⇓ More from ICTworks

How Can Open Source Software be Sustainable in International Development?

By Wayan Vota on February 11, 2016

open-source-software

Recently Chris Fabian made the very interesting declaration. Since the UN’s charter is to produce public goods, he says “we do everything Open Source and Public Domain at Unicef Innovation because it’s 2016,” and doesn’t invest his organization’s resources in proprietary software.

Previous, Karl Brown made a similar declaration in his 10 Theses on Power and Efficacy of ICT4D Indulgences. His seventh thesis states that “NGOs funded with public funds should not be writing proprietary software.” Both Chris and Karl were critical in the development of the Digital Development Principles and pushing for Principle 6: Use Open Data, Open Standards, Open Source, and Open Innovation.

Principle 6 has helped convince a whole cadre of techies and aid workers that international development programs should only be investing in Open Source technologies. Enough that we can say that the international development community is now moving towards Open Source solutions.

Open Source Costs

However, Open Source should not be thought of as “free” software. Developing and maintaining Open Source solutions does require significant resources. Like proprietary code, someone needs to develop and maintain the code, maintain version control, keep documentation updated, train and support new users and developers, and market the product to others.

In fact, the cost of building and maintaining an Open Source product is about the same as creating and maintaining a proprietary one.

How Can Open Source be Sustainable?

That brings us to one small problem with Open Source technologies: making them financially sustainable. Building a good software product in alignment with Principle 4: Build for Sustainability needs more than great code; it needs a business model for its full lifecycle.

As far as I can tell, there are only three Open Source business models that work in the international development context:

  • Get project funding to write good, documented code that you hand over to an organization that then maintains and grows the code with dedicated internal resources. See RapidSMS. Unfortunately, successes like RapidSMS are rare, and this usually means developers have to jump from project to project, building new solutions for each client, and GitHub becomes cluttered with abandonware when projects end and clients move on.
  • Build proprietary interfaces to Open Source software and sell those services to organizations that don’t have the resources to do it themselves, and hopefully support the Open Source code base in the process. TextIt did this with RapidSMS, and many, many companies do this with Open Data Kit. Yet, these are not true Open Source solutions, so they don’t fit Chris’ and Karl’s declarations.
  • Build a complex solution that while technically is Open Source, anyone who wants to use it still needs to hire you to customize and configure it for them. See DHIS2. This is the model that most Open Source software efforts aspire to, but very, very few achieve. A major challenge to this model is that you need to build a market for the Open Source product and for your consulting services, when the product itself may be immaterial to your clients.

Of the three models, I see the first one as the greatest risk to international development. As Herb Caudill recently pointed out in The Revolution Will Not Be Open Source:

The world of ICT4D has generated lots of abandonware – software projects that fizzled as soon as the donor’s attention shifted and the free money ran out. But there’s nothing more sustainable in the long term than a profitable company providing a service that people are willing to pay for.

In fact, Herb went farther in his post to make an interesting prediction that should make us question our love of Open Source software. He says:

The ICT4D revolution will not be Open Source, but it will be created by scrappy startups like Magpi, iFormBuilder, TechChange, Souktel, and DevResults, by people who are unapologetic about building a business for the long term, and who have bet their very livelihoods on their ability to create proprietary tools that people will pay for.

And the issue of business models for Open Source solutions isn’t just in the international development space. Even commercial companies are finding that Open Source is not a business model on its own.

Join the Debate

As you can see, the tension between Principle 6: Use Open Source and Principle 4: Build for Sustainability is multifaceted and nuanced, which creates lively debates online and in-person.

Either way, please help us think through how to make sure Open Source software solutions are sustainable. We do not need any more abandonware – be it Open Source or proprietary.

Filed Under: Software
More About: , , , , , , , , , , , ,

Written by
Wayan Vota co-founded ICTworks. He also co-founded Technology Salon, MERL Tech, ICTforAg, ICT4Djobs, ICT4Drinks, JadedAid, Kurante, OLPC News and a few other things. Opinions expressed here are his own and do not reflect the position of his employer, any of its entities, or any ICTWorks sponsor.
Stay Current with ICTworksGet Regular Updates via Email

47 Comments to “How Can Open Source Software be Sustainable in International Development?”

  1. Absolutely agree with Herb’s point about scrappy start-ups. Like the cases he mentioned, I’ve dedicated my life — and my personal finances — to building a sustainable business that furthers social aims. Our product, SurveyCTO, builds upon and contributes to open source, but it is proprietary because it is funded by users, not by donors. Every user pays a little bit to unlock a huge amount of value, and we remain agile and market-driven. Everybody wins. This idea that ICT4D has to be open-source sounds nice, but it’s not only hogwash in practice — it’s downright dangerous. It has caused countless well-meaning organizations to blow countless millions in building (and re-building, and re-building) tech that goes nowhere and does almost nobody any good. The private sector can be a force for tremendous good in the technology space, and the sooner the donor world comes to understand that, the better off the ICT4D space will be.

