Grants:PEG/Retrospective 2009-2012

I’m Kevin Gorman. I’ve been editing Wikimedia projects as a registered user for about two years. My username was originally kgorman-ucb, but is now simply User:Kevin Gorman. (I edited as an IP user prior to creating an account – my first IP edits were around 2004.) As an editor, I am primarily active on the English Wikipedia, although I pop on to other projects from time to time. I am significantly involved in the US Education Program and other real-world Wikimedia related outreach and co-moderate Gendergap-l (all as a volunteer.) From June to October of 2011, I was also an unpaid communications intern at the Wikimedia Foundation.

Scope

This report is also available as a PDF

This is a review of the Wikimedia Foundation’s grants program from fiscal year 2009-2010 through fiscal year 2011-2012. It is intended to serve as a review of the processes of the grants program, highlighting areas where the processes surrounding the grant program have done especially well, highlighting areas where they have historically been weak, and offering suggestions as to how they can be improved moving forward. It is not intended to serve as an analysis of the success of individual grants except inasmuch as examining individual grants can shed light on strengths and weaknesses present in the processes of the grants program.

The scope of the review is explicitly limited to grants given through the Grants:Index page. Participation grants given through Participation:Support are not in the scope of this review, though participation grants that were handled through Grants:Index are included. Grants given through the Funds Dissemination Committee are not in the scope of this review. The Grant Advisory Committee is also not in the scope of this review.

This review was commissioned by the Wikimedia Foundation (and I have been paid to complete it.) The staff of the Wikimedia Foundation have played an important role in the preparation of this report by providing background information, historical records, and their own thoughts and views about the program. Interim and final drafts of this report were also shared with Foundation staff, in the interests of double-checking facts and figures and addressing additional questions that came up during my research. However, they have not exercised editorial control over the contents of this report. A number of grantees including Petr Brož and Petr Novák of WMCZ, Mile Kiš of WMRS, Bala Jeyaraman and Srikanth Lakshmanan of the Tamil community, Sarah Stierch, and several additional grantees who wish to remain anonymous provided valuable insight that allowed for the creation of this report.

The opinions expressed herein are my own – not those of the Wikimedia Foundation.

Executive summary

The Wikimedia Grants program started in February of 2009 as a way to facilitate the valuable programmatic work of volunteers. Since its inception, the program has received more than 186 applications and disbursed more than $1,905,900 of movement funds to 154 grantees. The program has grown steadily, and has successfully facilitated valuable programmatic work by directing movement funds to partners who are better positioned to carry out on-the-ground work than the Foundation itself may be.

Total amount of money (in USD) distributed by the WMF grants program by fiscal year

The increasing expertise of the Wikimedia Foundation’s grant-making team combined with its steadily increasing resourcing has resulted in better and more consistent results as the program has progressed. If risks are proactively addressed and the program continues to have strong leadership and sufficient resources, the grants program is likely to facilitate an increasing amount of important programmatic work as time goes on.

Major milestones in the history of the WMF grants program

Structure

This review begins by briefly examining the history and performance of the grants program year by year for its first three years of existence, pulling out a couple major grants from each year for additional focus. It then moves on to more detailed case studies – one of Wikimedia Czech Republic’s Mediagrant, and one of the grant-related experiences of Mile Kiš and Wikimedia Serbia. The final sections detail risks and recommendations culled from earlier sections, my interviews with grantees and Foundation staff, and my analysis of publicly available information about the grants program.

The final sections contain explicitly actionable items that address problems that have already occurred or are likely to occur in the grants program, and provide some non-risk related actionable suggestions about how the grants program could be improved. The program in its earliest phases experienced significant growing pains, but their causes have mostly been addressed. Most of the unaddressed risks and procedural imperfections that affect the current and future performance of the grants program can be mitigated to an acceptable level.

Summary of risks and recommendations

The most significant (but hardest to mitigate) category of risks are those related to broader problems within the Wikimedia movement. Internal tensions within the Wikimedia movement are likely to have a significant negative effect on the functioning of the grants program, but there’s little that can be done in the specific context of the grants program to help that. Language and cultural barriers will certainly represent a severe constraint on the effectiveness of the grants program, but other than direct outreach to targeted organizations in underserved areas, this too is hard to address. Some risks in this category have the potential to be catastrophic and are thus worth being aware of, even if they are impossible to effectively mitigate.

The next category of risks has potential to do severe but localized harm. These risks are not likely to cause broad harm to the program, except through reputational harm. These risks could be financially costly, and may damage (or temporarily eliminate) WMF’s ability to do outreach in particular geographic areas. Almost all risks in this category can be significantly mitigated, although very few can be completely eliminated. Significant financial malfeasance by a major grantee would be costly, could cause the Wikimedia Foundation significant reputational harm, and could damage our ability to do outreach in a particular geographic area. A grant carrying an unpredictable but untenable administrative burden could damage (or even destroy) the capacity of the grantee organization to do productive work, and could potentially also drive off large numbers of content contributors. Risks in this category can almost always be significantly mitigated, and since they can cause significant harm, should be.

The next category of risks has the potential to do relatively minor localized harm to the program, or the potential to stop individual grants from reaching their full potential. If a grantee uses an inconsistent top level financial report, then it may take WMF staff more time to verify that the grant was used well than it otherwise would. These problems are generally not earth-shattering even on a localized basis and rarely have broad effects, but since they are frequently incredibly easy to mitigate, they should be addressed as staff time permits.

The next category in the risks and recommendations section addresses issues that amount to simple fixable oversights. Some of these oversights have caused significant localized harm to the program in the past, but everything in this category has an easy fix – often, a simple technical fix. Most or all of these problems should be fixed immediately, whether by my suggested method or another. None of the fixes should be very time consuming.

The last category in the risks and recommendations section suggests improvements to the program. These recommendations do not address any particular risk, but are ways in which the program may potentially be improved moving forward. They should be considered by the Wikimedia Foundation’s grants team on a case by case basis; some may be worth implementing immediately, and most should be worth implementing in one form or another as more staff time becomes available.

Fiscal year 2009-2010

Overview

The Wikimedia Grants program launched on February 18th, 2009. Its first year was, in many ways, an experiment. As can be expected from an experimental project, a number of significant successes occurred, and a number of problems are retrospectively visible. The program suffered from underdeveloped supervision and transparency and accountability requirements, insufficient staffing, and a number of other weaknesses. Most of the program’s initial weaknesses have been addressed in subsequent years. A number of initial grants resulted in better than expected outcomes, which in some cases produced excellent long-tail returns. Despite its weaknesses, the grants program facilitated important programmatic work in its initial year, and demonstrated the merit of its concept.

