Keywords

In February 2020, just pre-COVID, a group of managers from community organisations met with us researchers about data for social good. “We want to collaborate with data,” said one CEO. “We want to find the big community challenges, work together to fix them and monitor the change we make over ten years.” The managers created a small, pooled fund and, through the 2020–2021 COVID lockdowns, used Zoom to workshop. Together we identified organisations’ datasets, probed their strengths and weaknesses, and found ways to share and visualise data. There were early frustrations about what data was available, its ‘granularity’ and whether new insights about the community could be found, but about half-way through the project, there was a tipping point, and something changed. While still focused on discovery from visualisations comparing their data by suburb, the group started to talk about other benefits. Through drawing in staff from across their organisations, they saw how the work of departments could be integrated by using data, and they developed new confidence in using analytics techniques. Together, the organisations developed an understanding of each other’s missions and services, while developing new relationships, trust and awareness of the possibilities of collaborating to address community needs. Managers completed the pilot having codesigned an interactive Community Resilience Dashboard, which enabled them to visualise their own organisations’ data and open public data to reveal new landscapes about community financial wellbeing and social determinants of health. They agreed they also had so much more: a collective data-capable partnership, internally and across organisations, with new potential to achieve community social justice driven by data.

We use this story to signify how right now is a special—indeed critical—time for non-profit organisations and communities to build their capability to work with data. Certainly, in high-income countries, there is pressure on non-profits to operate like commercial businesses—prioritising efficiency and using data about their outputs and impacts to compete for funding. However, beyond the immediate operational horizon, non-profits can use data analytics techniques to drive community social justice and potentially impact on the institutional capability of the whole social welfare sector. Non-profits generate a lot of data but innovating with technology is not a traditional competence, and it demands infrastructure investment and specialist workforce. Given their meagre access to funding, this book examines how non-profits of different types and sizes can use data for social good and find a path to data capability. The aim is to inspire and give practical examples of how non-profits can make data useful. While there is an emerging range of novel data for social good cases around the world, the case studies featured in this book exemplify our research and developing thinking in experimental data projects with diverse non-profits that harnessed various types of data. We outline a way to gain data capability through collaborating internally across departments and with other external non-profits and skilled data analytics partners. We term this way of working collaborative data action.

By ‘data for social good’, we mean using contemporary data analytics techniques to fulfil a social mission or to address a social challenge. Data analytics is understood as the process of examining data to find patterns and insights that can aid decision-making and offer courses of action (Picciano, 2012). We define non-profits as all those organisations and community groups operating to pursue a social mission and that do not operate to make a profit. Individual non-profit organisations are thought of here as each pursuing their defined social mission, but also contributing to a collective social mission of achieving a more equitable and just society. While non-profits are often using data to track their operations and aid reporting, we emphasise the data that non-profits could use to further their work and goals. This includes mainly:

  1. a.

    internal data generated routinely from non-profits’ own operations or new data they might collect (e.g., to inform evaluation). Such data could be used, or re-used, for insights by individual non-profits or in data sharing collaboratives with other organisations and

  2. b.

    external open data generated through government agencies or made available by other organisations.

We take a pragmatic stance here as we write at a specific point in time and from our home country context (Australia), which we acknowledge is a high-income country with neoliberal ideology influencing social policy. Non-profit data analytics is a fast-moving field where practices and legislation will change. Other countries and regions have their own nuances. Globally, the non-profit sector is on a journey with data collection and computational data analytics. This is influenced by policy that drives competition and demand for accountability and measurement, as well as a desire to use sophisticated techniques for social good. This journey will continue into the future.

This moment feels like a critical juncture for non-profits and data analytics. Current strategies and decisions taken within the sector will significantly influence both the nature of non-profit data analytics and the philosophy underpinning it, but perhaps most crucially, it will influence who has the capability to work with data and to what ends—towards what understanding of social benefit. We believe that non-profits need to have data capability to shape the future of the sector and affect the difference non-profits can make in the world. The sector can be knowledgeable, confident and advocate for suitable data practices, or—lacking capability—be forced to passively accept data practices determined by other powerful actors like government and ‘Big Tech’.

