Thank you for any ideas you'd like to share!
Thank you for this suggestion, it is certainly a topic of great interest within WMF, the Wikimedia Movement and across the web. Thanks too, @Nosebagbear, for your reflections. We’ve tried similar themes in the past and it feels like an area to continue to explore together.
Here is a message idea we’ll workshop further and want to share here for early reactions. Feedback welcome; @Folly Mox and @Nosebagbear, what do you think?
As a non-profit, we are passionate about our model because at its core, Wikipedia belongs to you. We want everyone to have equal access to quality information - something that is becoming more rare online/harder to come by online/harder to recognize online.
I like this idea, and have a suggestion that's a little different in tone:
It's hard to know what to trust online these days. Disinformation, grifters, and scammers are everywhere. Wikipedia is different. It's not perfect, but it's not here to make a profit or to push a particular perspective. It's written by everyone, together, for no other reason than that they want to help create a free repository of quality information. That's something we all need, like a library or a public park. Please consider donating $2 to support our work today. —Ganesha811 (talk) 17:12, 21 July 2023 (UTC)[reply]
I like this because it challenges the popular conception that Wikipedia "moderators" are paid, hired or "vetted" by some authority. Not a criticism of your wording but I like using the phrase "Wikipedia is written by its readers". — Bilorv (talk) 09:01, 22 July 2023 (UTC)[reply]
User:SPatton (WMF), this is a genuine question: is this draft idea related to the disinformation / generative AI theme? No big deal if it's a segue into a different topic – I'm not expecting my suggestion to be workshopped immediately – but it seems non sequitur. Folly Mox (talk) 17:25, 21 July 2023 (UTC)[reply]
I was thinking more in the sphere of AI is spooky and unknowable. AI learns a lot of its information from Wikipedia, so a better Wikipedia means smarter, safer, and more accurate AI. Worse AI sends police to the wrong address and they shoot your dog. But, like, with good words and not over the top paranoid. Folly Mox (talk) 17:37, 21 July 2023 (UTC)[reply]
Hi @Folly Mox, yes! My response was related but obviously the language I suggested was less direct. We appreciate the nudges to talk about Wikipedia’s powerful role in the large language model space, and just want to make sure we accurately portray Wikimedia Foundation’s work on it.
Thanks for your suggestions too @Ganesha811, I love a lot of your imagery and tone. We will line up some of these ideas for testing in August. I’ll follow up next week with a sample of AI / misinformation messages for y’all to look at. SPatton (WMF) (talk) 18:45, 28 July 2023 (UTC)[reply]
User:SPuri-WMF Thank you Ganesha. Lots of great ideas here, I like the opening line a lot. Also agree with @Bilrov that it is great to highlight that the content is written by everyone/Wikipedia readers. We will line up some of these ideas for testing in the coming weeks and will follow up with some updates here. 20:25, 27 July 2023 (UTC)[reply]
Thanks for starting the conversation early! I like almost all of the team's suggestions above and appreciate your willingness to take on board the community's feedback, even though we raised less money last year than most years. One quibble: I would cut the word "humbly" wherever it appears; self-aware humility, ironically, has the opposite effect. —Ganesha811 (talk) 01:04, 17 July 2023 (UTC)[reply]
Hi @Ganesha811, thank you very much for the input and encouraging words. Would it surprise you to learn that we tested that concept of ‘humbly’ once again, head to head, in December, and found that including it was a statistically significant ~ 11% increase in donations?
That’s an outsized impact for a single adverb, but every so often, we hit upon a particular ‘power word’ that resonates with readers. Interestingly, when that happens, it sometimes gets picked up in the language donors reflect back at us; such as this comment sent in from a donor:
I noticed it, I read it and what caught my attention is the word “humbly”. To know that your organization is just trying to provide a free service accessible to all and to ask humbly is what made me want to give. I’m not being beaten down to donate, I was just asked politely, respectfully and humbly.
But I’m not just here to post old test results, we’re here to chart a course together. I will queue up a fresh test of humbly / no humbly in early August, and bring the results back to this page. We’ll also gather more donor feedback on this wording and share that back here to inform the discussion.