  2. Mike McKay says:

    I think RTI International’s Tangerine (http://tangerinecentral.org/) is a nice example of a mostly successful open source model (disclaimer: I wrote the first version and fought with the lawyers to get it open sourced). 35 different organizations have used it in 55 countries. It’s complex, but we have managed to get pull requests from other organizations and integrated their code. The challenge mostly comes from its success – where do we get the money to support so many different users or what do we do when a user needs to deploy their own instance to the cloud, but they don’t know what that means? Tangerine is currently developing a subscription model for support and hosting to fill this gap.

    OpenMRS is another good example and probably category #1, with the major difference being that it involved many organizations form almost the very beginning. It’s been a collaboration of many public health organizations & universities (PIH, MVP, Univ of Indiana, and many many others) putting their own resources into it. It’s been operational for more than 10 years in what looks like about 75 deployments https://atlas.openmrs.org/.

    Finally, a lot of what we do in ICT4D is very context sensitive and domain specific. So to fit a solution from one project into another is a bad idea (we have THE solution for education – it must be used everywhere!). Historically, the real added value in open source projects is not in shrink wrapped products but rather in tools that can be combined together to do great stuff (web servers, encryption tools, scripting languages, operating systems). A lot of value could be added by encouraging us in ICT4D to share and use tools (versus full on products) for things that we are building. For example, our domain (ICT4D) tends to see more problem with unreliable connectivity than typical Silicon Valley startups. Open Source databases like PouchDB are great solutions for partially connected data collection – but ICT4Ders aren’t working together to optimize it for our typical use cases (how about p2p syncing?). What about open source tools or templates for cranking out the sort of reports that donors always want? How about a tool to interface with Whatsapp? What little solutions have we all written that should be pulled out into a standalone open source tool that someone else can integrate or build upon for their larger product/project?

  3. This new orthodoxy around open source, a concept that the aid sector only vaguely understands, actually stifles innovation. With all due respect to donors like Chris and Carl, they are expecting startups to focus on ideological principles rather than the single thing that matters to a new company: product/market fit. At a startup, you don’t have time to build an open source governance structure, vote on every decision, and hope that volunteers complete their work on time. In fact, this is why the large open-source projects in our sector, ODK and DHIS, are buggy, innovate at a glacial pace, and still require IT staff to operate a single instance of the product.

    Anyone who is thinking about imposing a rigid open-source doctrine for projects, as UNICEF seems to be drifting toward, should read Herb’s entire post. As he points out, there are almost no private sector companies that make user-facing software that is open source. Yet all of the ‘proprietary’ software firms in ICT4D, who are meeting our sector’s other sacred tenet of being financially sustainable, extensively use and contribute to open source code. To avoid a new era of abandoned ICT4D open source projects, we need a more nuanced approach.

  4. Matt says:

    While we are distinguising between open source and free, let’s also make sure we are not conflating open-source and charity. Aptivate (www.aptivate.org) are not a charity, yet all our software is open source.

    Anyway, I think there is an interesting dimension to this that is rarely mentioned.

    The tension appears to be between an ethical drive on the one hand (“public money should be used to crate public resources (in this case, that translates to open source), not enable individual companies to get rich”) and a practical drive on the other hand (“private companies with proprietary software are more likely to be sustainable”).

    Let’s leave aside for a moment what we think of these two drives (I have enormous problems with the latter as most small software organisations go bankrupt – proprietary or open source or otherwise – as do most small businesses in most sectors!).

    What is implied but I have never yet heard is the obvious result of bringing these two together – How can we use public money to make open-source software more sustainable.

    I don’t have an answer to this although I have some ideas. But I would really love a group of software-people and development-practtioners and ICT4D folk to discuss this and come up with some recommendations!

    • Wayan Vota says:

      I really like your point here about it being philosophy vs. practicality”

      The tension appears to be between an ethical drive on the one hand (public money should be used to create public resources) and a practical drive on the other hand (private companies with proprietary software are more likely to be sustainable).

      • James BonTempo says:

        I’ve long maintained that, given that the cost difference—and I mean the total cost, if we’re being honest—between open source & proprietary can be a complete wash, the decision to go one way or the other is largely a philosophical one.

  5. Wayan Vota says:

    And then there is this:

    Gender bias in open source: Pull request acceptance of women versus men
    This paper presents the largest study to date on gender bias, where we compare acceptance rates of contributions from men versus women in an open source software community. Surprisingly, our results show that women’s contributions tend to be accepted more often than men’s. However, when a woman’s gender is identifiable, they are rejected more often. Our results suggest that although women on GitHub may be more competent overall, bias against them exists nonetheless.

    • Matt says:

      That’s really interesting will go read more on this.

      It reflects a more general bias in technology and business (there have been experiments done in the UK exploring bias against both gender and race where identical CVs have been submitted with names changes and – surprise surprise – “John” is more successful than “Wendy” or “Adewale” with identical CV/resumes). [I don’t still have the references I’m afraid!]

      One would naively hope that the open source sector would be a bit more enlightened but it seems not! :\

      • Wayan Vota says:

        Agreed. I don’t think OS projects are any more (or less) biased than the technology sector in general, though one would hope more gender welcoming.

  6. Jon says:

    While I strongly support building a business model around your innovation, I do not believe that finding a path to profit (or even simply stable revenues) is the only solution to sustainability. There are tons of software-as-a-service platforms that exist in tandem with providing their code as open source, from Ubuntu to Docker to Drupal and WordPress. There are entire industries that thrive because of these open source solutions – webhosting would be an even more horrible place (and used to be) if every site was selling you on their custom CMS, which would lock you in forever with their platform. Today, if you want to migrate WordPress from one webhost to another, it’s something you can do over lunch.