This book is meant for non-profit leaders, managers, practitioners and board members who want to see what can be done with data and discover how organisations like theirs can become capable with data. It is also for researchers, as we show how partnering with non-profits can help us to contribute to social justice and to knowledge about data for social good. The book is deliberately targeted at the practice and researcher nexus.

This first chapter sets the scene by introducing concepts, challenges and our rationale for why non-profits should engage with data analytics. It is by no means comprehensive in its understanding of international data initiatives in the non-profit sector, especially not in relation to data law and guidance in different country contexts. For that, we recommend seeking out local expertise, as that area is subject to variation by country or region, and subject to change as practice is only forming.

The Non-Profit Sector and Data

The non-profit sector comprises organisations with different legal and operational structures, including charities, philanthropic foundations, voluntary and community organisations, community groups, social enterprises and co-operatives (Salamon & Sokolowski, 2018). Some non-profits generate profit but re-invest it for social purpose. The sector has different names internationally, including the charitable and non-profit sector (Canada); third sector, social economy, voluntary sector (UK); third and social economy (Europe); not-for-profit sector, community sector (Australia and New Zealand); and charitable, voluntary and philanthropic organisations, civil society (US) (Lalande, 2018; Productivity Commission, 2010; Salamon & Sokolowski, 2018). Non-governmental organisations (NGOs) are non-profits that tend to work in other country contexts (Vaughan & Arsneault, 2013).

While non-profits generally operate to address social purposes not suitably addressed by government or private organisations (Vaughan & Arsneault, 2013), the social welfare role of non-profits can vary even within countries. Indigenous cultures including the Maori of Aotearoa (New Zealand), for example, have different understandings of social and community life that influence what is considered acceptable work for community organisations. Western notions of volunteering, separation of family and community, and who should provide community services should not be regarded as automatically aligned with Indigenous Peoples’ cultural understandings (Tennant et al., 2006).

In high-income countries, non-profits are significant providers of community services, including health, mental health, social care, education, environmental protection and disaster relief programmes. They contribute significantly to national economies; for example, employing around 13% of Europe’s workforce (Salamon & Sokolowski, 2018). Charities alone employ one in ten workers in Australia (Social Ventures Australia and Centre for Social Impact, 2021). Beyond service provision, non-profits contribute to generating a sense of community, “giving expression to a host of interests and values—whether religious, ethnic, social, cultural, racial, professional or gender-related” (Salamon & Sokolowski, 2018, p. 56) and, importantly, act as social policy advocates (Salamon, 2014). As such, non-profits are key actors in the policy community. They influence what are recognised as societal challenges, provide evidence about fruitful solutions and influence how the work of their sector is done (Vaughan & Arsneault, 2013). Government is a major funder for non-profits in high-income countries via contracts to provide welfare services (Salamon & Sokolowski, 2018). This increasingly leads to governments dictating the terms of engagement. Consequently, it is imperative that the non-profit sector is capable in contemporary organisational practices and innovations so it can influence social policy through data-supported knowledge and ideas.

In countries where policy is imbued with neoliberal ideology, including the UK, Australia and New Zealand, increased provision of public welfare services by non-profits started in the 1980s–1990s (Tennant et al., 2006). During this time, many traditional voluntary organisations became non-profit businesses. Additionally, the trend of non-profits supplying welfare services accelerated following the 2008 Global Financial Crisis. The marketisation of the non-profit sector led to competition for funding between organisations, forcing increasing corporatisation. Some now refer to a not-for-profit industrial complex (Incite! Women of Color Against Violence, 2017), with concerns non-profits are forced to subordinate their social mission to respond to funder-determined priorities in order to survive.