Is there other wording you’d be interested in exploring here as we’re brainstorming new messaging tests? E.g. I’m honored to ask you, or I have the pleasure to ask you …
I find that very interesting; clearly, not everyone finds it offputting! Thanks for sharing the data, and for running a fresh test. One word that you could try is "grateful"; We are so grateful that people choose to support this work... Or, if an adjective is simpler to slot in directly, "sincerely". As I said, this is a quibble, not a dealbreaker, so if you end up deciding that "humbly" should stay, it wouldn't be a cause for alarm. —Ganesha811 (talk) 17:06, 21 July 2023 (UTC)[reply]
I've always advocated for communicating the feeling of gratitude in corporate messaging, instead of pride / humility / honour. Even if it doesn't mesh well with the point of the message ("grateful to ask for your donations again this year"), gratitude is a universally positive energy that doesn't place the parties on an uneven level. Folly Mox (talk) 17:29, 21 July 2023 (UTC)[reply]
I appreciate there was some movement by the WMF last year in response to volunteer outrage. I oppose banners because I don't approve of how WMF funds are being spent. But, if banners are to exist, then they should be sparingly shown, unobtrusive, factually accurate and avoid crisis messaging about running out of money, regardless of what effects this has on revenue. My primary concern is making sure Thomas does not donate what he cannot afford. Wikipedia does have an existential crisis, but one of volunteer labour, not money. This crisis is mostly unacknowledged, while our fundraising model is known to most readers. I've proposed it before, but this is what I would like a banner to look like:
Wikipedia is facing a crisis: a crisis of volunteer shortage. A typical reader views [x] articles and spends [y] hours per year on Wikipedia. Consider giving just one hour of your time signing up with an account and clicking on [link e.g. user homepage for SuggestedEdits tasks] to help us fix the errors in our articles and give back to the website that wants to make the sum of all human knowledge freely accessible. You can also give financial donations to the Wikimedia Foundation, which runs Wikipedia and other educational websites, by clicking [here]. Read our [donation FAQ] for more.
While I wholeheartedly agree with the sentiment that we should try to recruit more volunteers, it should be noted that the Foundation has previously tried recruiting editors via banners, starting way back in 2010. Now that we have a better newcomer task suggestions system and mentorship systems for newcomers, I'd be curious if they worked any better. Part of the issue that people often sign up to edit because they're interested in a particular subject, rather than generically interested in editing, and these kind of banners and tasks aren't very good for that. Anyway it's worth a try. Steven Walling • talk05:32, 20 July 2023 (UTC)[reply]
Thank you for your ideas about how fundraising could be an opportunity to encourage more people to edit. Over the past year, the fundraising and Product teams have been experimenting with ways to do just that. After readers make a donation, they land on our “Thank You” page, which thanks them for their donations. The team ran an experiment in the fundraising campaigns in Latin America, India, and South Africa where we added a call-to-action on the “Thank You” page to create an account and start editing. Here’s an example of the invitation to edit on the donor thank you page. Since the initial experiments in LATAM, India, and South Africa, we have rolled out this invitation to edit on the thank you page out to donors in those countries, as well as France, Italy, Japan, Netherlands and Sweden! You can read more about the experiments here. Increasing editors impacts the existing volunteer community so we’re excited to have this conversation together here. The invitation to edit has not yet been rolled out to the English campaign, but the team is certainly interested in exploring that idea. We’d love to hear from volunteers on the idea of experimenting with including this CTA to the thank you page, potentially in a few brief “pre-tests” as a starting point to learn about the impact of inviting donors to edit. What do you think?
