This page is meant to be a high-level summary of arguments for and against the two competing models for payment-processing on the non-chapter Wikimedia sites: 1) The Wikimedia Foundation handles all payment processing on the non-chapter sites, and 2) A mixed model, in which the Wikimedia Foundation plus some/many/all chapters payment-process on the non-chapter sites. While it started as a draft by Sue as part of her own Recommendation, edits by everyone are welcome. The goal is to create a neutral summary that does not reflect any one person's viewpoint. It is meant to be high-level. In-depth analysis, if any, should be placed on the talk page.
|Future of fundraising discussions - Index
Guiding Principles discussion
Fundraising models/future discussion
Wikimedia Foundation resolutions
Wikimedia chapter statements
Arguments for and against centralized payment processing by the Wikimedia Foundation.
1) TRANSPARENCY. All Wikimedia fundraising activities need to be truthful with prospective donors. We need to tell people what we intend to use their money for, and we need to report in a timely fashion on how it was actually spent. The Wikimedia movement's experience thus far tells us that in a decentralized payment processing model, it is more difficult to be transparent with donors. Transparency is easier and simpler using a centralized payment processing model.
In our annual campaigns, all parties use the banners and appeals developed by the Wikimedia Foundation, which emphasize the international nature of the projects, the work done by volunteers around the world, the Wikimedia Foundation’s technology needs, and the work done by the Foundation’s staff. This material is designed for use by the Wikimedia Foundation, and when it's used by chapters with no or few alterations, we run the risk of confusing donors about how their money is actually being used. A mixed model of Foundation-and-chapter payment processing would have two choices: it could either use centralized messaging, which runs the risk of being misleading when used in chapter geographies, or it could use more customized, specific messaging, which runs the risk of being less effective (if it impedes donors by being too long or detailed, or by introducing decision points). The centralized model seems to be the solution that best supports transparency while not depressing donations: it is honest, authentic, and simple.
National Chapters can be more transparent because they are best placed to make decisions about grants for people doing things in their country. Not least because they are best placed to know what sort of transport and other costs are realistic. There is also the advantage that decentralising this means that we subdivide the grant assessment work. So it is practical for chapters to use trusted elected volunteers. Once you scale up to a global grant giving body you lose that local knowledge, and you cease to be on a scale where volunteers can do this as a hobby, so you have to pay to employ staff to do this. Because those staff are working on a global level they have to do more research in order to scrutinise less effectively. Chapter trustees will have a vague idea of what costs may be reasonable even in a markedly richer or poorer area of their own country. But when you start dealing in costs across borders you find it takes longer to do any sort of reality check let alone one that a local could do. There is also the issue of translation. The more we decentralise the process the more likely we are to have grant applicants and grant assessors speaking a common language, that saves on translation costs either at the applicant's end or the assessors, it also reduces the risk of people getting away with things because their mistakes are seen as translation errors rather than for example, misunderstanding the charity's remit. Transparency works best on a smaller scale. If a chapter has a budget that is a small fraction of the global budget then donors can attend an AGM in their own country and question the people who are responsible for spending their donation. In a global set of accounts the costs of events in Roma and Lima would risk being combined into one more agglomerated process, inevitably this has to involve less transparency than having the Peruvian chapter accountable for an event in Lima and the Italian chapter for an event in Rome.
2) EFFECTIVENESS (MAXIMIZING REVENUES). Some people have believed that chapters would be more effective at payment-processing in their geographies, compared with centralized payment-processing. However, the evidence thus far suggests that overall more money is raised --per day, or per unit of site disruption/user annoyance-- when the Wikimedia Foundation payment processes, than when the chapters payment process. (Please note that this has nothing to do with messaging: the chapters typically rely on Wikimedia Foundation messaging for the most effective appeals, and even if that were not the case, messaging can be handled separately from payment-processing.)
How do we know this? In 2011, the Wikimedia Foundation payment-processed donations for eight countries in which chapters had payment-processed in 2010. In those countries, the Wikimedia Foundation increased donations an average of greater than 100% in 2011 compared with 2010. (This compares against an average increase of somewhere around 35-40% in the 215 countries in which the Wikimedia Foundation payment-processed both years, and an average increase of somewhere around 40-50% in the four countries in which chapters payment-processed both years.) The reason for the very high increase in some of the countries that switched was that the Wikimedia Foundation made big improvements in ease of donating for users, offered new payment methods, and other factors. In countries where chapters had already achieved a basic standard of usability (e.g., the Netherlands and Hungary) the increases were clearly below the global average.
- A preliminary comparison of money raised by geography, 2010 vs. 2011, can be found here. Please note the caveats in the document. See Fundraising and Funds Dissemination/Data for further summary and discussion.