    But this is beside my actual point – as Chris says, it’s 2016. Are we going to continue to struggle against collaboration, against sharing best practices? Are we going to burn donor money by rebuilding everything from scratch again and again, and then keeping it proprietary? Open cource code also means that we can build and extend shared platforms, creating better complete systems while pursuing our own diverse project goals. It means leveraging huge communities of developers around the world who are providing volunteer or professional time for free to a project, and enjoying benefits of scale in bug-hunting, usage, and even translation.

    And then if your project’s funding dries up, or even if your business model peters out – your contributions to the space live on. For me, that’s sustainability.

  7. Wayan Vota says:

    Here is an interesting and related post from Sean McDonald of FrontlineSMS

    Frontline and the missing middle mile
    For us, open source started as a way for us to lower barriers to access and to signal our intentions to the broader community. Unlike many others, we weren’t trying to peer into user data or pursue a commercial agenda, we were trying to build something truly useful. The lesson that we learned, though, was that once you prove your tool is useful, you have to quickly begin prioritizing:
    – Are you serving the open source community, comprised mainly of developers?
    – Or, do you serve non-technical end users, who prove value but don’t contribute code?
    – Is your priority growing your open platform?
    – Or, is it proving that it’s useful to people without technical skills?

  8. Greg Bloom says:

    Aren’t successes rare, period?

    It’s helpful to point out that open source does not obviate the need for a business model.

    Yet Herb’s post (and, by proxy, this one) does not do this. Rather, it comes close to suggesting that sustainability, private property, and rent extraction are all inextricably linked. I disagree. Reasonable people disagree. Yet, I would rather someone outright make that case outright, rather than wrap a load of self-justification within a blanket of concern-trolling.

    The more useful inquiry, i think, is to recognize that the rent-extraction paradigm is sustainable for a small number of people who stand to benefit from it, while another paradigm — a public good paradigm — is still struggling to fully emerge, having already demonstrated some real, meaningful, transformative successes. How many more successes might come from clear thinking and right action by people who are seriously grappling with these dilemmas for the purpose of constructing actual public goods that are actually public and actually good?

    • Herb Caudill says:

      It’s self-righteous language like this that drives me crazy about the current ICT4D orthodoxy around open source. If you’re taking public money in exchange for consulting hours or training or hosting, you’re “building a public good”. If you’re just selling software because that’s what you have to contribute, you’re “extracting rent”.

      We’re all trying to make a living. If you’re able to create a long-term business model around open-source software, more power to you. I’d just point out that my business model has the advantage of being very straightforward and doesn’t obfuscate how the software is actually being paid for. It also ensures that I’m accountable to my users for building what they want and need, and not what a third-party donor is willing to fund. And a business model that involves selling software can, if it scales sufficiently to drive prices down, be a lot more affordable than a business model that involves a never-ending supply of imported consultants.

      I think it’s time for us all to come down from this high horse and recognize that different models work well in different contexts. This is a practical question, not a moral one.

      • Greg Bloom says:

        I don’t know which language you’re referring to. The part about reasonable people being able to disagree?

        I think we could probably reframe along two points:

        “Despite the fact that many for-profit companies have been known to screw their users, hurt the environment, and stifle innovation across their fields at large, *some* can be constructive forces for good for some people in some contexts. Meanwhile, despite the fact that the large majority of open source projects fail, *some* can succeed under certain circumstances, generating a tremendous benefit to society.”

        Then we could have conversations along these lines: in what kinds of contexts can for-profit companies be best held accountable for undertaking public interest work? And what kinds of models (charitable and/or commercial) can sustainably produce open source software?

        Those would be useful conversations.

      • Greg Bloom says:

        I wrote a response here and submitted it but it doesn’t seem to have appeared. Text below:

        I don’t know which language you’re referring to. The part about reasonable people being able to disagree?

        I think we could probably reframe along two points:

        “Despite the fact that many for-profit companies have been known to screw their users, hurt the environment, and stifle innovation across their fields at large, *some* can be constructive forces for good for some people in some contexts. Meanwhile, despite the fact that the large majority of open source projects fail, *some* can succeed under certain circumstances, generating a tremendous benefit to society.”

        Then we could have conversations along these lines: in what kinds of contexts can for-profit companies be best held accountable for undertaking public interest work? And what kinds of models (charitable and/or commercial) can sustainably produce open source software?

        Those would be useful conversations.

  9. Jaume Fortuny says:

    The revolution will be open source or won’t be.
    If it is not open source, it won’t be a revolution. It would be a “submission to the market”.

    I understand that DevResults is doing what it does thanks to free software. And also by the use of “open finances” (or as Herb says “….In order to pay the bills, we’ve so far had to focus on deep-pocketed aid agencies, foundations, and international NGOs….. “).

    Open source facilitates and empowers. And it becomes a powerful tool in the development (also in Western countries, where also suffers from inequalities).