Details

The Wikimedia Grants program was announced on Meta on February 12th, 2009, and started accepting grant applications on February 18th. From February to April the grants program was administered by Erik Moeller, deputy director of the Wikimedia Foundation, as one of many responsibilities. From April until September, the grants program was administered by Jennifer Riggs, Chief Program Officer of the Wikimedia Foundation (again, as one of many responsibilities.) After Jennifer resigned, administration of the program lapsed back to Erik for the remainder of FY09-10.

In fiscal year 2009-2010, the program received 34 submissions, of which 24 were initially approved, 5 were rejected, two were withdrawn, and three were withdrawn or cancelled after they had originally been approved. 13 Wikimedia chapters applied for grants (of which 11 had at least one grant approved,) and no non-chapter entity applied.

Grant applications totaled $159,878. A total of $113,585 of funding was approved for grants in the year, meaning that ~66.4% of all requested funding was approved. Almost all approved grant applications received the level of funding they requested. Approved grant applications requested an average of ~$4,200 in funding each, whereas grant applications that were not funded requested an average of ~$6,500.

Grant decisions were made in as few as 33 days, and as many as 373. Funded grants took an average of 145 days to approve on Meta, while grants that were rejected or withdrawn averaged 306 days to have their status shown on Meta. 145 days is an exceptionally long period of time to expect a grant applicant to wait before learning if their project has received funding. It is possible that some grant decisions were made and conveyed to the applicant substantially earlier than they were posted on Meta. It is not possible to tell from publicly available information how often this occurred (or if it at occurred at all.) Even if all grants that took longer than 175 days to be listed as approved and an average time-to-approval was calculated without taking those grants into account, the average time-to-approval would still be 89 days. This waiting period is normal for many grant-making programs, and is not considered out of line with grant giving best practices. However, a waiting period this long would pose a significant burden to many applicants, especially applicants used to dealing with the comparatively agile processes of the Wikimedia world. At its launch, the grants program had relatively lax accountability and transparency requirements compared to recent years, and they were not stringently enforced. In FY09-10, 21 of 25 executed grants resulted in on-wiki reports, only 14 of which were reasonably comprehensive, and almost all discussion related to grant applications, approved grants, and grants reports was held off-wiki. Grant reports were accepted by both email and postal mail; there was no expectation to post a grant report on Meta until February of 2010. There was a written set of guidelines about what reports were expected to include, but they were not enforced. Without access to email logs from this time period it is hard to directly determine what level of interaction Foundation staff members had with grantees, but the evidence that is available to me seems to show that many grants had relatively little Foundation-side follow-up, that only intermittent attempts were made to make sure that grantees made appropriate use of their grants, and that few attempts were made to ensure that lessons learned were preserved and made good use of.

Of the 21 executed grants with on-wiki reports, 15 reports had clear indicators of the success or failure of the grant in the eyes of the organizers. (5 reports either didn’t have clear indicators of the success or failure of the project in the eyes of the organizers or had outcomes too complex to easily classify. One grant has not yet reported because it is not yet actually finished.) Of the 15 grant reports with clear indicators of outcome in the eyes of the organizer, 2 grants failed to meet their expectations, 9 met expectations, and 4 exceeded them. 8 of the 15 reports with clear indicators of outcome in the eyes of the organizers contained enough information about the grants to conduct an independent evaluation of the success of the grant as gauged by meeting or exceeding the goals of the grant as stated in its initial application. In those eight cases, I believe that the organizers accurately judged the success (or failure) of their grants in meeting their original goals. (A couple of grants were bigger successes than was apparent at the time the reports were written, due to long tail benefits in subsequent years not predictable at the time the grants or reports were made. )

Although most approved grants in FY09-10 were successful in meeting their stated goals, it is unclear if the goals stated in grant applications received any meaningful degree of scrutiny before being accepted. If the stated goals of grants don’t receive adequate scrutiny, then a grant could ‘succeed’ while still being an inefficient use of movement funds. In subsequent years, the amount of scrutiny given to grant applications has increased significantly, which seems to have diminished this risk to an acceptable level.

Many chapters applied for multiple grants simultaneously, sometimes for activities that were interdependent. A situation where the funding status of one grant may affect the feasibility or budget of another grant is generally undesirable. (In at least one instance in FY09-10, the rejection of one grant did adversely impact the outcome of another grant given to the same chapter.) It would be preferable if grant applications from the same entity in the same timeframe were consolidated, especially where the successful execution of one activity is dependent on the successful execution of another activity. That way, applicants can make all necessary adjustments if not all grant money is approved during the application process, and not after the disbursal of funds.

Key grants within the year

Wikimedia UK’s kick start grant had unexpectedly significant results; it led to Liam Wyatt’s successful Wikipedian-in-Residence posting at the British Museum. Liam’s residency was the first of its kind, and provided a model that many subsequent GLAM collaborations have adopted. Content access projects were only a small part of the overall grant, but unpredictably generated Liam’s productive residency and laid the groundwork for many future GLAM collaborations. This grant highlights well one of the main purposes of the Wikimedia Grants program – enabling creative work by the community that would not be achievable without financial support that may provide outsize results.

Wikimedia Czech Republic’s $1,250 grant for a camera with accessories produced interesting results, which are further discussed in the later section that discusses the series of media grants that Wikimedia Czech Republic has received in greater detail. It highlights many of the benefits and drawbacks of grants that involve equipment purchases intended for Wikimedia movement members to manage over time. (The initial camera grant substantially succeeded its stated goals; by the time the initial grant report was submitted in June of 2010 more than one thousand photos taken with the camera had been uploaded to the Wikimedia Commons, primarily of quite high quality.)

2010-2011 overview

Summary

FY2010-2011 was the second year the Wikimedia Grants program existed. More than five times as much grant money was distributed in FY10-11 as in FY09-10. The program began to show signs of maturity; accountability and transparency requirements were greater than they were in FY09-10, and more attention was paid to grant applications and reporting. The program sponsored a wider variety of activities than it had in the previous year, and issued considerably more grants than it had in its first year. The grant process was more selective than it had previously been, but was not excessively so. Additional staff time and community participation allowed for more in-depth decisions to be made regarding the partial funding of grants, which increased the ROI of grant funds.

Details

During the first two months of FY10-11, responsibility for the grants program transitioned from Erik Moeller to Barry Newstead, the newly hired Chief Global Development Officer of the Wikimedia Foundation. Like Erik, Barry had many other responsibilities as well, though he was able to dedicate significantly more of his time to the grants program than Erik had been. Winifred Olliff was hired as a program assistant for the global development team in November of 2010, and began to assist with grant program related work. In April of 2011, Asaf Bartov was hired as the head of Global South Relations by the Wikimedia Foundation and took over from Barry as manager of the grants program. Asaf was able to dedicate substantially more time to the grants program than either Barry or Erik had been able to, and the increased amount of staff time in FY10-11 resulted in noticeably increased results.