Absolutely, I'm in favour of encouraging donors to volunteer in this way. Please roll out the new donation thank-you page when you can. However, I don't think volunteering is an add-on, secondary to the goal of getting readers to donate, and just one step above filling out a survey on donor demographics. I think it's (like finances) critical to the continued existence of Wikimedia and (unlike finances) far below the level it needs to be at. — Bilorv (talk) 16:54, 21 July 2023 (UTC)[reply]
That is, alas, very clearly only half true. Both are needed - and the use of banners (and reader and community tolerance thereof) is a limited resource. The tests should try both, but this can be a case where being okay at both is worse than good at one. We know we can raise money through banners - acquiring volunteers who stay for more than 1-2 edits through banners is less confirmed. Nosebagbear (talk) 16:35, 24 July 2023 (UTC)[reply]
That's as may be, but consider that recruiting new editors is also fundraising, in a very real sense. If I tried to calculate the value of the time that I have spent on Wikipedia, I suspect it would run easily into the hundreds of thousands of dollars worth, if not more, depending on the exact calculation method. That's far more than I would have ever directly donated to it. We need volunteers much more than we need cash. Money doesn't do NPP or AFC or check for copyright violations or do any of those things we actually need involved volunteers to do. SeraphimbladeTalk to me23:33, 24 July 2023 (UTC)[reply]
Hi all, I just wanted to chime into this interesting discussion with the factoid that we have already worked in a small mention of editing to a banner from last year’s campaign:
There are no small contributions: every edit counts, every donation counts.
But I’d be curious to explore this theme further. Here’s some sample language we’ve developed, what do y’all think?
Wikipedia relies on volunteers: If Wikipedia and its sister sites are useful to you, support us with a donation or support us with an edit.
From the beginning, Wikipedia has been a volunteer effort. The articles you read are written by volunteers, and the non-profit that supports our projects relies on voluntary donations. Every edit counts, every donation counts.
It will always be challenging to position two calls to action in a single appeal; but we know that most readers don’t actually end up making a donation, and it would be gratifying to get more from our banners and drive actions we need for our projects to thrive.
The lines above are generally pretty good, doing a better job at indicating avoidance of funding bias rather than that more money would somehow improve the content quality directly. I don't oppose the fundamental nature of there being fundraising banners, but looking at the other aspects noted above, I'd concur with Ganesha's point. Unless increased donations not merely could but are more likely than not to improve on disinformation/LLMs, I wouldn't include any mention of them. While a day or two of editor-focused testing might give some interesting data, I wouldn't support it being looped into the actual campaign beyond the level we've seen in the past. Nosebagbear (talk) 10:17, 18 July 2023 (UTC)[reply]
Hi @SPatton (WMF) and MeganHernandez (WMF): -- More experiments like this would be fantastic. To be comparable, they should link directly to an edit funnel (account creation --> tutorial?) which deserves A/B testing for % that complete the funnel / make an account / make a first edit, like our donation funnels. We could also use a way to flag new accounts that come in via this route, so that standard mod tools can tag them appropriately (to distinguish them from other sources of new accounts which may have other cohort dynamics / may be more likely to be brigades or coordinated spam). And the sorts of editors responding here in support of such banners (myself included) may be glad to sort recentchanges by such a flag and give extra attention to those editors, to see what we can learn. – SJ +14:07, 28 July 2023 (UTC)[reply]
Hello @Sj, thank you for your enthusiasm about the idea! I’m fully in support of closely examining the edit funnel and impact of a Thank you page experiment. We looked at the edit funnel last year, when the Growth team helped with this Thank you page experiment, and the Growth team has already committed to helping with similar analysis this year if the community is interested in testing this on English Wikipedia.
Approximately 7% of donors in these markets showed interest in editing immediately after donating, based on the estimated click-through rate from the Thank you page.
Great to see the nubmers. On the 7%: "Try editing" is a screen down, and the button to click further down still. I'd also be surprised if "try it, it's easy!" is a better hook than "jump in, it's fun", "we're always looking for new contributors", "we're created by editors like you", or just an assortment of potential topics "help write or share photos about astronomy, botany, history [search for a topic]". would love to see tuning experiments. A message that everyone sees in a banner would also avoid the dropoff of requiring people to have a credit card and donate money first.
The landing page achieved a 45.1% account creation rate, which is a promising result compared to other channels.