This suggests that the Wikimedia Foundation will in general be a much more effective payment processor than new chapters who are just beginning to process payments. It also suggests that there is no major inherent gain to be had from local payment processing. Many different variables determined the percentage increases in different countries (e.g. Germany ran for longer, France was still sorting out some basic issues last year, WMF worked to maximize the number of days of editor appeals and minimize the overall number of fundraising days). But the numbers strongly suggest that once basic standards of usability are reached, that the gains in fundraising efficiency come from messaging and creative resources, not having a local processing system. WMF has reached those basic standards and so have a few chapters.
More on effectiveness: The argument has been made that donors are likelier to give to a chapter than to the Wikimedia Foundation, because a chapter can offer specific incentives that the Wikimedia Foundation cannot. It is true that some chapters have the ability to offer some additional specific benefits, such as the UK's Gift Aid program or tax deductibility, that the Wikimedia Foundation cannot. (The Wikimedia Foundation currently is able to offer tax deductibility only for the United States and the Netherlands, whereas six chapters are currently able to offer tax deductibility for their countries.) However, the Wikimedia Foundation has been unable to find any evidence suggesting that tax deductibility is an important incentive for donors giving small amounts of money, and the Wikimedia movement's experience thus far shows no correlation between tax deductible status and amounts donated during the annual campaign. For example, in Italy in 2010 Wikimedia offered tax deductibility via the chapter, and in 2011 we did not. And yet Italy was a very high-growth country in the 2011 campaign. By contrast, in the UK we did not offer tax deductibility for donations in 2010, and in 2011 we did. And yet, in the UK we saw a higher-than-average increase, but not as great an increase as in the countries where we shifted from only taking dollars to also accepting local currency. This is a very small amount of data, and of course there are many factors affecting revenue raised in both countries apart from tax deductibility. But it does suggest that tax-deductibility is not a significant driver of donations. In sum: we don't have evidence that shows us that tax deductibility or other similar incentives are a significant enough driver of overall giving in the annual campaign, for this to be a major consideration tipping the balance towards decentralized payment processing.
More on effectiveness: It could be argued that chapters are better able than the Wikimedia Foundation to offer locally suitable donation methods. The Wikimedia Foundation recognizes that local payment options are important to effective collection of donations in the local geographies, and that’s why it has purposely chosen to work with service providers that can facilitate local payment mechanisms such as local bank transfers, local card types, and other means for payment. This year, in the three months between the Board’s letter to the chapters and the launch of the campaign, the Wikimedia Foundation worked with a firm called Global Collect to expand the local payment options we offer for donors. This resulted in the Wikimedia movement being able to offer donors many more payment methods (deployed or nearly deployed methods include one new credit card (JCB), six online direct debit options, direct debit in 12 different countries, two e-wallet products, two cash products, local bank transfer in over 50 countries and credit cards in 70 new local currencies) than we had been able to offer donors in 2010. The Wikimedia Foundation was not able to implement direct debit in a couple of countries that had it last year, but we will have those working in the next couple months.
On an initial analysis of the 2011 fundraiser it looked as if people were more willing to give to the Foundation than to a chapter in their own country. Closer analysis has shown a more complex picture.
- The most important thing is to make it easy to donate - accept the currency and banking methods that people are comfortable with.
- Pretty much everywhere we did better in 2011 than in 2010.
- Our worst results on a like for like basis were in the two countries where the WMF took over from a chapter and couldn't improve on the donation experience.
- Tax incentives are complex and very different around the world. Though they are clearly important for us in the UK, the jury is out in some other countries.
- There are probably some parts of the globe that would respond better to a global appeal than to one from "their" Chapter. Palestinians, Scots, Basques and Catalans may all be worth testing for a global or national as opposed to Chapter fundraiser.
In conclusion, if a chapter can get appropriate payment processing in place and the Country level identity has general support of the population, there is potentially a >20% advantage in having the chapter fundraise instead of the WMF. Tax incentives are one of the areas that need proper analysis and of course a differently worded email to those who have previously asked for a tax receipt as opposed to those who didn't.
3) EFFICIENCY. The Wikimedia movement aims to raise the maximum possible amount of money from donors while minimizing administrative costs as much as possible, in order to reserve the largest amount of money possible for programmatic activities. Experience tells us that for chapters to payment process during the annual campaign is needlessly expensive, because it duplicates work already being done by the Wikimedia Foundation and services used by the Wikimedia Foundation such as Global Collect.