In fiscal year 2010-2011, the program received 66 submissions, of which 46 were approved, 9 were rejected, and 11 were withdrawn. Of the 66 submissions, 50 came from Wikimedia chapters, 4 came from proto-chapters (chapters not yet formally incorporated,) 11 came from individual members of the Wikimedia movement, and one came from a non-movement organization. Many chapters applied for multiple grants – the 54 chapter submitted grant applications represented 15 full chapters and 2 proto-chapters. Every chapter and proto-chapter that applied for a grant in this fiscal year had a grant approved.

Grant applications totaled $1,156,070 ($606,070 excluding one very large request.) A total of $512,080 of funding was approved for grants in the year, meaning that ~44.3% of all requested funding was approved (~84.5% excluding the one unusually large denied grant request.) Approved grant applications asked for an average of ~$11,900 each, whereas grant applications that were not funded asked for an average of ~$22,000 each.

Grant decisions were made in as few as 11 hours after they were posted on Meta and as long as 182 days. On average, grants that were approved were approved in ~45 days, whereas grants that were rejected or withdrawn took an average of ~170 days to resolve. Deficient grants were left open by WMF staff for a longer period of time than approved grants were so that the applicant would have a chance to repair the perceived deficiencies. These timeframes are roughly half the length of those in FY09-10, and were likely made possible as a result of greater staff attention to the program. They represent a substantial improvement to the functioning of the program; a 45 day wait, although still not ideal, would inconvenience grant applicants far less than the previous 90 day wait did.

The grant process was substantially more selective in FY10-11 than it was in FY09-10, both in terms of overall approval rates and in terms of funding percentages for grants that were approved. There are two apparent factors that likely contributed to this. Firstly: the program’s profile grew as it went into its second year; it received more applications from more entities for a broader variety of activities with a higher average requested amount. This increase in both diversity of activities and average amount probably accounted for a large part of the overall increase in selectivity. Secondly: the program received substantially more staff time than it did in its first year, which meant that grant applications received substantially more scrutiny than they had in FY09-10. An increased level of scrutiny as a result of both increased staff time and an increase in public participation in grant discussions on-wiki would have allowed for an increased ability to determine when a grant could be adequately executed with only partial funding in comparison with what was possible FY09-10. (This also could have accounted for part of the increase in overall selectivity.)

FY10-11 grant recipients were subject to generally similar theoretical requirements regarding accountability and transparency (although on-wiki discussion was far more normative in FY10-11 than it previously was,) but the requirements were enforced more stringently than they were in FY09-10. On-wiki grant reports were submitted for 42 of the 46 (~84.8%) approved grants during the year. (The missing four grant reports accounted for $7,700 of funding, ~1.5% of total grant funding for the year.) Of the 42 submitted reports, 34 of them were reasonably comprehensive. Most initial reports were substantially more detailed than was true of FY09-10 grants, and WMF staff intervened seeking additional clarity in many (but not all) FY10-11 grants that initially had inadequate grant reports. The combination of increased initial report quality and increased WMF intervention points clearly towards evolving and increasingly strong standards for transparency and accountability in comparison with the first year of the program.

Of the 42 grants with on-wiki reports, 33 of them contained clear indicators of the success or failure of the grant in the eyes of the organizers. (The other 9 either contained insufficient information, or had outcomes too complex to easily classify.) Of the 33 reports where the success or failure of the grant in the eyes of the organizers can be deduced, 21 grants met the expectations of the organizers, 6 significantly exceeded them, and 6 failed to meet them. Most of these reports contained enough information for me to make an independent evaluation of the success or failure of the grant against its stated goals, and in almost all instances where reports were detailed enough to allow me to make such an evaluation, my assessment of the grant’s performance was the same as the organizers’ assessment.

As in FY09-10, it appears that the stated goals of grant applications did not always receive significant scrutiny, and almost certainly were not always appropriate ways to measure the success or failure of the grant. Grant applicants should be strongly encouraged to include falsifiable metrics of success in their applications whenever possible.

Key grants within the year

A number of grants were given in FY10-11 to cover parties for Wikipedia’s tenth birthday in various cities across the world. The purpose of these parties meant that their success is generally difficult to quantify. Despite this, reports on these sponsored events generally indicate that they were well-received, attended by more people than anticipated including people who had never edited Wikimedia projects before, and that some received positive press coverage. The sponsored 10th anniversary celebrations were attended by at least 750 people, at a total cost of less than $6,320 (some of the events underspent their budgets.) Although a total sponsorship of ~$8.40 a participant may seem steep, it is lower than many other Wikimedia events have incurred, and since it facilitated both outreach and real-world connections between existing Wikipedians in areas that may not have regular meet-ups, seems likely to have been worthwhile.

A large grant was given to the John Ernest Foundation in support of Wikisym2011. This was the first grant of any size given by the Wikimedia Grants program to a non-Wikimedia movement organization. Wikisym is an annual conference focused on research into wikis, including into Wikimedia Foundation wikis. In 2011, eleven WMF-focused papers, one WMF-focused demo, and seven WMF-focused posters were presented. More research about WMF projects was presented at Wikisym than anywhere else in 2011. Wikisym is a good example of where a WMF grant to a non-movement group (for a mission-aligned event) can be very worthwhile. The Wikimedia Grants program currently has very limited visibility outside of the movement. It is likely to prove difficult to raise the profile of the Wikimedia Grants program outside of the movement, but increasing the Grants program’s engagement with non-movement members (for movement-aligned goals) is likely to be worthwhile.

2011-2012 overview

Summary

FY11-12 was the third year of the grants program. The processes surrounding the grants program showed substantial maturity. Thanks to an increasing amount of staff time, grant applicants (both successful and unsuccessful) had far less time to wait to find out the status of their grant than in any previous year.

Average time (in days) until a grant decision is made broken down by year

Grant applications received more thorough and more consistent scrutiny than they had in any previous year. Many grant reports have not yet come due, but those that have have generally been more comprehensive than in any previous year. The Grant Advisory Committee also became function during the first part of FY11-12, which greatly increased the amount of community feedback grant applicants received.

Details

Asaf was the primary point of contact throughout FY11-12, with Winifred assisting throughout the year and taking on more duties towards the end of the year. Asaf settling into his position coupled with Winifred spending an increasing amount of time working with the grants program meant that FY11-12 was the first year that the theoretical accountability and transparency were rigorously enforced. Asaf and Winifred were able to spend far more time trying to ensure adherence to the requirements than previous grant staff had been able to, although the amount of backlogged administrative work that they had to work through meant that they only fully cleared the administrative workload debt in late 2012 (after FY11-12 had ended.)

In fiscal year 2011-2012, the program received 67 submissions, of which 55 were approved, 4 were rejected, and 8 were withdrawn. Of the 67 submissions, 36 came from Wikimedia chapters, 5 came from proto-chapters, 19 came from individual Wikimedians or non-incorporated groups of Wikimedians, and 7 came from non-movement organizations. The 39 chapter submitted grant applications represented 20 chapters over-all. Every chapter and proto-chapter that applied for at least one grant in this fiscal year had at least one grant approved. FY11-12 introduced annual plan grants, a new category whereby established chapters could request funding for an entire year in a single grant (which accounted for the drop in applications from FY10-11.)