Perhaps that's an avoidable dropoff. what happens if you immediatly show people a "suggested edits" carousel and invite them to edit? (with a link to a mentor, collapsed into a single "ask for help" line, right on that screen) Have them make an edit or two before prompting for account creation. [many ways to gloss this, with a pseudo account or not, but MW is designed to work with no access-control at all!]
4.6% of the accounts created right after donating started editing within 24 hours of their creation.
I'd think the metric would be "made an edit within 2 minutes", and working on a variation on the first-edit funnel that gets close to 100%. I'm thinking "voting on POTY" level of ease and immediate gratification. – SJ +23:23, 3 August 2023 (UTC)[reply]
The percentage of people who funneled from donation to editing was lower than we originally hoped, but I think it’s understandable that people who create accounts after donating are less motivated to edit compared to people who create accounts organically. Do you agree, or are you concerned by any metrics in this funnel?
Although we didn’t conduct an A/B test for this specific experiment, the Growth team releases most of our features as part of A/B tests to properly measure the impact. I'll explore the possibility of conducting an A/B test if the community is interested in testing this on English Wikipedia. Additionally, I'll investigate the feasibility of flagging these accounts for patrollers and advanced editors while being mindful of privacy concerns.
Besides A/B testing and flagging accounts, are there other improvements you would like to suggest for this experiment? I’ve started a discussion on the associated experiment page: Encouraging donors to edit (Thank you page experiment) If you are interested in this experiment, please join the discussion.
As a reminder, the Foundation does experiment with ways of recruiting new volunteers to the movement throughout the year, the donor Thank you page is just one potential way to bring in new editors. You can learn more about current work onboarding new volunteers on the Growth team page, or subscribe to our newsletter. Thanks! - KStoller-WMF (talk) 22:45, 3 August 2023 (UTC)[reply]
Appreciate all of this, thakn you KStoller. I'll comment on the New-experience page you linked. I thought we were brainstorming a banner campaign that would be separate from any thank-you page campaigns, but both are important and would surely have overlaps. (Very different average audience taking the first step, however). The nice thing about a family of topical banners is that we can develop messaging w/ relevant wikiprojects, reach out to people who care about the topic, and now have a well-targeted set of open tasks, and community waiting to engage w/ those interested. Photographers, foodies, local historians, gardeners -- every hobbyist group -- likely has a couple of related wikiprojects and, while a generic "come and edit!" might feel like a slight lift only if they have time, an invitation to indulge a hobby and meet fellow enthusiasts would be on the other side of the pendulum: a draw even when they have other things competing for interest. I don't know if we've tried to measure how much a given banner campaign burnishes our overall reputation + association with ways people can give back to their communities of practice, rather than drawing down on the site-wide reputation, but this would ideally become a strong source of the former. – SJ +23:28, 3 August 2023 (UTC)[reply]
Thanks for the feedback, SJ! This thread should remain banner focused, so I'll follow up more about the thank you page ideas here: Talk:Growth/Newcomer_experience_projects
Despite over 50 million banner impressions, the banner only led to 492 "constructive activations" (a constructive activation is defined a new account editing within 24 hours of registration and that edit not being reverted within 48 hours). We A/B tested a few changes to the banner language to improve click-through rate, with limited impact. Ultimately we concluded that the Growth team should utilize our time testing other methods for new editor recruitment and continue our efforts to ensure we engage and retain new editors.
That being said, I think you are spot-on about readers being more motivated to edit if we connect them to a topic or project that interests them. We haven’t explored this idea as it relates to recruitment and banners, but the Newcomer homepage provides suggested edits that are targeted based on the account holder’s interests.
Two potential messages for banners encouraging Wikipedia readers to edit. Appreciate these details! That feels a bit clinical for a welcoming invite or call to arms. I agree, looking at the banners you mention, that might not be an effective way of recruiting. ;) The banner had 50+ words and no image, wasn't the sort of pretty banner that makes people glad to see it at the top of a page (minor + offsetting the minor nuisance of banner and lost vertical space), and didn't have a motivating call to action. Of the images in that presentation, the top of the landing page (L) w/ a button to click might make a more compelling banner message than the block of text (R).
