Foundations and open-source projects

Most open-source projects start off as small, volunteer-only operations. As such, the prospect of growing large enough to warrant setting up a formal governing organization may sound like a victory with little downside. Yet, in reality, establishing a governing foundation or partnering with a fiscal sponsor carries with it several complex questions about fundraising, workloads, and tax exemption. In an informative session at OSCON 2016 in Austin, a panel drawn from some of the best-known non-profit foundations in the open-source software world discussed the ins and outs that developers face when interacting with foundations.

The panel consisted of Open Source Initiative (OSI) board member Deb Bryant, Danese Cooper from the Apache Foundation and Drupal Association, Karen Sandler from the Software Freedom Conservancy (SFC), Stefano Zacchiroli from Debian and OSI, and Deb Nicholson from the Seattle GNU/Linux conference (SeaGL) and OpenHatch. Sam Ramji of the Cloud Foundry Foundation had been scheduled to participate, but sent his regrets. Bryant, who introduced and later closed the session, asked each panelist to give a brief introduction, then wanted to reserve as much time as possible for questions from the audience.

Most of the panelists are well-known enough that a formal introduction would be superfluous, but they did serve to frame the following discussion. Cooper noted that she had "been in a lot of foundations" over the years, and may have written more OSI-approved software licenses than anyone. Sandler commented that she can say from experience that non-profit foundations are decidedly non-profit, and while they look exciting, running them is hard work and surprisingly easy to mess up.

Zacchiroli noted that he serves as an OSI board member, but has also been a "customer" of open-source foundations—when he worked as Debian Project Leader. He encouraged developers to remember that the goal of their project is (likely) technical in nature. "What you really want from a foundation is for it to take over the other tasks, so that you can remain focused on the technical things." Nicholson commented that she has now worked for non-profits for over 20 years, and that one lesson she has learned is that there is rarely a good reason to "roll your own."

Dough

The first audience question was about money. Specifically, the audience member asked, how does an open-source foundation go about ensuring that it does not compromise its mission when it raises funds? Cooper replied first, saying that Wikimedia is the most successful foundation at fundraising, and it does so using direct "asks" to site visitors each December. Not everyone has such an easy option, she conceded, but added that it is important to note that Wikimedia employs a lot of A–B testing to find the right way to ask. She cautioned developers about 501(c)(6) foundations (also known as trade associations), because they rely heavily on large corporate donors who expect "pay for play" arrangements: in exchange for a large donation, the donor gets one or more seats on the governing board.

Nicholson said that the first big mistake many foundations make is not doing due diligence in the beginning to establish what their mission is. If the mission is something vague, she said, "like 'uh, make the cloud ... awesome'" there will likely be trouble down the road. The second big mistake, she added, is not diversifying the funding. As a result, "you'll always be beholden to that 40% or 90% donor." She also said that if a program comes up that the project feels strongly about, but which is too tangential to the project's primary goal, it should consider spinning that program off into a separate project to not distract from the mission.

Zacchiroli said there are big differences between corporate-funded and public-facing (in terms of membership and support) foundations, and that the community has yet to find a reliable model for funding public-facing foundations—meaning a way to consistently raise enough money for one or more full-time staffers. There are some exceptions, however; Sandler added that SFC recently pivoted from corporate to individual donors, and has been happy with the outcome. Bryant agreed, noting that OSI had restructured so that the board is no longer selected by corporate donors. "Before that, we had roughly two donors. To their credit, they stayed out of the way, but things are better now."

Staff

The next question came from Donna Benjamin, who related that she has worked with the Drupal Association and Linux Australia, two foundations that take decidedly different approaches to having paid staff. She asked the panel how a project can ensure that there is balance between the volunteers, who often have "sweat-equity" in the project, and employees who are being paid for their time.

Zacchiroli concurred that the step of hiring an employee is "the big change" for most foundations, and reiterated his advice that a foundation serves the project best when it takes on the auxiliary tasks, not the core technical tasks. Thus, "to be accountable to the project's main goal, the employees should take on the accessory jobs." Nicholson agreed, adding that asking volunteers "who wants to file the taxes?" is not likely to be successful. "Although," Cooper added, at least somewhat in jest, "that job is pretty easy when you have very little money." She went on to note that the Apache Software Foundation is the exception to the general rule, since it employs several part-time staff members.