Grant applications totaled ~$1,376,772. A total of $1,280,299.05 of funding was approved for grants in the year, meaning that ~92.9% of all requested funding was approved. Almost all approved grant applications were funded for the full amount requested. Approved grant applications asked for an average of ~$23,278 each, whereas grant applications that were not funded asked for an average of ~$9,268 each. Excluding grants for annual plans – which were available in FY11-12 for the first time – approved grant applications asked for an average of ~$7,644, which continues the trend of approved grants asking for less money than rejected grants. (It seems fair to exclude grants for annual plans from this calculation; chapters eligible for annual plan grants are well-established, and annual plan grants are typically discussed with WMF staff beforehand. So far, no annual plan grant application has been rejected.) By straight numbers, the grants program was substantially less selective in FY11-12 than it was in FY10-11, which seems likely to be an artifact of the increasing amount of assistance given to applicants in framing their grants coupled with the fact that many grantees had already navigated the grant process at least once and the creation of annual plan grants.

Grant decisions were made in as few as 12 hours after they were posted on Meta and as long as 198 days (discounting one grant that was left open for 556 days to allow the grant organizers as much time as necessary to potentially revise the grant.) Grants of any status took an average of ~36.4 days to approve (~51.1 days to approve not excluding two outliers.) On average, approved grants were approved in ~34 days (~42.5 days excluding one outlier,) whereas grants that were rejected or withdrawn took an average of ~48.2 days (~90.5 days not excluding one outlier.) Deficient grants were left open by WMF staff for a longer period of time than approved grants so that the applicant would have an opportunity to correct the perceived deficiencies of the grant. The timelines for both approved and rejected grants were substantially shorter than they were in previous years – approved grants were approved in approximately 75.6% of the time it took in FY10-11, and rejected and withdrawn grants had their statuses finalized in approximately 28.4% of the time it took in FY10-11. The Grants Advisory Committee became operational around the start of FY11-12, and the organized community scrutiny that it created resulted in lower waiting times and more feedback to applicants.

In many situations, extended wait times may negatively impact the outcome of a grant, or sometimes even its feasibility. Grant applicants must often make time sensitive decisions that may involve more funds than they have access to without an approved grant (or at least more funds than they feel comfortable personally expending without confirmation as to whether or not they will be reimbursed. Since its launch in FY09-10, wait times for the grants program have consistently been improving. FY11-12 was the second consecutive year of significantly improved wait times. In its initial year of operation, successful grant applicants had to wait an average of nearly five months before knowing if their project was good to proceed; in FY11-12 they had to wait only ~36.4 days on average. This is an incredibly significant improvement to the functioning of the grants program; wait times for grants no longer represent a substantial impediment to the successful execution of grant-funded activities. (Since grant waiting times have become more consistent, it would be worth including an explicit mention of the approximately expected waiting time on Meta to give potential grant applicants a better idea of how long the process takes.)

Not all grants from the fiscal year have reached their reporting deadlines yet, but most grants that have done so have submitted a report in a timely fashion. Since only around half of all grants from FY11-12 have come due and submitted their grant reports, it seems premature to try to compare metrics from this year with those from past years. Of the 37 grants that have passed their reporting deadline, 28 grant reports have been accepted, 7 are in review, and only 2 are past due (less than one third of one percent of grant funding is overdue for a report.) All but one report submitted so far have been reasonably comprehensive, and grantees' assessment of their own success or failure has universally agreed with my own. Most of these reports contained enough information for me to make an independent evaluation of the success or failure of the grant against its stated goals, and in almost all instances where reports were detailed enough to allow me to make such an evaluation, my assessment of the grant’s performance was the same as the organizers’ assessment. FY11-12 was the first year where the stated goals of grants frequently received explicit scrutiny, which is a notable improvement to previous years.

Key grants within the year

An $800 grant was given in FY11-12 to provide prize money and cover miscellaneous costs for a media contest being run across the Tamil language projects. The grant’s stated goal was attracting 3,000 appropriately licensed media files and 100 new contributors; it ended up attracting 15,000 appropriately licensed media files and 251 new contributors. The uploaded media files have an article-space usage rate of 53.63%, quite high for a recent contest. The grant significantly raised the profile of Wikimedia Commons among the Tamil photographic community, and attracted a number of new quality contributors. This grant is discussed in more detail in the interviews section. WikiConference India was supported with a $40,000 grant. The conference was, as a whole, successful, with more than 700 attendees in attendance. The event was – by far – the largest event put on by the community in India up until that point in time, and was very ambitious. The report of the organizers contains much valuable information. On the talk page of that report, Asaf summed up many of the lessons learned from WikiConference India. Future organizers of major events like WCI would be well-advised to read over both the WCI report and Asaf’s summation.

Case Studies

Wikimedia Czech Republic Mediagrant

There are many lessons that can be learned from the Mediagrant given to Wikimedia Czech Republic, and I feel that it is worth examining in depth. The Mediagrant highlights some of the good things about the grants program, and also many of its serious dangers. It is worth noting that the Mediagrant is not at all a typical grant – I am including a detailed study of it because it is in many ways unusual, not because it is a representative outcome of the grants program.

In preparing this case study, I have spoken extensively about the grant with Asaf Bartov and Winifred Olliff of the Wikimedia Foundation, Petr Brož (Chmee2) and Petr Novák (Che) of WMCZ. I have also read and reviewed most public information about the grant, including a good portion of the public information written in Czech. (I had to rely on Google Translate for some of the Czech information, although I had a number of exchanges of particular interest translated for me by a native Czech speaker who is not involved in the Wikimedia movement.)

The Mediagrant

Wikimedia Czech Republic received a 21,000 euro ‘Mediagrant’ in December 2010, which was intended to promote the production of high quality multimedia within the Czech Republic. The grant was intended to increase the production of freely licensed high quality multimedia in the Czech Republic. The grant aimed to do this in several ways – directly reimbursing certain expenses incurred by volunteers while creating media, purchasing a variety of equipment useful in the production of media (building off an earlier camera grant) for the use of members of WMCZ, and providing training to members of WMCZ about how to take higher quality media. At least five of the Mediagrant ‘sub-grants’ had been funded in previous years as stand-alone grants. The Mediagrant can be viewed as a micro-grant program administered by WMCZ specifically targeted at multimedia funded by a macro-grant from the Wikimedia Foundation. In addition to the approved elements, the Mediagrant as initially envisioned by WMCZ would have included funding for one paid employee to handle the administrative details of processing micro-grants. (Since the grants program does not normally cover salary costs, this was not approved.)