However even in that example, you found modest A/B tests that showed 50%-80% improvements. So I expect there are a couple magnitudes of improvement possible with compounding iterations, and another magnitude from simplifying "click->make account->land on homepage->try to edit" to "click->try to edit". I hope we can make it possible for communities to effectively run these sorts of A/B tests on small-cohort banners for readers. A fixed team's time is limited, but community groups have separate time constraints, and drive to generate enthusiastic messages that resonate with fellow enthusiasts. Warmly – SJ +00:36, 8 August 2023 (UTC)[reply]
User:Edward-Woodrow please, that's not constructive. Whatever your feelings on salaries / pensions / etc., the Foundation is here in good faith trying to meet us halfway. Let's not sabotage the working relationship with comments like this. Folly Mox (talk) 17:11, 21 July 2023 (UTC)[reply]
The WMF does not need eight or nine digit figures to maintain Wikipedia and its sister projects. The vast majority of these funds are not essential or even particularly beneficial for Wikipedia or its readers. If readers part with their money because they're led to believe that it's going to support Wikipedia when it's not, then that's not a fundraiser. That's a scam. Thebiguglyalien (talk) 03:55, 25 July 2023 (UTC)[reply]
@Edward-Woodrow, Thebiguglyalien: It's worth noting that the Foundation has just announced the second round of grantees for the m:Knowledge Equity Fund. That's about $1 million flowing to organisations outside the Wikimedia universe, without broad community oversight. One might add the tens of millions funneled to the Wikimedia Endowment at the Tides Foundation, which is quite opaque and has never yet published audited statements detailing the Endowment's revenue and expenses. AndreasJN46615:56, 4 August 2023 (UTC)[reply]
This is not a suggestion for a message, but a comment on the fundraising results. It appears that online donations to the Foundation are down. (I searched for data providing just how much income the Foundation receives from these appeals, vs. soliciting for large donations, but failed. All I have to go by is the comment above that these donations are down by $10 million or 30% from the previous year.) However, it's not clear whether these donations are down because of the change in the language of the banners, the economy, or some other cause such as an organized campaign against Wikipedia or the Foundation. Nevertheless, if the donations are down because these banners are using language more in line with our suggestions -- which might be the case if this cooperation continues for this year -- then speaking for myself I can live with that. Better that these banners reflect our values & get less money, than misrepresent them in order to get more.And please remember, we who are contributing to Wikipedia, Commons, Wikisource, & all of the projects -- creating the content people come to these websites to use, thus creating the funding the Foundation depends on -- we are not paid. In some cases, we actually have to pay out of our own pockets for the material needed to create that content, so it could be said that we are paying to contribute. (And yes, I know there are people paid to make edits. This is not encouraged by the community; in many cases these paid edits create more problems for us, so they are a negative.) -- llywrch (talk) 19:00, 25 July 2023 (UTC)[reply]
Extended content
Major Changes Needed Wikipedia has a serious reputation problem with a massive segment of the population. Many of the current affairs and public figure pages use highly biased secondary sources, which, in turn, leads to highly biased Wikipedia pages. Because of the almost universal bias in the secondary sources and academic fields, this bias has propagated through to create an equivalent bias on Wikipedia. At this point I can hear many of you shouting “no,” or “that’s not true,” at your screens right now, but you need to remember - perception is reality. I know from personal experience how frustrating it is to try to improve a Wikipedia entry only to get brigaded by established editors who mass-revert any changes to make sure that no one person reverts more than three times, enabling them to circumvent the rule on edit warring. The result being that a massive segment of the population feels that Wikipedia is horrible at best, and complete trash at worst. This ties into the fundraising issue because if any of you think that someone from that 50% of the population is going to contribute to Wikipedia as long as this continues, you’re fooling yourselves. The point being that Wikipedia needs to take a step back and review every aspect of how it develops its content, its source requirements, and find a way for Wikipedia to deliver accurate and completely factual information without ANY interpretation of the the information at all. It might make the material more “dry”, however it would then leave it up to the reader to decide how they want to interpret the information without spoon-feeding an ideology to them. Steven Britton (talk) 23:56, 25 July 2023 (UTC)[reply]
@Scbritton: I am highly sceptical that any source amending process would manage to, say, halve the % of the population that thinks our current affairs/public figure articles are biased. Accusations of our varying biases are frequently made with reference to fully known statements without any of that pesky interpretation.
