Documenting a failure mode for NGOs making free software

There is a potential governance pitfall when NGOs develop free software.

Let's say one accepts resources to build some software. These resources might include grant money and volunteer time. Explicitly or implicitly, the resources are intended for pursuing some specific objective. In the cases I'm thinking about, this is that the widespread availability of the software will advance some goal shared by the NGO and its donors.

Keeping NGOs focused

What can go wrong in *any* NGO is that undersupervised staff try to divert the NGO's resources towards other worthy or less worthy causes; this is a problem in NGOs of any stripe. To some extent it can even be illegal, e.g., it is a breach of the trust in which the donors' funds are held, or it's ultra vires the NGO's own powers under its governance arrangements. In recent decades, managerialism has undermined the safeguards against this problem.

That's the tension in the background: it costs resources to keep the NGO's resources focused on the agreed objectives as opposed to other goals its staff might be incentivised to pursue.

Software in NGOs

But in free software, there's an additional risk. Free software is often self-hostable. Not everyone wants to self-host, so there'll be a market for a hosted version of the software. The NGO may wish to enter this market itself, as indeed may its staff.

Even if the staff had no commercial ambitions of their own, there are inevitable trade-offs between the interests of the NGO's central hosted version of the software and those hosting their own, and the NGO's staff will be in a position to put their thumbs on the scales.

This could lead to either the self-hosted version becoming a second-class citizen (e.g., poor installation documentation), or, what is less likely, that the NGO's own flagship installation of their own software isn't getting enough attention and serving as a good example of the organisation's work.