The Mediagrant was posted on Meta on December 9th, 2010 by Danny B, and was marked as funded on December 22nd, 2010 by an anonymous user (probably Barry Newstead) editing from an IP address controlled by the Wikimedia Foundation. The grant was discussed (apparently extensively) off-wiki in email conversations between first Danny B and Erik Moeller and later Danny B and Barry Newstead before it was approved. No public record exists of these conversations, and much of their contents were not shared with members of WMCZ other than Danny B. Nowadays most initial grant discussion takes place on-wiki, so this situation seems unlikely to occur again, but it is still worth noting that the use of non-transparent channels did have a significant negative impact on the success of the Mediagrant, as described below. The grant application as posted on Meta contained a lot of information about the grant, but from talking to all involved it appears that very significant details involving the structure, purpose, and limitations of the grant were decided upon in email conversations and never posted publicly, even after the Mediagrant was formally approved. Since the Wikimedia Foundation did not have the staffing to keep a close enough eye on what was being done with grant money to be able to notice if deviations were made from the initially agreed upon conditions, Danny B, being the only member of WMCZ who was aware of the initially agreed upon conditions, was able to act as the de facto sole arbiter of what was allowable. Regardless of Danny B’s performance in this role (and to properly evaluate that would require a far greater examination than I have performed,) it is greatly undesirable to have a single individual functioning in this role, especially for such a large grant. Single points of failure are almost always undesirable in any project. In the future, this situation should never be allowed to be repeated.

The Mediagrant is structured so that its ‘eye’ must approve all requests for funding. The eye initially consisted of three people – Danny B, Milda, and Che. The administrative workload that the volunteers who formed the eye had to bear was significantly higher than they had anticipated, since they had formulated the idea of the grant assuming that a paid employee would complete much of the grunt-work involved in handing out micro-grants. Although the administrative workload of the grant was unpredicted, it was certainly predictable – especially after the idea of a paid employee was rejected. The grant should not have been approved without an adequate plan in place to deal with the administrative burden that it poses.

Although many of the constituent elements of the Mediagrant had previously been funded as separate grants, not all project leads were consulted about the idea of consolidating their projects into the Mediagrant, and the Mediagrant was not obligated to and did not in fact maintain the structures or practices of the grants that it was subsuming. Interference or perceived interference in the established practices of pre-existing grants generated a lot of bad blood and negative feelings among members of WMCZ, and resulted in some people who had previously been doing valuable programmatic work leaving the Wikimedia movement entirely. These incidents, coupled with disputes between chapter members about what was worth spending money on (and disputes about who had the authority to make such decisions) resulted in an almost year and a half delay in the progress of the grant.

The initiation of the Mediagrant essentially paralyzed Wikimedia Czech Republic for more than a year, damaging their ability to conduct valuable programmatic work and driving away productive content contributors. Petr Brož (Chmee2) described the problems associated with the Mediagrant as “two quite large mistakes from Wikimedia Foundation which almost disintegrated Czech chapter.” The blame for the situation certainly doesn’t lie solely with the Wikimedia Foundation, but the factors that contributed to this paralysis were predictable and preventable, and attention should be paid to ensure that the lessons painfully learned from the Mediagrant do not have to be learned a second time elsewhere.

After the start-up pains, WMCZ has returned to stable governance and regained its capacity to perform valuable programmatic work. The Mediagrant has begun to flow fairly smoothly and produce large quantities of media (though the administrative burden is still an unwelcome problem.) The Mediagrant has funded the costs associated with the production of more than 12,000 photos so far, and many of them are both high quality and of high priority subjects. For a decent number of the photos, it seems likely that if the Mediagrant did not exist, Commons would have no photos of the subjects. So far, slightly more than a quarter of the grant’s total budget has been consumed. Gauging whether or not this has ultimately been a good use of movement funds is out of my scope.

One other unexpected upside of the grant deserves mention – in an effort to make handling the administrative workload associated with handling micro-grants easier, Petr Novák (Che) wrote a tracking application, teh-tracker. It is open source, and available here. It has all of the significant features that one would expect micro-grant tracking software to have, and has a laudable level of transparency. Other chapters that are running or are considering running micro-grant programs would be well advised to consider using this tracker.

Mile Kiš (Wikimedia Serbia; GAC)

Mile Kiš has a lot of experience with many aspects of the Wikimedia Grants program. Mile has submitted a number of grant proposals on WMRS’s behalf (several of which were funded.) He has also sat on the Grant Advisory Committee (GAC) since July of 2012. Asaf suggested that due to the breadth of Mile’s experience with the grants program, he would likely have valuable insights into how it can be improved. Having talked with Mile, I believe that his experiences are fairly typical of someone who interacts with the grants program, so I think it is worth examining some aspects of his experiences in depth.

Grants given to Wikimedia Serbia have, by and large, been significantly successful. They have allowed Wikimedia Serbia to hire a full-time employee and establish their first central office. They’ve started to launch GLAM collaborations (including aiming to place Wikimedians-in-Residence,) with one successful collaboration so far, have organized a number of photo-tours and a large number of community events, and run a successful Wiki loves Monuments competition.

Grant Advisory Committee

Mile has been a member of the Grant Advisory Committee (GAC) since July of 2012. The GAC provides feedback to grant applicants about how to improve their grants, and provides recommendations to the Wikimedia Foundation about which grants should eventually be accepted. The GAC currently consists of 24 members, with an uncapped and non-exclusive membership. The GAC was originally an exclusive committee, but membership was opened up in July of 2012. Currently, about half the members of the GAC are involved with Wikimedia chapters, and about half are not.

Before talking to Mile, I had already concluded it would be good if an advertised off-wiki channel existed for community members to give potentially contentious feedback about grant applications to the Wikimedia Foundation’s grant-making team. Although as much grant-related business as possible should be conducted in public, realistically, there are situations where public communication of concerns about a grant application will cause unnecessary drama. While talking to Mile about his experiences with the GAC, the subject of off-wiki communication between GAC members came up. Mile pointed out that situations exist where conversations between members of the GAC about concerns related to a particular grant may be similarly harmful if held in public.

After talking with him more about these situations, I think that Mile and I agree that, while the GAC should aim to conduct as much of its business as possible in public, there are some situations where public communication would create unnecessary drama with no upside. In these situations, members of the GAC shouldn’t hesitate to engage in off-wiki communication. Such communication already happens, but since there is no centralized or formal means of off-wiki communication for the committee, such communication is difficult. The creation of a mailing list for the committee to allow members to engage in easier (and more transparent) off-wiki communication in situations where such communication is warranted should be considered.

Mile believes that it is important for members of the GAC to understand the grant-making strategies and goals of the Wikimedia Foundation to a greater extent than they currently do, and suggests offering (or even mandating) more extensive training to members of the GAC in support of this goal. WMF staff who are involved in the grants program do have a lot of experience with best practices in grant-making – far more than most new members of the GAC will have. Since the GAC plays a significant role in the grant-making process, it makes sense to try to find a way to share these experiences with the GAC. I would suggest that a first step towards this goal could be the fleshing out of the GAC Learnings page on Meta-Wiki. A lot of information about the WMF’s grant-making practices and experiences could benefit members of the community who are considering applying for grants, so it makes sense to have as much information as possible available to both groups.