But beyond that suspicion, "Wikipedia" doesn't review its policies - the editors that make it up do. So if you think our sourcing rules are inherently flawed, we can't just will them into being - you need to start a discussion over at VPI to find improvements that don't cause more issues than they solve.
More relevantly, this isn't really the place - the WMF has no control over non-legal content policy so the only bit that could be relevant to them would be the area of "should we target our fundraising banners to those inclined to trust us, and if so, how?". To some degree I guess we do that already - those who think we are liars all, probably don't think we've given them $3 worth of value. Nosebagbear (talk) 15:26, 26 July 2023 (UTC)[reply]
Sadly, based on what've you've said here, it sounds to me like Wikipedia is, and will remain afflicted with Groupthink for the foreseeable future. Note that this isn't an attack against you personally; I'm just stating that I highly doubt any discussion I begin will be met with anything other than hostility at this point. Steven Britton (talk) 02:31, 27 July 2023 (UTC)[reply]
Pcoombe (WMF), your preamble for this discussion said "Starting in August, we'll kick off messaging testing. " But it appears that tonight, July 28th, you have got this fundraising exercise underway? I just had a Wikipedia query (unlogged-in, using en.m.wikipedia on Android) displaced by the full page Appeal overwrite. AllyD (talk) 19:56, 28 July 2023 (UTC)[reply]
Hi @AllyD, in the paragraph right above the one you quote, we mention that we’ll start “technical and payments tests” in July. That’s what we’re running this weekend: a “low-level” traffic test of our payment forms, showing to 5% of English Wikipedia readers.
We haven’t done any messaging testing as promised; we will start that next week with some of the suggestions that have been posted here.
I had interpreted the technical and payments tests to be technical infrastructure testing, possibly involving some volunteer users; a productionised broadcast appeal to 5% of the site users over a weekend is something which I would regard as a live fundraiser, and will certainly be perceived as such by all these site users. AllyD (talk) 06:04, 29 July 2023 (UTC)[reply]
@AllyD I've had several of those too over the past week on my Kindle, and if you look at the daily fundraising data here:
they may have brought in significant amounts of money, too – several days in the second half of July registered more than half a million dollars each.
Note however that per m:Fundraising the French email campaigns started on July 19th, and the Indian email campaign started on July 18th. So some of these July takings will be due to those (the WMF could probably tell us the split).
For reference, the Indian sample emails provided on Meta look like this:
Emails 1-3 seem fine, email 4 is a bit odd unless we (or actually Jimmy, but I assume he's going with what Advancement thinks!) genuinely didn't/don't think the target wouldn't be met. Most of email 4 is actually really good, and even that bit is hardly the end of the world. Nosebagbear (talk) 13:04, 3 August 2023 (UTC)[reply]
Hi @AllyD and @Andreas, I’m happy to provide some more details on the banners that you saw. For starters, I’ll clarify that these short tests (which we internally call “pre-tests,” though I understand that’s jargony) are a normal course of action for us. Every year, we run tests of different lengths and durations in the lead up to the end of year English Wikipedia campaign.
We are particularly excited this year to devote many of these pre-tests to workshopping and adopting suggestions that come from this collaboration page. Generally these tests are useful for a couple reasons:
We do technical work throughout the year to add support for new payment methods, and run short tests to get data at scale before deploying for the end of year campaign.
Some optimizations need more methodical, slower development over the course of many tests, such as the growing support for recurring donations we highlighted in last year’s Fundraising Report.