    The open source products are the essence of human collaboration where we take the best of ourselves and put our knowledge for mutual benefit. With this way to work, we are able to evolve adding value and improving all together. And, in this way, the funds applied can achieve a broader and more ethical “Return On Investment” for development.

    The mistake is to think that the products have to be “closed” and must have a price for being sustainable.
    (In one other aspect we could think about which should be the price: The price that supports the operating costs? The price for financing other actions also? The price of the personal enrichment of the creators?)

    But then, how can we achieve sustainability?

    First, looking for sustainability through services: setting up, commissioning, maintaining, security, etc. The examples can be found in successful pairs like: wordpress.org/wordpress.com, Ubuntu/Canonical, … (Open and Free Solution/Payed Service)

    And second: through policies that force companies and individuals to go that route. Until now, when we present projects to be financed with public funds (UN, National Aid and Development Agencies, NGOs, donors …) we can read the clauses on gender equality, non-discrimination by race or religion, the existence of certain partnerships, etc. …. Here is where we should see clauses with requirements on the use of open source technology for the pursued solutions. And the resulting solutions should remain free and open to the use of others because of these clauses.

    Ah! And just in case. Let’s force to “Open the Data” which is also is moved thanks to public funds.

    Revolutions have always gone against the powers and the rules. And “Open Source” does it too.

  10. Mike says:

    I’m with Herb on this one. ‘Open source’, ‘proprietary’ – the right tool and approach for the right job. As a practical matter most everything will be *both* going forward in some way, although the admixture will vary: your products and services will as a practical matter be a result of the utilization of a combination of such tools and approaches. No need to preferentially regard (or denigrate!) either approach – all that really matters in the end is that you meet the needs of your users. Usability trumps ideology in my book. How you get there might be important, but getting there is what matters most in the end, no?

  11. Agree with Mike and others who suggest that open source can be a part of the solution. Disagree with those who insist that it be 100% open source (or “all open source or nothing”). The idea of wrapping difficult-to-use technical solutions with services that generate revenue, okay, that’s great for consultants. And the aid community is full of lots of consultants, so that’s fair enough. But if you want to empower non-technical users to be able to accomplish their work without technical gatekeepers and scads of consultants slowing them down at every turn, then I think that you need to have professional vendors who invest seriously in solving problems — and then get to charge something for doing so. And look, if your beef is with private enterprise, then look at other economic systems (more command-and-control, communist, etc.) that have fared much more poorly in terms of producing social value. This probably isn’t the place to debate the relative merits of different economic systems, but the more ideological responses do seem to slip in that direction (the evils of profits, etc.).

    • Freso says:

      »But if you want to empower non-technical users to be able to accomplish their work without technical gatekeepers and scads of consultants slowing them down at every turn, then I think that you need to have professional vendors who invest seriously in solving problems — and then get to charge something for doing so.«

      This is where you’re missing the point, AFAICT. There’s no reason that those professional vendors you talk about should not be producing open source software, and be paid to produce that. In fact, that’s how a lot of successful open source software projects work. Drupal has tons of small to medium size private businesses around the world making Drupal solutions for other people/businesses/organisations, and they get paid (and get paid well, last I checked), but the solutions they deliver to their customers are open source. And since they get paid to work on this, their customers indirectly pay for the maintenance and continued development of Drupal itself as well. This is probably one of the most common open source business models, but for whatever reason wasn’t mentioned in this article. I would also say it’s the most sustainable one.

      As *was* said in the article though: open source ≠ free (as in pizza).

      • James BonTempo says:

        Wouldn’t the customized solutions developed using/on top of Drupal only benefit the larger community if they were regularly contributed back as enhancements to the core (that were accepted) or as new &/or updated modules? In my experience, it is not uncommon for things to be built on Drupal & never be contributed back or shared w/the community at-large (oftentimes they’re too niche & specific to be of broader use). So to suggest that all work that consultants do using Drupal helps to further its development and ensure its sustainability doesn’t seem quite right to me—some (most?) of it remains hidden within the walls of the organizations that paid for it.

  12. Greg Bloom says:

    I’m finding Nadia Eghbal to be a helpful corrective to this conversation: https://medium.com/@nayafia/

    • I was thinking I had just read something great on Medium about this and you found it here! The main points I took from it are:
      1. Successful open source projects tend to be developer tools (think Node.js, React.js, Docker) that requires either a well-funded sponsor to support it (like Facebook -> React) or lots of developers who contribute and are funded by the paid work they already do for their company, which tend to be in the proprietary software business (which is to say, 99% of the software business).
      2. The internet relies more on open source infrastructure than ever before, but fewer people are contributing code. It would be much better if donors who wanted more interoperability between platforms would fund a high quality, shared infrastructure or data exchange standards. Or at least support organizations like ours who are trying to do this all on our own.

  13. Referring to open source as though it’s one thing oversimplifies the debate. A lot of very smart people have been writing about this issue – in particular, the difference between open source standards and products. In particular, I’ve found this (http://www.adamhyde.net/leadership-diversity-and-sector-change-in-open-source/) from Adam Hyde to be really insightful, especially this passage:

    “This is where open source is getting in its own way. Open source is about employing the right tools, facilitation included, to expand your developer base. Thats fine if you are building the plumbing – open source libraries, under the hood services, backend functionality etc can all be built well using this kind of method. However if you want to be part of a team that creates wide scale change through a user-facing product, then I would suggest that open source as a culture-methodology is not the way to do it. Open source is, at its heart, a techno-meritocratic culture. Great for building under the hood tools (builders building for other builders), terrible for producing products.”

