Lessons in building social sector data capability
Posted on 11 Apr 2023
Over the past four years, we experienced some glorious flights in our data-for-good program — and took a few tumbles. In this article, we explain Our Community’s project-focused approach, key lessons, and what might come next for the Australian data-for-good landscape.
- Provide more long-term technology-specific funding to help progress towards systemic change
- Support and solicit programs that can facilitate the process of finding and connecting not-for-profits with data experts
- Encourage not-for-profits to learn by doing using data resources tailor-made for the social sector
Aiming for the sky
Our vision of lifting the social-impact sector on the wings of data science has yet to be fully realised.
Not-for-profit organisations with significant in-house data capability are rare. Supporting services—volunteers, consultancies, and intermediaries—remain “wildly fractured”, difficult to find and daunting to engage. Add the scarcity of technology funding and ambient concern about the risks of data-intensive technologies, and it’s no wonder that any not-for-profit that can merely hop from the ground has cause for celebration.
In 2019, we received a grant from Equity Trustees to improve this state of affairs in the Australian not-for-profit sector. We have helped (and continue to help) tens of thousands of organisations in many facets of their operations, from governance training at the Institute of Community Directors Australia (ICDA) to low-cost fundraising with GiveNow. It was logical to extend our advice and training to data science and use our unique vantage point in the sector to connect experts with organisations in need.
Four years in four seconds
We started by community building and establishing not-for-profit data needs. Then the pandemic reshaped priorities, so we consolidated a suite of resources focused on project scoping.
Four years in forty seconds
We kicked off with a range of data science activities to gather information and iterate—not unlike the Wright brothers, who conducted numerous wind-tunnel experiments before their aviation success. We ran tutorials, one-on-one consultations, learning partnerships, and Melbourne’s first datathon-for-good (in tandem with Community Hubs Australia). We interacted with dozens of not-for-profit organisations and grew a meetup community of over a thousand people passionate about using data science to help the social sector.
At the start of 2020, we took stock of the challenges we faced. What we couldn’t have foreseen was the havoc that would be caused by COVID-19. Burnout in the not-for-profit sector magnified. Organisations that had expressed interest in leading data capability initiatives now had reduced capacity to pursue them. They had to learn how to deliver programs remotely, often amid unprecedented spikes in demand for their services.
The core mission always comes first. With that in mind, while other parts of our organisation looked after the sector’s basic needs — surviving in the pandemic era, remote work policies, etc — we developed data resources that not-for-profits could access as they rebounded.
We released a data capability framework to kickstart conversations about what not-for-profits were already doing with data, and what they could do. Then we tried to make the data project journey more approachable by breaking it into a set of discrete steps, including templates for creating a data project brief and position descriptions.
We designed new tutorials that covered project scoping and data quality — identified as key needs in our learning partnerships — and transformed the tutorial series into self-paced online courses. We continued to provide free one-on-one consultations to not-for-profits across the country. Finally, assisted by our friends at WhyHive, we ran an all-day workshop devoted to scoping data projects.
Lessons learned
Here are three things we’ve come to believe about developing data capability at scale:
Most not-for-profits need to start with a project focus
You might have noticed I didn’t use the words “maturity” or “strategy” in my summary of our approach.
Many of the organisations we support weren’t ready to tackle (or weren’t interested in tackling) ambitious activities such as building data maturity, improving data governance, or implementing privacy measures. They’re often only working at a program level or just trying to start using data more effectively in one area of the organisation.
That’s why we designed our resources with a project focus. Successfully delivering a project helps an organisation to develop its data culture in an incremental and responsive way. Rather than build an airport and manufacture a fleet, we just wanted organisations to rent a plane and get in the air to see what they could learn — and to get excited!
Scoping is an early — and high — hurdle in data projects
Even with a project focus, there’s up-front effort: forming partnerships, deciding on scope, refining questions. If these are prepared with care, the data analysis component is often quite quick.
Through dozens of consultations with not-for-profit organisations, we saw that there was no shortage of interest in building data capability, but they either didn’t know how to approach it or they had an idea of the process but wanted reassurance. As time went on, we addressed this by developing a step-by-step data project pathway. We also zeroed in on the scoping stage with a hands-on workshop.
A major benefit of clearly scoping a data project is that organisations can more easily find the right help. There’s a wellspring of data professionals seeking to use their time to make an impact. Our community-building efforts aimed to create opportunities for not-for-profits to recruit paid staff and volunteers from this pool, but this depended on having appropriate projects scoped and ready. The US-born global not-for-profit DataKind, which has codified and shared its approach, places a similar emphasis on project scoping and design prior to recruitment.
Ultimately, not-for-profits are responsible for propelling their own projects, but we wanted to provide the most favourable winds possible.
Face-to-face time cuts through uncertainty and competing priorities
After COVID-19, we adapted our in-person tutorial series to an online, self-paced format. There are obvious advantages to such resources, including that people can track their progress and pick up where they left off, which helps when data projects are often stop-start.
But people especially carve out time for video consultations and in-person workshops. One participant said it best: attending the workshop was “an opportunity to stop and spend some time thinking about the aim, goal, hurdles and solutions” that they might not have done if not for the face-to-face commitment.
