There is a tendency of data for good sector to fund pilots of projects without ever funding the realization of those pilots. Instead of thinking in a structural, sustainable, long-term way about the biggest problems facing our world, we too often try to solve the world’s problems with new technologies implemented as quick fixes.

This tendency, which has us investing only in the short term with no thought or plan for taking projects forward, is called pilotitis. Pilotitis is pervasive across the social impact sector. It leaves us with projects that are orphaned or abandoned, and wastes our money and time.

Pilotitis comes from human nature. We embark with good intentions, with pilots that are designed for entirely good reasons — solving a problem, testing hypotheses, using a promising new technology, and seeing if you can be agile in your pursuit of a solution. 

But as you move out of the pilot phase and start actually pursuing your goal, you suddenly start to find difficulties getting the financial support you need. That’s because the people who invest in social impact projects tend to get caught up in pursuing short-term objectives. You might find that you and your funders will launch digital interventions together, you’ll get a great headline that generates buzz, you’ll celebrate that hype — and then your funder moves on. 

And if this happens to you over and over again, chances are you’ll move on, too. Welcome to pilotitis. 

At its worst, this tendency is not only wasting time and resources — it actually kills better work. If a number of projects are competing for attention, resources, or influence, often the shiny, short term ones win out because the teams behind them can invest all their energy in the things that make it shiny: the marketing, in the push, in making the first launch successful. You don’t need to take the cost of putting in long-term infrastructure, support and maintenance contracts — all the boring, long-term things that you’d need to make it work. 

This isn’t just a theoretical problem —there are famous cases of pilotitis that became so hairy, national governments have actually had to intervene. In 2012, Uganda had so many pilots of mobile health tech that the whole system was fundamentally overwhelmed: Every hospital, every public health body in Uganda had three, four, sometimes five approaches — often from very well-meaning, global north charities and philanthropic bodies — saying, “We can invest this in this app, or in this way of doing things.”

They had so many pilots that it actually hindered Uganda’s ability to run the basic health service in the country. So the national government put a moratorium on the development of any more mobile health projects.

I suspect the cycle is in tune with new technologies. Back in 2012, when the Ugandan government made its ruling, it was the age of the smartphone. The promise of new mobile technology and apps created this feeding frenzy, this excitement. And this hype is a big part of what instigated this problem. Today, all the hype is around artificial intelligence. This new era of pilotitis in our sector will be around AI, the next big data science.

We’ll likely never cure pilotitis completely. It’s human nature. What we can do is minimize it, and make sure it doesn’t disrupt too much. 

If we want to do better, we should start by finding something that excites us in the unsexy act of longer-term maintenance and support of good projects. Not everyone in my sector — and the people who fund my sector — knows about the Ugandan case study, so we need a better education about the consequences of getting it wrong. We also need to make it easier to fund the long tail parts of the work that makes it successful long term, like support, maintenance, documentation, testing, community building.

Pilotitis leads to duplicative work: in some cases, ten or twenty of the exact same type of project getting just a little bit of funding. To solve that, we need better relationships between funders. Our funders need to collaborate and cooperate with one another, they need to be transparent about what they’re funding so that they can identify duplicates. They need to start saying, “Instead of us each funding a slightly different version of the same thing, let’s all unite around one technology, and maybe even open source that tech so that others can benefit.”

At the moment, the sector works a lot with one-to-one programs. You get funding to do a great thing with data, and great, that project works. But there is no learning across projects, and the projects don’t build from one another’s successes, failures, breakthroughs, and findings. 

If you want to cross a river, pilotis will have you throwing stones in the water until you have enough at the bottom to be able to wade across. That could take years — if it ever happens at all. We need to be artisans. We need to work together to architect a bridge. A bridge will use far less stones. A bridge will help us get across the river much faster. 

Collaboration is key. At the moment, data.org is launching the Epiverse program for epidemiology. One of the drivers for that, for us taking up this challenge, is that in the world of epidemiology, often the tools to create the models are made locally, within the team. This means the same tools are recreated again, and again, and again at each university. We’re helping the field take a step back, allowing us to invest as a community in open-source tools, properly support the maintenance these tools require for long-term sustainability and then make these tools available for free to the community. 

To solve big problems, you need multiple disciplines involved. There is no app that can solve climate change. There is no app that can solve the COVID-19 pandemic. Some problems are just too big. So many of the important problems in our world are too big to be solved by a single discipline, which means you need to work across disciplines, bringing together experts from the physical sciences, social sciences, technology, industry, the social impact sector, policy, and government. Interdisciplinarity is the commitment to investing in and working with, people across these disciplines, creating partnerships and connections that have expertise and fluency in more than one area. 

It’s a powerful tool towards helping to fight against pilotitis because if you don’t understand the subject matter, or hold just one narrow view on the problem, you’re more likely to be mobilized by the flashy rather than the effective. 

We saw a lot of examples of taking the narrow view on the problem at the very beginning of this pandemic when there were hundreds of attempts by tech startups to build apps to diagnose Covid. But a report from the Turing Institute found that the majority of these apps didn’t work because they didn’t involve a single doctor in the development process. They didn’t take into account the kind of nuances that only a professional knows — and once again, we have another high-profile example of a proliferation of pilots that get abandoned. And this is where interdisciplinarity can help. You need interdisciplinary scholars and professionals working within teams to make them more sustainable, to help them go beyond the pilot, to be implemented, and to be successful.

Just like you need doctors to help out the techs building COVID-19 diagnostics, so the reverse is also true: getting technical expertise into the decision-making process of what gets funded is very important to help mitigate against pilotitis. But that’s easier said than done: Many funders don’t have in-house technology and data experts with the skills and expertise to advise on funding proposals.

In a way, data.org was set up to try and fix that. One of the jobs we have is working with multiple funders as their out-of-house tech experts. We provide our partners with the tech expertise to help them guide investment into their products. Bringing experts into the decision-making process will invariably steer funders away from focusing too much on pilots that don’t go anywhere, placing the focus where it actually needs to be: on sustaining long-term projects that actually solve problems.

So this is the data.org prescription for curing pilotitis: don’t be seduced by the hype, invest for the long term, collaborate across funders to unite around shared, open solutions, and build interdisciplinary teams, where technologists and subject-matter experts work through the world’s complexity together. With this approach, we can achieve the agile learning and growth that a promising new pilot provides, and build on that momentum for maximum and sustained impact.