@SPatton (WMF): Thanks. But just out of interest, how much money did you take with these English Wikipedia pre-tests last month, and for how many days/hours did they run? Any ballpark figure appreciated. I mean, if it's something like $10,000, then nobody here will lose a word over it – and it would be in your interest to say so. But if it's $500,000+, then that's substantial and almost equivalent to a full English Wikipedia fundraising day in December. Regards, AndreasJN46615:47, 4 August 2023 (UTC)[reply]
The number of donations and total revenue we earn from any given testing window varies widely based on length of testing, % of traffic limiting, and what we are testing. Some tests generate only a few thousand dollars in revenue, whereas there are others where that figure is in the hundreds of thousands. All revenue collected in pre-tests contributes to our annual goal for that region, and is reported in our annual Fundraising Report. - SPatton (WMF) (talk) 19:10, 8 August 2023 (UTC)[reply]
The current 2023 banners are way too large and intrusive. It falsely gives the impression that somehow Wikipedia is at risk of being bought up due to lack of funding, and when people look deeper they find out that Wikipedia is sitting on over $100M, and has funding secured for years to come -- which means people are going to be turned off donating indefinitely. At a time where the wealth gap has widened considerably quickly in recent years, and large corporations have been reporting record profits, (especially tech companies like Amazon), for Wikipedia to pull this sort of stunt runs the risk of a major backfire. Smarten up and remove the "Wikipedia is not for sale" line! Thoric (talk) 21:45, 8 August 2023 (UTC)[reply]
First round of new banners incorporating copy suggestions
Hi all. As promised, here are some variants we’ve worked up based off the suggestions on this page to talk about disinformation and the role of Wikipedia in ensuring accurate information; as well as the ‘perfectly imperfect, human nature’ of our work.
I’d like to start by testing these as replacements for the ‘middle part’ of our banner, keeping the opening lines and final call to action intact, and then work the language up based off results.
Feedback on all these banners -- get rid of the "Wikipedia is not for sale". It falsely implies that Wikipedia will need to be sold if donation goals are not met -- when a small amount of research shows nothing could be further from the truth -- which means this is a lie, and it will cause potential and existing donors to feel mislead, and may lead to a backlash. Please reconsider this banner. Thoric (talk) 21:51, 8 August 2023 (UTC)[reply]
Personally, I've never considered that interpretation of "Wikipedia is not for sale", though I agree it would be concerning if readers are viewing it that way. The line originated last year on Twitter when someone suggested Elon Musk should buy Wikipedia to fix its "bias". Jimbo replied "Not for sale", which got an enormous amount of public engagement. It was then suggested for a fundraising banner by Jimbo on-wiki and modified on last year's collaboration page, as seen here. —Ganesha811 (talk) 00:10, 9 August 2023 (UTC)[reply]
Banner 1
Wikipedia is not for sale.
A personal appeal from Jimmy Wales
Please don't scroll past this 1-minute read. This Friday, August 4, I humbly ask you to reflect on the number of times you visited Wikipedia in the past year, the value you got from it, and whether you're able to give $3 back. If you can, please join the 2% of readers who give. If everyone reading this right now gave just $3, we'd hit our goal in a couple of hours. $3 is all I ask.
It's hard to know what to trust online these days. Disinformation and scammers are everywhere. Wikipedia is different. It's not perfect, but it's not here to make a profit or to push a particular perspective. It's written by everyone, together, for no other reason than that they want to help create a free repository of quality information. That's something we all need, like a library or a public park. And because Wikipedia and its sister sites are supported by a non-profit organization, there’s no danger that someone will buy Wikipedia and turn it into their personal playground.
If Wikipedia has given you $3 worth of knowledge this year, please donate now, it really matters. Thank you for your generosity!
Feedback on Banner 1 What do you like and do you have any alternative versions we could test? What would you change and how?
Most of these comments also apply to the other three banners below. "$3 is all I ask" is fine as a sentence but feels redundant after the prior sentences. I would cut "create a free repository of quality information" - the sentence is stronger without it (I note with amusement that I was the one who wrote the sentence originally!). The next sentence could then be modified to read "Wikipedia is something we all share, like...." Finally, I would split out "It really matters." into its own sentence, which strengthens it. —Ganesha811 (talk) 18:03, 4 August 2023 (UTC)[reply]
Banner 2
Wikipedia is not for sale.