Our participants also found it valuable to learn from people with the same struggles, questions and decisions as them. If you’re part of an effort to uplift data capability, remember how indispensable face-to-face interaction can be, and see what happens when you get participants in the same room.
Taking the next leap
How can the broader data science community meet the Australian social sector’s needs in future?
Connecting not-for-profits with supporting services
For starters, we could make supporting services easier to find, compare and engage.
Our resources help organisations to ideate and design a data project, but a gap emerges once a project is scoped: if it can’t be done internally with existing skills or training, there is no easy-to-use platform to find and connect with experts, be they volunteers, freelancers, or consultancies such as Clear Horizon and SVA Consulting.
One approach might be a self-service directory of data analysis, data science and impact evaluation people who want to work with not-for-profit organisations. An example of what this might look like is the Dovetail Network, a directory of digital agencies that serve UK charities, or — even more lightweight — this spreadsheet of US corporate tech-for-good initiatives. We mocked up something similar using Google Forms, aligned to our data capability framework. However, maintenance is an ongoing issue, and it ought to be designed to inspire action. If you’re considering such an initiative, get in touch and let’s talk.
Encouraging and supporting not-for-profits to learn by doing
Pro bono organisations in the mould of DataKind, Civita, Good Data Institute and Data4Good do excellent work, but they’re no replacement for in-house subject matter experts with data skills. We’ve seen hackathons and corporate volunteering produce exciting starts, too — only to stall because there is no capacity for implementation within the organisation.
Not-for-profits aren’t oblivious to this problem. In annual surveys of the sector by Infoxchange, the proportion of NFPs who considered it a priority to explore new technologies and innovations dropped from 22% in 2020 to 12% in 2022 — the worst Covid years, when the sector’s focus narrowed. Even so, NFPs expressed an increased desire to improve staff digital skills.
Common challenges in this space, such as lack of in-house capability and the inability to outbid corporates for talent, are opportunities for staff development. Embarking on a data project is a great way to learn by doing.
Many participants in our tutorials and workshops are what you could call the ‘data people’ of not-for-profits. Some have technology backgrounds, but often they end up working with data by accident or necessity. Upskilling these subject matter experts results in sustained capacity. But it doesn’t require a learning odyssey. Not-for-profit organisations shouldn’t be afraid of rapid iteration. Try and fail, celebrate small wins, build knowledge, build expertise.
There’s a continued need for free or low-cost data analysis and data science material tailor-made for the social sector. One missing topic is entry-level programming. Introductory Python materials are commonplace, but none focus on the social sector. Such a course could include exercises on topics like fundraising (e.g. donor segmentation), needs analysis from public datasets, and impact measurement (e.g. classifying survey responses). This could generate interest for not-for-profit professionals and help them apply lessons learned within their area of expertise.
More technology-specific funding to progress towards systemic change
Few grants are available for building digital and data capability in not-for-profit organisations.
If businesses and government tend to struggle to make systemic change, how can we expect not-for-profits to do it — especially if they lack the time, resources, skills or capacity to deliver data projects internally?
Systemic change begins by enabling and developing many ‘data people’ in not-for-profit organisations, who over time can band together to make good things happen with their skills and experience. We’ll also continue to need intermediaries to fill the gaps and connect the sector: more groups like Our Community, Seer Data, and Infoxchange.
Above all, we need more long-term funding to enable this work, and to develop a cohort of “public interest technologists” who can provide end-to-end skilled support. Our data-for-good activities were funded by Equity Trustees. The Paul Ramsay Foundation is another grantmaker helping to progress data capability in the Australian social sector.
Grantmaking is a powerful lever for change. Our future social-sector data efforts will be focused on embedding best practices in our SmartyGrants grants administration platform. In particular, we aim to help with outcomes, automation and risk management. Meanwhile, other parts of Our Community will take our lessons into their work with not-for-profits.
Ingenuity relies on perseverance
Set against the continuum of human flight, today’s data-for-good landscape resembles the era when we dressed in feathered cloaks and flung ourselves from the tops of mosques and mountains.
More generously, we’re closer to the invention of the box kite in 1885. After a gusting wind sent Lawrence Hargrave — pictured at the start of this article — soaring above a beach south of Sydney, he described his invention for everyone to learn from. Hargrave cells, as his box shapes became known, appeared in aeroplane designs across the Pacific, speeding the innovation that culminated in the motor-powered Wright Flyer of 1903.
One-hundred and eighteen years later, in 2021, Perseverance, a rover, landed on Mars and deployed Ingenuity, a helicopter. In a nod to its aerial lineage, NASA engineers had stowed a fragment of Wright Flyer’s fabric wing under Ingenuity’s solar panel. The helicopter’s 39-second maiden voyage was the first powered controlled extra-terrestrial flight. Hargrave was a link in that chain.
Our collective efforts in the social sector are building towards something similarly meaningful. That isn’t to say we’ll be funding charities to service any aliens we find on the red planet, but that by sharing our learnings with candour and supporting the vision of leaders in the sector, we can reach collective heights we could never have imagined by working in isolation.