An audience member then asked how Mozilla—which employs many developers and other paid staffers—has its corporate structure set up. That audience member's project receives frequent requests for paid support contracts, it seems, so juggling interest from businesses and the non-profit nature of the project's foundation was of particular interest. Brian Behlendorf, a member of the Mozilla board, was in the audience as well, and he replied that it is hard to generalize from Mozilla's example. "Browsers have an almost unique opportunity to monetize through the default-search-engine arrangement," he said. The healthiest approach, he said, is to have a handful of friendly consultancies that the project can point interested businesses toward. But there need to be several, he added, and they should not be too tightly tied to the foundation.

Sandler added that it is legal to form a for-profit underneath a non-profit foundation as long as it does "related work" as defined by the IRS. But, she said, too often the for-profit business starts taking up so much time that the employees no longer have time to work on the non-profit tasks. "At that point," she said, "you'd might as well just have a consultancy."

Government and governance

Next, an audience member inquired if there were ways to get funding from government grants or other government programs. Cooper replied that Wikipedia used to have a grant program, but that after it figured out "how to do direct asks right," it determined that the grant program was not worth it. "You have to have a whole grant-writing team that does nothing else, and that becomes self-justifying," she said. In addition, government grants often do not mesh well with open-source projects because they typically require the project to write proposals long before they determine what they will be working on; software development simply changes too quickly. There are a few "preferred grants," she said, in which the government agrees to work with a project on a trusted-partner basis. Those would be beneficial, since they would provide so much flexibility, but they are quite rare.

Sandler replied that she has written a lot of grant proposals, and that she agrees with the audience member that free-software projects would be a good fit for many government grants, particularly those projects who focus on technology related to issues of freedom. Nicholson pointed out that there are non-governmental organizations that offer grants, too, but that working with them can be hard. "They tend to be very old organizations and very 'legacy,'" she said. "You'll really have to get pictures of puppies and children that are sad and say 'we'll fix that'" in order to have much success.

Zacchiroli then added that the situation in Europe is rather different than that in the US. In particular, he said, many government projects are getting big enough that they can contribute code to open-source projects directly. Cooper added that Mozilla and a few other organizations within the open-source community also run grant programs, which may be an easier path to take. Bryant added that there is a lot of overhead involved in dealing with the government, and that a better approach would be to find a humanitarian organization that already works with the government and partner with it.

Another audience member then commented that non-profit foundations in the open-source community are different than those in other fields, because they act as a sort of governance body themselves—by setting norms for the community and offering guidance. With that in mind, they asked, how do the various organizations represented on the panel feel when the foundations come into disagreement?

Zacchiroli replied, asking "I guess what you are saying is that we have different organizations acting as authorities on licensing and such things?" If so, he continued, "I think it's a strength to have several organizations." Considering the amount of money floating around open source, he said, it is beneficial to have multiple organizations reinforce what constitutes "good" licensing. "It means you would have to take over multiple organizations before you could sneak a proprietary license in. I don't think of it as duplication of effort, but resilience."

The last question before time ran out was about joining an umbrella organization versus forming a standalone foundation. The audience member pointed out that the IRS has created the new "1023-EZ" form to streamline the process of getting official recognition as a tax-exempt non-profit. With that available, does it still make sense for small projects to join umbrella organizations? Nicholson replied that it can be a major problem to spin up a non-profit foundation before really knowing what form the organization needs to take. Consequently, it is better to spend a year or two under an umbrella organization to determine what works and what shape the organization should take. Sandler added that it is important to read the termination policies offered by the various umbrella organizations.

With that, the session ended. Bryant thanked the panelists and the audience alike. She noted that the group had offered a three-hour seminar version of the same panel in the past, and hoped that the shorter version proved useful as well. It was certainly a well-attended session, and there was no shortage of questions, both of which suggest that many within the broader open-source community remain hungry for real-world advice on the details of non-profit foundations.