In general, Mile found that most of the rest of the processes surrounding the GAC functioned adequately, although he was concerned that low levels of participation by GAC members on many grants might delay the timely processing of grants, and felt that all involved would benefit from more frequent communication between the GAC and the WMF.

Potential problems during the grant process

There are many points during the grant-making process that have the potential to cause harm if handled poorly. Since Mile has experience in almost all areas of the grant-making process, I decided to talk to him about a few potential problem points that I had previously identified.

Rejection of a grant

Grant applicants put large amounts of time into writing their applications without knowing whether or not they will be accepted. Some grants represent weeks – or months – of planning, and grant applicants naturally believe in the merits of their own projects. It would be impossible for all grants to be accepted, but the rejection of a grant has the potential to be demoralizing to the volunteers involved, especially if the grant involved an unusually large amount of planning, or if notice of the rejection was not handled with care.

The first grant application that Mile participated in was for a planned Wikimedia Management Congress that would’ve been a three day event with at least sixty participants aimed at helping Wikimedia chapters professionalize. The grant application was withdrawn around five weeks after it was submitted, when it became apparent that it would not be funded. Putting the application together took a huge amount of effort from Mile and other members of Wikimedia Serbia, since submitting the application required completing most of the initial planning for a large event.

Mile agrees with me that any grant rejection has the potential to be disruptive, especially when the volunteers involved had already enthusiastically committed to the project, but felt that the non-funding of the Wikimedia Management Congress grant was conveyed in a way that minimized fallout, and was handled well by both the Wikimedia Foundation and the involved volunteers. Other volunteers I have talked to who have been involved with ultimately unfunded grants have, for the most part, agreed that rejected grants are generally handled in a fashion that minimizes the chance of negative consequences. However, the potential for harm in the grant rejection process remains (and has sometimes been realized,) so vigilance around rejection procedures should be maintained.

Delays in the grant process

The approval time of grants has been substantially reduced during the existence of the grants program, with the average approval time of funded grants down to only 33.5 days in FY11-12 (the last full year for which records are available,) from a massive 145 days in the first year of the program. I have felt that this is one of the most meaningful improvements the program has made – a fast approval time lets grant applicants begin conducting their work as soon as possible, and minimizes the inconvenience the grant process poses to them.

Mile’s thoughts and experiences as a grantee match up with my own thoughts about the subject. In his experiences, enthusiasm and desire to do something among volunteers are strongest at the beginning of a project, and if major delays occur, that productive energy may be wasted instead of being successfully channeled into valuable programmatic work, especially if grant applicants are not kept up to date about the cause of any delays they experience. He feels that delays are generally handled appropriately in the grants program at the current point in time, but that due to their potential for harm, continued caution is warranted, and stresses that good communication with grant applicants is key.

Risks & Recommendations

Most of the growing pains it experienced during the programs initial years of operation have been addressed, but the Wikimedia Grants program does have a number of significant risks moving forward. Some of these risks can be fully resolved with little effort, while some of them are very difficult or impossible to even mitigate. Risks and recommendations related to the functioning of the grants program, even when addressed in other sections of this report, have been excerpted here.

Movement level risks

These risks relate to broader problems within the Wikimedia movement. They have the potential to cause significant harm to the functioning of the grant program as a whole, or even to interfere with its existence. These risks are very hard to effectively mitigate, and are very likely to occur.

Internal movement tensions are likely to in some ways limit the efficacy of the grants program

The Wikimedia movement experiences significant internal tensions that affect most areas of the Foundation’s programmatic work. Although internal tensions have so far only affected the grants program intermittently and have been of limited consequence, internal tensions have the potential to cause greater disruption moving forward. This risk is significant and is difficult to address. This risk is not at all exclusive to the grants program, but it may be greater in the grants program than in areas that do not directly involve the distribution of movement funds.

Community members who are not on good terms with the Foundation may feel discouraged from applying for grants because they believe they would not receive fair consideration, or may choose not to apply because they distrust the Foundation or believe that accepting money from the Foundation would alter the power dynamics of their relationships. (I don’t believe that there is a significant risk that the Foundation would inappropriately discriminate against community members based on previous conflicts at the present time, although this could change in the future if the WMF undergoes a substantial cultural shift.) These risks are hard to address without coming up with a way to lower movement tensions in general, which is a challenging issue. The chance of this being a problem will be affected by the community’s impression of the independence and strength of the GAC, as well as the community’s impression of the treatment that grant applicants who disagree with the WMF receive. If all goes well, it should reduce with time, as the grants program builds a longer track record of success and fair treatment.

The grants program also has the potential to inflame existing intra-community tensions. Community feedback is solicited on grant proposals, and is normally given in public. Wikimedia community discussions are often contentious, and grant discussions have occasionally been contentious in the past, and will likely occasionally be so in the future. Relationships between community members may be negatively impacted if a community member feels that their grant application has been denied due to the feedback of another community member. It is impossible to eliminate this risk while maintaining an openness of discussion and process in-line with the Wikimedia movement’s values, but it can be mitigated. When discussions become heated, Foundation staff members who deal directly with the grant application process (currently Asaf and Winifred) should not be shy to actively attempt to defuse tensions.

I would also strongly suggest accepting private feedback on grants, and making it obvious on public pages that private feedback is accepted. Public feedback has significant advantages and should be strongly encouraged as the primary feedback pathway, but some feedback, if given in a public forum, may provoke unnecessary drama. In this vein, I would also strongly suggest the creation of a private mailing list for members of the GAC, so that discussions that are necessary but are also likely to spark unnecessary drama if held in public may be held in private.

Language and cultural barriers are likely to in some ways limit the efficacy of the grants program

Currently, the grants program directs an overwhelming amount of its funds towards highly or very highly developed countries

Wikimedia projects exist in 285 languages, but only a few of these are spoken by Foundation staff members who are involved in the grants program. More of them are spoken by GAC members, but a majority of languages covered by existing Wikimedia programs are spoken by neither Foundation staff members nor members of the GAC. Grant applications in additional languages can be accepted with the use of professional translators by the WMF, but even if applications in all languages can be accepted, people are unlikely to know that the grants program exists or to feel comfortable applying for grants if information about the grants program is not available in a language they are fluent in.

This is a challenging situation to effectively and economically address, especially with the limited bandwidth of the Wikimedia movement’s volunteer translators, and with the general visibility problem the grants program suffers from. Although not ideal, as an interim solution it may be best for the Wikimedia Foundation to directly reach out to chapters and other identified partner organizations or potential partner organizations in areas of the world currently under-served by the grants program. Having the grants program listed in relevant Foundation directories may also help begin to address this problem. (An effort to do so has already been started by WMF staff.)