Originally, some people felt that chapters acting as payment processors for the annual campaign might be cost-effective, particularly to the extent the work could be done by volunteers. We know now that isn’t true. Volunteers simply don’t have time to create and maintain appropriate financial, technical, administrative, legal and communications systems and processes supporting the annual campaign. Chapters that tried to do that work in past years using volunteers found themselves unable to do the necessary work  on time, which exposed the movement to unacceptable risk and resulted in many payment processing chapters deciding to incur expenses such as the hiring of staff and engagement of third-party payment processor companies. These are unnecessary additional costs, given that the Wikimedia Foundation is already creating and maintaining appropriate financial, technical, administrative, legal and communications systems and processes supporting the annual campaign. Further, centralized payment processing enables economies of scale in the relationship with third party companies such as Global Collect and PayPal, reducing costs for collection of payments and enabling better bulk processing rates, more-favourable contractual terms, and overall better relationship management and support.
 These systems and processes include: developing a solid understanding of the laws regulating fundraising activities and ensuring compliance with those laws; creation of, and adherence to, policies such as gift policies and donor privacy policies; creation of a website for donors as well as implementation of an online payment system and a donor database; maintenance of site uptime during the campaign so donations are not lost; creation of systems for privacy safeguards, fraud protections and security measures preventing data theft and phishing; creation of systems for answering donors inquiries and processing refunds; creation of systems for thanking and sending receipts; creation of systems for managing payment subscriptions; creation of annual plans and annual reports; creation of marketing materials explaining what the donations will be used for and reporting back on how they were used; creation of financial accounting procedures, controls and auditing practices sufficient to ensure adherence to appropriate standards and all relevant laws and regulations, and purchase of insurance portfolios to manage potential risks.
4) EASY TRANSFER OF FUNDS INTERNATIONALLY. The Wikimedia movement is international in nature, and our fundraising practices need to support the easiest possible transfer of money internationally in support of the movement’s priorities. The Wikimedia movement's experience thus far tells us that processing payments in a single location rather than multiple locations is the simplest and most effective way for us to facilitate the easy transfer of money around the world. There are very few laws restricting the Wikimedia Foundation’s ability to move money out of the United States -- fewer than restrict the movement of funds out of other countries such as for example France, where the proportion of money that can be moved is capped. Centralized payment processing out of the United States, or only allowing chapters to fundraise in countries which don't restrict international transfer are the two options that give the movement maximum freedom to move money around internationally to meet the movement’s objectives.
5) ROBUST CONTROLS. People have occasionally suggested that payment processing should be decentralized to minimize risk of fraud or misuse of funds. However, centralized payment processing by the Wikimedia Foundation poses less risk of fraud or misuse of donor funds, compared with decentralized payment processing, because 1) the Wikimedia Foundation has extremely robust controls, and 2) the laws of probability show us that the smaller the number of payment processors, the lower the risk.
Robust controls: The Wikimedia Foundation has been actively strengthening controls since inception in 2003, and particularly since 2007. In 2008, the Wikimedia Foundation engaged as its auditor KPMG, one of the world’s Big Four audit firms, and since then it has actively worked to implement KPMG recommendations eliminating controls gaps and tightening up its practices. Its Board has engaged in active Board development during that same period, improving its governance systems consistent with non-profit best practices. It employs an experienced executive team including an ED, CFA and GC each with more than twenty years of professional experience.
National Chapters can have more robust controls because they are best placed to make decisions about grants for people doing things in their country. Not least because they are best placed to know what sort of transport and other costs are realistic. Once you scale up to a global grant giving body you lose that local knowledge and the staff have to do more research in order to scrutinise less effectively. Chapter trustees will have a vague idea of what costs may be reasonable even in a markedly richer or poorer area of their own country. But when you start dealing in costs across borders you find it takes longer to do any sort of reality check let alone one that a local could do. There is also the issue of translation. The more we decentralise the process the more likely we are to have grant applicants and grant assessors speaking a common language, that saves on translation costs either at the applicant's end or the assessors, it also reduces the risk of people getting away with things because their mistakes are seen as translation errors rather than for example, misunderstanding the charity's remit.
Laws of probability: Let’s say for the sake of argument that each organization processing payments in the annual campaign has a 1% chance of experiencing fraud. This means that centralized payment processing, with one payment processor, has a 1% chance of experiencing fraud. The probability of fraud rises with each additional payment processor: with two it is 1.99%, with three it is 2.97%, with ten it is 9.56%, with thirty it is 26.03% and with fifty it is 39.5%. This is important because the reputations of payment processors are linked, because they are all processing payments for the same campaign. If one entity experiences fraud or malfeasance, everybody's reputation will be damaged.