    In addition, there are some open questions about whether open licensing (standards, data, code) is actually delivering on the promise of open participation. Valerie Oliphant and I wrote “The Open (Data) Market,” about exactly that. https://medium.com/@McDapper/the-open-data-market-92f9557fd63d#.te0ic6a93

    There is a lot of value in open source, no question, but the best way to make a broad impact is through tools that deliver value to the people who need them most – and those are only sometimes developers. The best way to meet the needs of a group of people is to work directly with them, and that means building business models that focus on end user value and revenue.

  14. Josh Woodard says:

    I am in the middle in terms of OS vs. proprietary. Each has their place. As Wayan notes, the answer is deeply intertwined with the issue of sustainability. I get the frustration that some people feel about donors investing in proprietary solutions given the fact that their track record is not so great. But the answer is not to kill the use of proprietary software at all in development–otherwise, you should all stop posting your project updates to Facebook–but rather to try to eliminate flawed investments into software (whether OS or proprietary).

    For instance, we want to avoid things like this:
    https://www.linkedin.com/pulse/donors-really-just-well-intentioned-crack-dealers-josh-woodard?trk=pulse_spock-articles

  15. Tony Roberts says:

    Great article Wayan. Thanks for stimulating this debate. In the context of international development I disagree with the assertion that open source should not be thought of as free. In Europe universities standardised on Moodle as the free and open source platform for online education. All European universities contribute staff time to maintain and develop the code. What could be more sustainable than that? There is no reason why international agencies like UNICEF can’t collaborate to maintain and develop free and open source apps and platforms for development use. I also dispute the claim in the article that “there’s nothing more sustainable in the long term than a profitable company providing a service that people are willing to pay for”. Whilst the fortunes of private companies fluctuate wildly markets boom and bust, universities and UN agencies endure. What could be more sustainable that that? I expand on these points here: http://appropriatingtechnology.org/?q=node/224

    • Wayan Vota says:

      Open Source isn’t free to produce, as you say so yourself in your post, therefore it should not be thought of as “free” software:

      It is not free of costs to produce and maintain of course, but all participating universities contribute staff time to maintain and develop the code. In this way the costs of maintaining the code base are socialised across multiple institutions

      Now it can be free (or not) to the end user, but that’s not what this post it about. Nor is it my assertion that profit is the only sustainability mechanism, that’s Herb’s point. Read closer next time 😉

      Overall, I agree that OS can be made sustainable by large institution buy-in. That is my very first example with UN adoption of RapidSMS, though note that none of the outside software developers who helped build it get any revenue, income, or employment from it now. So great code but they have to make a living doing something else now – the very downside to example 1.

      • Herb Caudill says:

        I think that’s a really good point about “free” – the cost is there, it’s just not covered by the users. You can pay for software with donations (in kind from the developers themselves, or from charity or government) or you can pay for it by charging the users. Depending on someone else’s money/time/goodwill is inherently a less stable model.

        I didn’t say profit is the only sustainable mechanism – read closer next time. 🙂 I said “there’s nothing more sustainable” than a profitable business. Successful companies, of course, can go out of business, but only if they’ve stopped creating value. And yes, institutions like universities and governments last (essentially) forever, but their priorities and their financial commitments don’t. And their commitments aren’t necessarily tied to whether they’re producing value for users or not.

        The most enduring open source model is the scratch-your-itch model, where there are enough *programmers* around who benefit from the software to ensure its long-term survival. But that model almost exclusively produces software that’s used by other software, or that’s used by programmers. There are successful projects in the ICT4D space that fall into that category, for example around the IATI standard – and that’s great. Where the model doesn’t do so great is with user-facing software.

        • Jon says:

          I think this is something that is changing rapidly. The “programmers making tools for programmers” narrative is all well and good, but I think it discounts a wide number of programs which are open source, and are aimed at “users” – sometimes niche, sometimes not. I’m biased, as I’m writing this on an open source web browser (chrome/chromium, but I could also be using firefox), on an open source operating system (ubuntu), where I have almost entirely open source software running to meet my day-to-day professional and personal needs.

          But I also say this as someone who’s running a project that connects open source tool developers working in the NGO space with users of their tools around the world. Developers have been crazy excited to connect directly with many of their users who they’d otherwise net be able to meet with and collaborate with them to find where there are gaps and confusing parts of their systems.

          Perhaps the problem, at least in the ICT4D space, isn’t with developers, but with grant-making organizations pushing for a feature or functionality, instead of a user-centered design process?

  16. Wayan Vota says:

    Sadly, I think developers are doing great design with the user, but that core user is seen as the grant making body or its implementing partner, not the constituents they are supposed to serve.

    Case in point: pretty much every single mobile data collection tool (OS or not) that only aggregates data up to org/donor and rarely/never back to data giver – be that a community member, teacher, health worker, etc.