A personal appeal from Jimmy Wales
Please don't scroll past this 1-minute read. This Friday, August 4, I humbly ask you to reflect on the number of times you visited Wikipedia in the past year, the value you got from it, and whether you're able to give $3 back. If you can, please join the 2% of readers who give. If everyone reading this right now gave just $3, we'd hit our goal in a couple of hours. $3 is all I ask.
I set up the Wikimedia Foundation as a nonprofit to host Wikipedia and its sister sites because at its heart, Wikipedia belongs to you. Being a nonprofit means there is no danger that someone will buy Wikipedia, turn it into their personal playground, and cut you out. We are passionate about our model because we want everyone to have equal access to quality information - something that is becoming harder and harder to find online.
If Wikipedia has given you $3 worth of knowledge this year, please donate now, it really matters. Thank you for your generosity!
Feedback on Banner 2 What do you like and do you have any alternative versions we could test? What would you change and how?
Banner 3
Wikipedia is not for sale.
A personal appeal from Jimmy Wales
Please don't scroll past this 1-minute read. This Friday, August 4, I humbly ask you to reflect on the number of times you visited Wikipedia in the past year, the value you got from it, and whether you're able to give $3 back. If you can, please join the 2% of readers who give. If everyone reading this right now gave just $3, we'd hit our goal in a couple of hours. $3 is all I ask.
When I set up the Wikimedia Foundation as a nonprofit to host Wikipedia and its sister sites, I envisioned a source of neutral, high quality information. In the age of AI, this vision matters more than ever. As the internet floods with machine generated content, Wikipedia becomes even more valuable for people looking for information they can trust.
If Wikipedia has given you $3 worth of knowledge this year, please donate now, it really matters. Thank you for your generosity!
Feedback on Banner 3 What do you like and do you have any alternative versions we could test? What would you change and how?
I would tweak the final sentence of the first paragraph; "Though the internet today is being flooded with machine-generated content, Wikipedia has always been written by people - people like you. Volunteers, who want to help people looking for information they can trust." —Ganesha811 (talk) 18:08, 4 August 2023 (UTC)[reply]
We have always had a healthy volume of constructive bots and partly-automated edits, and now also have content generated by large language models that has been curated by editors. I wouldn't demonize those things in the message, just focus on the value of "neutral, densely cited, contextualized information". I don't personally like emphasis on "high quality" information "you can trust" because this is a reference that anyone can edit, where anyone can check the references + history themselves. Cf:
Warning: Please be aware that any information you may find on Wikipedia may be idiotic, misleading, offensive, extraterrestrial, dangerous, or illegal. Wikipedia is not uniformly peer reviewed; while readers may correct errors or remove erroneous suggestions they are not obligated to do so. If you need specific advice (medical, legal, psychosexual, financial, arcane, &c.) please seek a licensed, bonded, and knowledgable professional. YOUR INDEPENDENT VERIFICATION OF ANYTHING FOUND IN WIKIPEDIA IS STRONGLY ADVISED.
Wikipedia is not for sale.
A personal appeal from Jimmy Wales
Please don't scroll past this 1-minute read. This Friday, August 4, I humbly ask you to reflect on the number of times you visited Wikipedia in the past year, the value you got from it, and whether you're able to give $3 back. If you can, please join the 2% of readers who give. If everyone reading this right now gave just $3, we'd hit our goal in a couple of hours. $3 is all I ask.
When I set up the Wikimedia Foundation as a nonprofit to host Wikipedia and its sister sites, I envisioned it as a home for high quality, neutral information. I didn't realize it would become a cornerstone of online education with readers around the world. Our nonprofit status helps us focus on providing high quality information for everyone without any risk that someone will buy Wikipedia and turn it into their personal playground.
If Wikipedia has given you $3 worth of knowledge this year, please donate now, it really matters. Thank you for your generosity!
Feedback on Banner 4 What do you like and do you have any alternative versions we could test? What would you change and how?