Breakdown of funds distributed by the WMF grants program by UN region

Program level risks

These risks have the potential to do severe harm to individual grants, the grant program’s impact in particular geographic areas, or reputational harm to the program as a whole. These risks don’t threaten the existence of the grants program as a whole, but may be financially costly. Most risks in this category can be significantly mitigated, although most cannot be completely eliminated.

Financial malfeasance on the part of a grantee may damage the impact, reputation, or functionality of the grants program

The risk of significant financial malfeasance from major grantees should be low, since the amount of money given to a grantee is generally commensurate not only with their planned work but also with their previous track record of success. As long as WMF staff continue to make reasonable assessments of the track record of grantees, then the overall risk of meaningful financial malfeasance should continue to be low. However, this risk cannot completely be eliminated. There are several major ways by which this risk could be realized.

In the first scenario, WMF staff would give a large grant (that is then misused) to an organization without first making sure that the organization has the necessary infrastructure and culture in place to prevent its misuse. As long as WMF staff continue to take into account the track record, reputation, and organizational structures of entities applying for large grants, this will probably not occur. I believe that the grants program (and the WMF as a whole) has strong enough practices and leadership in place that, barring a large cultural shift, this is unlikely to occur.

In the second scenario, WMF staff would give a large grant to an organization that, although sufficiently vetted, then misuses the money. This is a more substantial risk than the first scenario, and one that is hard to combat. Financial malfeasance can occur in long-standing organizations with previous track records of success. The current practices of the grants program (which include, among other things, paying attention to leadership transitions and apparent organizational instability) will mitigate the risk of significant financial malfeasance occurring, but cannot eliminate it.

Given past experiences in other areas, it is likely that one of the forms of financial malfeasance most likely to exist among established volunteer organizations is the existence of conflicts of interest between the roles a member plays in the organization and their day job, or nepotism in the disbursal of grant funds. The appearance of unaddressed financial malfeasance or corruption on behalf of a grantee, even if it doesn’t actually exist, could be almost as damaging to the grants program as actual unaddressed financial malfeasance or corruption.

Avoid single points of failure (especially in communication)

There have been multiple situations in the past where the success of a grant has been materially harmed by the fact that important information was conveyed by the Wikimedia Foundation to one person at a grantee organization using email or other nontransparent methods of communication and that important information has not been available to any other members of the grantee organization. There have also been multiple situations in the past where confusion has arisen because one staff member of the Wikimedia Foundation has said something or agreed to something (such as a grant reallocation or deviation) using email or other nontransparent methods of communication and this important information has not been available to other staff members in the future. Single points of failure in any system are undesirable, and wherever possible grants should be structured so as to avoid them.

Where possible, grants should be structured in ways that guarantee that the single point failures in communication that have occurred in the past do not occur in the future. One way to do this is by having conversations on-wiki, although there are many grant related situations where this is not possible.

Where the use of email to a grantee organization is required, at least two members of the grantee organization should be included on all emails unless there is a highly compelling reason not to do so. This is true even in situations where the majority of work on a grant is expected to be completed by one individual – no matter how reliable individual people may be, things happen. If personal circumstances force someone to step back from a grant they have been managing with little or no warning, no critical information should be irretrievably stuck in their inbox. (Obviously, this problem may not be surmountable when dealing with grants given to individuals.) Similarly, since future accessibility of email sent to Wikimedia Foundation staff members has proven problematic in the past, the Foundation grants team should consider creating an archival @wikimedia.org email address and including it on the CC list of all grant related emails.

Encourage grant applicants to structure their grants so as to minimize administrative burden

Some grants have created significant administrative burdens on their recipients that were not anticipated ahead of time. In numerous cases, these unanticipated burdens negatively affected the outcome of grants. In several cases, unanticipated administrative burdens were so significant so as to negatively impact the functioning of an organization – in one case, effectively stopping a chapter’s programmatic work for a year and a half. WMF staff who are experienced grant-makers will often be better able to predict these administrative burdens than volunteers who may never have previously been involved in projects of similar scope. WMF staff should be especially careful to evaluate the potential changes in administrative workload that can be caused by partially funded grants, since grant recipients may be less likely to carefully do so. In at least one instance, the partial funding of a grant resulted in a chapter’s functioning being significantly negatively impacted.

WMF staff should (and have been) pay significantly more attention to the potential administrative burdens of projects than they did during the early years of the program. For complex grants (or grants that seem likely to have an unusually high administrative overhead,) WMF staff should explicitly convey to grant applicants how significant the administrative burden associated with a grant is likely to be, and where possible, WMF staff should make suggestions to grant applicants about how to mitigate these burdens. WMF staff should reject grants that have unavoidable administrative burdens that outweigh the benefits the grant would present. WMF staff should also consider rejecting grants that have unusually significant administrative burdens unless the grant is coming from an organization with a proven track record of managing significant administrative burdens.

Ensure grantees have a clear and consistent point of contact with the Foundation

In the past, members of grant-receiving organizations have not always known who at the Foundation to contact when situations come up. Grantees should always know who they should contact at the Foundation about grant issues. When possible, Foundation staff members contacted by community members who have questions about grant issues should refer them to the designated contact person for their grants, rather than trying to handle the issues themselves. For grants given to organizations (especially large grants,) Foundation staff members should always ensure that multiple relevant members of the grantee organization are aware of how to contact them and know that contacting them is welcomed if needed.

Take action to increase the visibility of the grants program

Across its more than three years of existence, fewer than one hundred entities have ever tried to make use of the grants program. Although grantees have been a remarkably diverse group, they represent only a relatively small fraction of the Wikimedia community. There are certainly members of the Wikimedia movement who are doing valuable programmatic work that could be facilitated by the grants program who are unaware of the existence of the grants program.

The grants program’s visibility among chapters is excellent, and the program’s visibility among non-chapter affiliated movement members who are heavily involved in meta work is also high, though neither are perfect. The program’s visibility among Wikimedia movement members who are neither affiliated with chapters nor heavily involved in meta work is very low. Taking action to increase the visibility and accessibility of the grants program among this group of movement members is a necessary and worthwhile task. Including more longtime movement members or groups of movement members (such as Wikiprojects) with established reputations and track records could greatly enhance the impact of the grants program.

Announcements should be posted about the grants program on the Village Pumps (or equivalent community forums) of all established Wikimedia projects (written in the language of the project whenever practicable.) A CentralNotice about the program may also be worth exploring – possibly a brief-running targeted CentralNotice repeated a few times a year.

Announcements should also be made regarding the grants program via Wikimedia-l and other high traffic mailing lists, stressing that non-chapter entities are eligible for grants. Most readers of Wikimedia-l probably already have at least a passing familiarity with the program, but since an additional announcement requires essentially no additional investment of resources, it would be worthwhile anyway.