    • Mike McKay says:

      By “pretty much every” I guess you are excluding every mobile data collection tool *I’ve* written. The whole focus of Tangerine is to collect data from teachers and teacher trainers and then show it to them over time so that they can track their progress (it’s also aggregated up for program managers of course). For Coconut Surveillance, our malaria surveillance tool (also open source), not only does the health worker have the data that they collected, but it is used to help them decide if the house they are visiting is in a high risk zone (based on data they have previously collected) and if it is the tool directs them to visit nearby houses. Every electronic medical record I’ve ever written uses data to give real time decision support.
      Wait a minute, you were trolling, weren’t you Wayan?
      You have a valid point in there though – most data collection systems are for M&E reporting to make charts for donors. If the developer sees those people as their user then that is what they build. That’s why you put the nerds on the plane and make them sit in the clinic in the middle of nowhere until they understand who the user is and what they are trying to do.

      • Wayan Vota says:

        As I recall, Tangerine’s first few versions tested kids on EGRA but data wasn’t visible to teachers on a per-child basis. Tangerine Class added that functionality later.

        • Gordon Cressman says:

          System-wide EGRA/EGMA assessments, continuous student assessment in the classroom, and continuous coaching to improve teaching practices are three different applications with different user cohorts. A separate version of Tangerine has been developed with and for each of these three major cohorts. They are (1) ministry of education, (2) teachers, and (3) teaching tutors. In each of these three cases, the users are consumers of the data.

    • Gordon Cressman says:

      It’s not difficult to point to exceptions, which may prove the point:

      Wireless Access for Health (WAH) in the Philippines built a broad partnership to introduce a locally developed open source EMR into rural health units, enhance the software, and scale it up. The local EMR was developed with and for local clinicians first. Beginning with four pilot health units, it’s now in well over 100. It’s now supported by a local NGO built for the purpose. The extent to which it improves M&E data for the Department of Health and USAID, a hoped for by-product, is secondary to impact on the quality of patient care. Clinicians use it because they find it useful daily.

      Coconut Surveillance in Zanzibar has been designed for the Zanzibar Malaria Elimination Program with program managers, surveillance officers, and malaria epidemiologists. The believe it is helping to prevent resurgence of the disease. The extent to which it provides researchers and PMI with better data for research and M&E is secondary to its core function of facilitating rapid and effective response and preventing outbreaks. Surveillance officers use it because they find it useful, daily.

      The Zambia Electronic Perinatal Record system in Lusaka was designed with and for local clinicians. The extent to which it provided valuable data for research helped to sustain it, but clinicians continued to use it because they found it useful, daily.

      The amazingly successful SijariEMAS patient referral system in Indonesia is another, but I’l stop there.

      I can think of many more examples designed primarily to provide data for project M&E. I’m not aware of examples in which any of these were sustained locally beyond their project function. I can also think of many, many examples, in which data and information flowed in one direction only – up the organizational hierarchy. Providing value to the data provider, and making feedback loops as small and low as possible are old ICT4D principles.

  17. It is great to see a discussion about this. I am not sure I can make it on the 29th, so I will add my contribution here.

    First of all let me make it absolutely clear that I think future software solutions will always be a mix of open source and proprietary. For example, I use a healthy mix of software on my personal computers. But the systems we build are all open source.

    When we at Akvo started working on internet based solutions for international development, nearly ten years ago now, we had a few principles we started with:

    – Everything we create, build and/or operate should be open source

    – We build only what isn’t available already, that is fit for our purpose

    – We build only Software as a Service, with apps for offline capability as appropriate

    – We have a sustainable business model

    I’ll step through the reasoning behind these principles.

    First, we believe that in the future the data collected and maintained in the systems we operate will be considered critical infrastructure for all organisations and countries that we support. We also believe that it is healthier for the organisations and countries to have infrastructure systems that are based on open source software to minimise the risk of unhealthy lock-in effects.

    Second, we believe that in the near future cloud-based solutions will by far be the preferred choice for the end-user, due to easy of maintenance, cost and speed of development. Essentially no organisations that we talk to ask us to supply our software on their premises. Nobody wants to run these types of services. They want to use them.

    Third, we made a strategic choice to only build Software-as-a-Service solutions. There is nothing wrong with wanting software to run on your own servers. For some purposes it is even better. But, we choose SaaS as we are a relatively small organisation and we want to get as much value out of the work we do as possible, and we believe the solutions we build are best provided as SaaS solutions. It is much easier for us to extend, monitor, maintain and update services when we run them in our virtual data centres, than trying to keep hundreds or thousands of installations on disparate hardware and different locations running smoothly. When scaling to working with hundreds of organisations and many thousands of users across the whole world, there is significant responsibility to provide good up-time, data reliability, backups, security etc. Which is easier to do in a SaaS system.

    This is important as we operate infrastructure data services for a dozens of countries, many large NGOs and UN organisations. We really need to operate these services professionally with as little downtime or hassle for the users as possible.

    Forth, we are a not-for-profit organisation, but we are also a not-for-loss organisation, in other words we must have a functional business model. We charge money for our services. Which is for the long-term sustainability of our organisation.