The risk of fraud is a matter of much more than just probability. Scale, impact on confidence and how an organisation responds are also important. In a decentralised system the scale of likely frauds is by definition lower as more individuals have control over smaller sums. But more important is the issue of relevance, in a decentralised system a fraud can be largely a problem in the relevant country and can be more easily dealt with. "Medium sized sum stolen from Country X branch of Wikimedia, fundraising in Country x will now be done by Global HQ until further notice" is not a nice story to have happen. But it is not a big story outside of Country X and the movement itself, nor is it necessarily a big financial problem that threatens the survival of the movement. A more centralised system where all eggs are in one basket may make the basket of mithril and be confident that fraud is unlikely. But if something did go wrong it would risk the confidence of all donors and of the whole movement and all the movement's money.
6) MINIMIZING LEGAL EXPOSURE. It is important for Wikimedia fundraising activities to be conducted in a manner that doesn’t create unnecessary legal exposure. I believe that centralized payment processing by the Wikimedia Foundation creates less legal exposure for the movement, relative to decentralized payment processing. When payment processing happens in a decentralized fashion in multiple countries, chapters are required to develop compliance programs to ensure they follow their local laws and regulations in areas such as privacy laws, data security compliance, and fundraising compliance. The Wikimedia Foundation is in a better position than chapters to manage these risks: it has one of the best law firms in the world (Covington & Burling) working with the top law firms in individual countries, to ensure compliance and a consistent strategy.
Further, when a chapter directly collects donations, this results in larger bank balances, which makes the chapter a tempting target for unfounded lawsuits and other legal exposure. This is of course also true of the Wikimedia Foundation, but the Wikimedia Foundation is in a better position through resources and infrastructure to manage legal exposure and defend meritless lawsuits.
A centralised model has the disadvantage of having a single point of failure. If the WMF was taken down by a court case the movement would be crippled. A strategy of not having all our eggs in one basket might have some extra legal costs but it would make the movement as a whole much more legally secure
7) INDEPENDENCE. There has been a lot of discussion about how best to preserve chapters’ independence when it comes to fundraising. Chapters have argued that being able to payment-process enables them to remain independent of the Wikimedia Foundation in a way that receiving grants does not. In truth, though, this is a fiction. Decentralized payment processing is not a path that leads to chapter independence from the Wikimedia Foundation. On the contrary, decentralized payment processing by a chapter intimately ties that chapter to the Wikimedia Foundation for permission to payment process, fundraising strategy, messaging and campaign coordination. Chapters that aspire to independence can and should fundraise off-wiki independently, as well as outside of the Wikimedia movement.
Being a legal entity which the public has given money to gives a chapter a degree of independence from the WMF but increases its responsibilities to local regulators, donors and members. This applies even if the images may be centrally supplied and the messages translated from common messages.
8) MAJOR DONOR RELATIONS. It could be argued that chapters have an advantage in building major donor programs in their countries, relative to the Wikimedia Foundation. This is almost certainly true: chapters speak the same language and share the same culture as others in their country, and proximity for face-to-face interaction is important in major donor cultivation. And it is true that major donor prospect identification and cultivation would be easier if the chapter were the primary initial point of contact for a given donor. This is a legitimate challenge, and worth trying to address. However, it likely isn't significant enough to, by itself, tip the balance towards decentralized payment processing.
9) RELATIONSHIPS. Whether or not a chapter is approved to payment-process during the annual campaign has been a source on ongoing stress and tension in the relationship between the Wikimedia Foundation and the chapters. The Wikimedia Foundation's fiduciary responsibilities make it impossible for it to allow chapters to payment process unless they have proved themselves to be competent and responsible stewards of donor funds, and yet as independent organizations the chapters resent and resist the Wikimedia Foundation's efforts to oversee them. This is a fundamental and irreconcilable tension which is causing great damage to relationships between the Wikimedia Foundation and the chapters on an ongoing basis, as chapters try to make the case that they are ready to payment-process, and the Wikimedia Foundation tries to find an appropriate balance between its desire to trust and support chapters, and its obligation to ensure donations are safeguarded. I do not know if relationships will improve over time if chapters know that payment-processing is off the table, but I do believe relationships will continue to deteriorate, if payment-processing continues to be seen as something chapters should aspire to do.
It isn't healthy for the movement that the Foundation and the chapters are at loggerheads. Out and out "victory" for either side would at best be hollow. The chapters need a global body to do those things that work best at a global level, the WMF needs volunteers and a healthy self organising community in as many countries as possible. The best result would be one where the Foundation accepts that the chapters should do the things that they are ready to do and can do most efficiently, and that as a consequence it should welcome chapters reaching the stage when they are ready to take on a particular responsibility. In turn chapters would probably be far more trusting of the foundation if it only centralised where needed, rather than by default.