Encourage grant applicants to use falsifiable success metrics

Many grant applications do include suggestions on how the success of the grant can be measured, but unfortunately the suggested indicators of success are often not falsifiable. Although it is often hard to measure the success of a project, being able to do so wherever possible is greatly desirable. Being able to evaluate the success of projects funded through the WMF grants program will make it easier for the grants program to learn from experience and improve its future performance. Chapters would also benefit from having an increased ability to self-evaluate. There are certainly grant projects that are worth funding whose outcomes may be hard or impossible to measure, but evaluation should be attempted wherever possible, and it may be a good idea to presumptively require falsifiable metrics of success for grants over a certain level (say, $5,000) absent strongly compelling reasons on a case by case basis.

Grant level risks

These risks have the potential to harm the outcome of individual grants or stop individual grants from reaching their full potential, but are unlikely to cause harm to the broader grants program. They have potential financial costs, but these costs are unlikely to be incredibly significant. Many of these risks can be mitigated or eliminated, but staff time will probably present a limiting factor in addressing them.

Strongly encourage all grant applicants to consolidate their applications wherever possible

There have been situations where the same grant applicant has applied for multiple grants simultaneously for activities that were related or interdependent. This can pose problems, especially in situations where the status of one grant may affect the budget (or feasibility) of another grant. It would generally be preferable if grant applications from the same entity during the same timeframe were consolidated as one grant if the proposed grant activities are closely related or are at all interdependent. With consolidated grants, applicants can make any necessary adjustments to their planning if all portions of the grant are not approved during the application process, and not after the disbursal of funds. In situations where multiple related grant applications are being considered at the same time from the same applicant, WMF staff should strongly encourage their consolidation. If, for whatever reason, the applicant doesn’t want to consolidate the grants, WMF staff should be mindful of potential effects of approving some but not all of the grants, and should actively solicit feedback from the grantees before approving or denying a set of grants.

Be cautious about equipment-focused grants

There is nothing categorically wrong with the idea of using grant money to buy equipment. However, the resources of the grants program (and of the Wikimedia Foundation) are limited, and grant money should be directed towards areas where it will have the most impact. Due to the drawbacks inherent in purchases of physical equipment, this will not normally include grants that mostly cover physical equipment. Grants intended to primarily buy equipment should be made cautiously, and only in situations where careful evaluation suggests that the value offered by the grant is significant enough as to be worth the costs.

The impact of equipment is generally limited, in large part due to its physical nature. It must be maintained, has a limited lifespan, is difficult to share between people who live in different places, and also poses special problems for accountability. The types of contributions made to the Wikimedia movement by expensive equipment are not normally irreplaceable; it is often possible to obtain the same quality of contribution through other means, such as by finding an existing member of the Wikimedia movement who has a similar piece of equipment, or by attracting a new member who does. The infrastructure required to manage the distribution of equipment can pose a significant burden on chapters or other equipment-holding organizations (in some cases, one that can have negative effects on their ability to function well.) And only rarely will the contributions allowed by a primarily equipment grant be of a nature that allows them to be leveraged as heavily as something like the results of Wiki Loves Monuments can be.

Some grants that are primarily equipment focused are worthwhile, which is why I am not recommending their categorical prohibition. The original WMCZ camera grant, which generated a large number of high quality photos for a relatively modest cost (many of which are of subjects not easily replaceable,) is an example of a worthwhile equipment-focused grant. An equipment-focused grant request that seems to present strong value should be considered, as long as it includes a plan for how the recipient organization will manage the problems inherent to such grants. Barring exceptional situations, equipment grants should not be given to individuals. (An equipment-focused grant application that seems to present extraordinary value made by a long-established Wikimedian operating in an area with no established Wikimedia affiliate organizations could be worth considering, but even such a grant would present substantial problems.)

The recommendations contained in this section are not intended to apply to grants that are not primarily equipment focused. Many large grants will include some amount of money for equipment. As long as the grant is not primarily focused on equipment purchases, this is not a problem – especially for organizations that are members of or have long term relationships with the Wikimedia movement.

Oversights

This section essentially consists of easily fixable problems overlooked in the management of the program to date. Some of the risks in this section have caused significant localized harm to the program in the past, and some have the potential to cause broad harm to the program in the future, but all risks in this section can be easily eliminated.

Fully protect grant applications after they have been funded

In a small number of instances, grant recipients made changes to their initial grant applications after they had already been approved and disbursed. In some cases these changes included substantial changes to the initial budgets of grants, and WMF staff were not always informed or notified of these changes. There are very few reasons why it would be necessary to make changes to a grant application after the grant has been funded; approved grant applications serve as a record of what the application was, not a record of the ongoing project. Most of the changes made to already approved grant applications would be better noted in the grant report than edited into the application. There is substantial value in casual observers, GAC members, and WMF staff being able to make the assumption that grant applications reflect the state of the grant when it was approved. Although a grant’s full history is always available in the page’s history tab, not all viewers will think to check the current state of the grant against its state at approval. WMF staff members should apply full protection to grant applications as soon as they change their status to “funded.” If a grant reallocation or significant budget change is approved by WMF, a WMF staffer should edit through the protection to make a note of this on the bottom of the grant application page. The changelog on the bottom of the page should serve as the authoritative record of WMF-approved changes. This should be communicated to all new grant recipients. (Using the grant application changelog as the authoritative record of approved changes would have the added benefit of avoiding any confusion about whether or not a change has been approved by WMF, which is an issue that has occurred in the past on at least several occasions.)

Require grant reports to be entirely hosted on Meta, or another Foundation site

In some instances, grantees have submitted grant reports that were hosted partially on non-Foundation websites. Grant reports for WMF-funded grants should be available for at least as long as WMF exists, and this may not always be true for reports hosted on external websites. WMF should require that all grant reports be hosted on Meta, or on another suitable Foundation site.

Require financial reports to use a consistent and easily readable top level format

Grant reports all include some form of financial report, but the format of it has varied widely. In some instances, financial reports have not included a sufficient amount of information. In others, financial reports have disclosed so much information that their usability suffered. WMF staff should produce a basic and easily readable top level format financial report that covers all information that WMF wants to universally collect. Grantees who want to provide more than the top level financial details should be encouraged – transparency is one of the key principles of the Wikimedia movement. For the sake of legibility and usability, grantees who want to provide more than top level details should still provide top level details in the same format, and should provide additional information in later sections, subpages, or collapsed sections of their reports.

Changelog

Some changes may be made post-publication. Grammatical changes will be made by simply replacing text. Minor content changes will be made by replacing text and noting the change here, major changes will be made by striking through the original text and noting the change here.

• Added a y-axis label to the time to funding bar graph; made a minor wording change in its caption where it appears here. Kevin (talk) 16:22, 14 March 2013 (UTC)
• Changed the header of the Wikimedia Serbia section so that it better reflects its content. Kevin (talk) 19:49, 15 March 2013 (UTC)