    We have chosen to charge for services from the beginning for several reasons. If nobody wants to pay for our services then we don’t actually think we provide sufficient value. We think it is important that those that use our services understand what it costs to build, maintain, run and service what we provide. We think it is unhealthy to hide these costs behind grants and provide them for “free”. We have been provided with grants, by a number of organisations and governments to build new services. One very important aspect of these grants has been that the software we build is open source and that we have a business model for the continued existence of the service. The grant climate is often fickle and we think it is healthier and less risk for everyone involved for us to operate like we do it. My colleague Peter van der wrote more about this in a blog that I recommend that you read.

    Finally we only build what isn’t available to run “off the shelf”. For example we offer some services based on WordPress. The Akvopedia knowledge portal runs on a standard Mediawiki installation. The whole Akvo platform have too many open source components to list, but it is completely built on open source and all our software is available on Github. That said, we run a fairly large and complex system, which isn’t really meant for individual organisations to download and install. We build it to scale to thousands of organisations and their users. However, if any of our users thinks that they would rather have IBM run an installation of the Akvo platform than us doing it for them, then we are of course completely happy with that. I think of it as an insurance policy for or users.

    To sum it up: we believe we have achieved a model that is healthy and can be sustainable. You don’t have to be a company to be a “scrappy startup”. This can work for a social enterprise too, like our foundation. Finally I don’t think there is a “one size fits all”, but we at Akvo think that open source software solutions have a strong place in a future, where data for country and organisational governance and the systems you maintain the data in, are considered as important as the infrastructure or other things you measure and track.

  18. Mikel says:

    Business models are built within, or create, markets. In international development, the market for technology is shaped by donors. Donors have an obligation for their funds to have the biggest impact possible, and set up projects for future success. Access to the code underlying services is one critical strategy. So is access to the underlying data. The principles are a recognition of some failures of proprietary software development to deliver lasting value in international development.

    It’s up to developers in these markets to be creative thinkers. Open source is not one thing, and you can participate in multiple ways, on multiple parts of the software stack — even Facebook contributes a lot to open source. Complaining that good UI isn’t developed in open source communities is a red herring — I’ve seen plenty of terrible proprietary UI. You have options within an open source regime — keep you UI proprietary as a service, or make your hosting service so useful, it’s easier to pay you to run it rather than yourselves.

    If you want to help a startup software developer in developing countries, encourage and strengthen open source. Leverage code and open knowledge — hard to imagine a vibrant start up industry starting without a good open source culture.

    • James BonTempo says:

      Hmmm… While it’s true that the market for International Development (capital I, capital D, International Development, driven by Official Development Assistance, as opposed to socioeconomic development that has its roots elsewhere; e.g., through purely private sector development) is shaped by donors, we’re being required to think about the sustainability of our work in the broader market. If we were only “playing” in the donor market, our sustainability plans would be focused on ways to convince those very donors that they should continue to fund our work (though we might have to “shop” the idea around every now & again if our existing donor(s) lost interest). But again, that’s not the case—we’re being asked (required?) to look beyond donor funding (wherever that might be) to support sustainability. So don’t we need to look beyond the donor market for our solution(s)? And now I’m reminded of a post Wayan wrote several years re: donor funding as a sustainability model…

      • Mikel says:

        Question is — who is we?

        Sustainability for the communities and governments that we are all trying to serve and support in international development? I think the Principles are designed with this well in mind.

        Or sustainability for business models of software developers and ICT4D advisors? Well, I think that’s up to working harder and being creative — and abandoning principles of our work is not going to help sustainability.

        • James BonTempo says:

          My simple point—though I clearly didn’t make it well!—was that the market for technology in development is shaped by much more than simply (official) donors of International Development. So I would say we need to look beyond that market & beyond those donors for models that will/could work. I’m just responding to the first two sentences of your original post, Mikel 🙂

  19. Mikel says:

    Btw, these quotes make my skin crawl….

    > The world of ICT4D has generated lots of abandonware – software projects that fizzled as soon as the donor’s attention shifted and the free money ran out. But there’s nothing more sustainable in the long term than a profitable company providing a service that people are willing to pay for.

    And there’s plenty of unprofitable companies which don’t make it. Or formerly profitable companies. There is abandonware _everywhere_: open source, proprietary, business markets, development markets. It’s the nature of software, and a good thing, that things are tried out and then lose steam. Perhaps the issue here is that ICT4D inherits too much love of hype from the tech industry.

    > The ICT4D revolution will not be Open Source, but it will be created by scrappy startups like Magpi, iFormBuilder, TechChange, Souktel, and DevResults, by people who are unapologetic about building a business for the long term, and who have bet their very livelihoods on their ability to create proprietary tools that people will pay for.

    Baloney. These orgs build and/or benefit from open source.

    Souktel: “We build open source applications or proprietary solutions for your own internal use. ”

    Let’s look at tools covered in a recent TechChange course https://www.techchange.org/online-courses/tc141-mapping-for-social-good/

    > Map your neighborhood in OpenStreetMap
    > Visualize your data with Google Fusion Tables
    > Decisions on data: OpenStreetMap and Google
    > Make beautiful maps with TileMill and MapBox
    > Track a crisis with Ushahidi

    All open source, or by companies with huge open source component (yes even G)

  20. John Mayor says:

    I really wish you wouldn’t allow Facebook, Twitter, and other commercial icons to clutter up the screen! One of the reasons I don’t use these “services”, is because of crap like this! I don’t need these icons in the way of my ability to see text, or images in a site!… nor interferring with my ability to enter text in a messaging window!… like the one I’m presently using! It’s just bloody stupid!… and probably why so many sights have little– or NO!– comments!

    Please!… no emails!

  21. John Mayor says:

    I could care less about Open Source!… but, I care very much about FREE AND OPEN SOURCE! And to acquaint your readers with the distinction between these, I would ask that they examine the URL, http://www.gnu.org/philosophy/open-source-misses-the-point.en.html.

    Thanks for your attention!

    Please!… no emails!

  22. John Mayor says:

    Inasmuch as the name of your initiative is the Open Source Initiative… and Richard Stallman and the Free Software Foundation have made CLEAR the limitations of the Open Source Movement in their work, titled, Why Open Source Misses The Point Of Free Software (see URL, http://www.gnu.org/philosophy/open-source-misses-the-point.en.html)!… I have NO INTEREST in the Open Source Initiative.

    Please!… no emails!

    Oh!… and P.S., get these stupid Icons (Facebook, Twitter, and etc.!) out of everyone’s writing area! Thanx!

  23. Karl Brown says:

    Since I was quoted at the top, I wanted to take some time to respond to some of this rich discussion with my thoughts.

    I should state at the outset that I don’t personally have any ideological preference for open source (or “free”) software, and in my daily job and life I use a combination of both. On the other hand, I should disclose that the company I work for, ThoughtWorks, releases almost all of the work we do in the global health space under open source licenses.

    In an case, I feel that in any sophisticated enterprise setting, you will end up with a combination of the two.

    One thing I have seen on several occasions, is abuse of donor (and country) funding by proprietary software companies and the use of the proprietary nature of the software to extract rent. Now, this does not apply obviously to all companies in this space, and there are certainly bad apples in the FOSS space, but one advantage of FOSS is that even if your vendor tries to screw you, you can always throw them out and get another vendor. You can’t easily do that with proprietary software. Another example was a vendor that built a system in East Africa funded by a certain large (not to be named) American bilateral. When it was proposed to move that system to another country nearby, the vendor was unwilling to do so – so you have public funds that paid for a custom solution, and that solution was not actually a public good.

    In my 10 theses article I proposed that NGOs should not be *writing* proprietary software with public funds. However, I did not say that:
    a) NGOs should never purchase proprietary software with public funds
    b) Private (for profit) firms should never develop proprietary software with public funds

    For a), for example, you could say “Well, Gates gave us money to build up our health information system, and we’d like to do analysis, we all know Excel, so can we buy licenses to Excel, even though its proprietary?” I would say “Of course!!!”

    For b), for example, you might say “Hi, I want to create a startup that is servicing this space and we want to be able to have a sustainable business, so we want to charge a license fee and keep the software proprietary” well, that seems perfectly reasonable to me as well, and several commenters above have done just that.

    What doesn’t seem to work, in my experience, is when NGOs – (and by this I really mean not for profit entities) – whose main job is programmatic and implementation work – try to create proprietary (or pseudo-open-source) software. They don’t have what it takes to run the software as proper proprietary product, so I don’t think in most cases they should. There are always exceptions, of course…

    Now, there is a very good point brought up in the article about sustainability (which is I suppose the view from the developer), and the flip side from the customer is perhaps total cost of ownership. But I don’t think it’s only FOSS that suffers from sustainability issues. Many proprietary products get shelved or forgotten or the companies running them decide to no longer support them. What I find most interesting now is the sort of hybrid approaches; freemium models, paid hosting/customization packages, support/warranty packages, etc. In any case, sustainability is not magically solved by choosing a license. There are tradeoffs and ways to do it poorly no matter what you choose.

    We all know the only constant in software is change. So, whether the software is open source or proprietary, there is always a need for money/resources to keep it going and keep it relevant to the changing environment. If you have a custom piece of software, you have to maintain it all yourself. If you have an open source product, then you participate in the community and – if the community isn’t doing what you like – you have the power to fork. If you have bought a proprietary product, you are ultimately at the whim of the vendor – if they are responsive and willing to make the changes you want – great. If not, you might have difficulties.

    In any case, when you are making a software decision, you should be thinking about all of these things – what is the initial/upfront cost, what is the long term total cost, how much control do I have over the direction of the platform, what is the product roadmap, what is the financial health of the organization supporting this work, and so on.

    Finally i just wanted to comment on the discussion above on “free” – e.g. “However, Open Source should not be thought of as “free” software.” I hope that those of us who talk about FOSS in this domain have not been guilty of conflating free as in beer with free as in freedom – but I do feel that the “free” as in “freedom” aspect of FOSS has been insufficiently covered in the discussion above, and that’s really what sets open source apart from proprietary – NOT the license cost.

    • Thanks, Karl. That was all extremely reasonable!

    • Herb Caudill says:

      Karl: That’s all very fair and well-stated. In particular, it’s very helpful to distinguish the worst-case scenario you describe (in which a donor commissions a custom software solution that then stays in the private domain) from the scenario I and others have been championing, in which bootstrapped software companies just make a living selling software. I think a lot of the heat in this conversation has come from those two scenarios being unintentionally conflated and people talking past each other as a result.