Accountability and reporting demands of government and philanthropic funders mean non-profits have had to collect increasing quantities of data. Funders influence or define the data to be collected and may even supply data collection systems. This scenario can stifle non-profits’ internal strategies about working with data and funnel their work towards reporting rather than using data to drive social change. To date, the sector is accused of over-emphasising easy-to-collect output data (e.g., about number of services delivered) rather than data about outcomes, impacts and the processes underpinning them (Lalande & Cave, 2017). Over time, as non-profits look for new ways to gain competitive advantage, interest in innovative data use has grown. Some larger non-profits invest in data professionals, while others contract with specialist consultants.

The danger with outsourcing data-related work is that organisational data and analytics become viewed as ‘too hard’ and internal know-how diminishes. We propose non-profits need to have data capability so they can appropriately drive their organisations’ data strategy for impact. More widely, collectively developing data capability at a sector level enables non-profits to influence government and funder priorities and investments around social challenges and data practices, informed by grassroots experiences. Here, we understand non-profit data capability as a holistic concept that involves interconnected combinations of resources. Data capability is hard to pin down to a checklist or benchmarking tool. It involves having the staff skills and roles, technologies, data management practices and processes that are appropriate for each non-profit in relation to its context of practice and enables effective use of data within that context. Thus, data capability for a non-profit is likely to evolve, potentially in response to changing organisation priorities, learning from trying out techniques and datasets, and in response to emergent data practices and norms of the non-profit field. Non-profit data capability has foundations in responsible data governance. We suggest it can be built through collaborating, experimenting and discovering with data. We extend our discussion about non-profit data capability and how to achieve it in Chap. 3.

Unfortunately, as related to business operations rather than direct service provision, data and information management tends to be underfunded in non-profits (Social Ventures Australia and Centre for Social Impact, 2021; Tripp et al., 2020). Ongoing lack of investment and expertise in social data analytics leads to problems with adopting innovation, resulting in a phenomenon termed the non-profit starvation cycle (Gregory & Howard, 2009). This is where ongoing focus on funding service delivery leaves organisations simultaneously under-invested in management and infrastructure, but also in staff skilled to understand what is required. Organisations are thus vulnerable to environmental shocks, as seen in reactions to the recent COVID-19 pandemic. A survey of Australian charities’ capability to deal with the pandemic found only 46% used cloud-based systems and only a third had systems and software for working at home. Deficits were mainly attributed to underfunding (Social Ventures Australia and Centre for Social Impact, 2021). A survey and report by Australian technology non-profit Infoxchange shows that the sector has not yet prepared for advanced data analytics or for automated futures, although investment in information technology and digital infrastructure and systems is improving and the skilled workforce is expanding (Infoxchange, 2020).

Collaboration between non-profits would enable cost-sharing for infrastructure and skilled workforce, but competition in the sector is a barrier. This has led to suggestions that government should incentivise or facilitate collective working (Social Ventures Australia and Centre for Social Impact, 2021). Some successful collaborative models exist; for example, Collective Impact initiatives, where community organisations work together to identify, address and monitor change about a social challenge. LeChasseur (2016), for example, describes a Collective Impact initiative to improve lives of low-income mothers and their babies. In Collective Impact, collaborating with data facilitates measurement of community-level social change as well as helping to assess the contribution of individual organisations. Some non-profits are involved in initiatives funded by Social Impact Bonds, where private investment can be gained to fund projects to improve social outcomes, with outcome data required in order to access premiums (Arena et al., 2016; Sainty, 2019).

Making Good Use of Data

The main goal of non-profits using data analytics is to inform organisational learning so adaptations can be made to achieve better outcomes. A range of reasons for applying analytics techniques to data to advance social missions are outlined by Verhulst and Young (2017), including for situational awareness and impact evaluation. Once attracted by the prospect of generating such analyses, the issue for non-profits might turn to how to adapt existing datasets, departments and staff into a system capable of generating insights from data.

Data analytics for non-profits is not solely predicated on having access to technology and applying computational techniques. Rather, it builds on having a foundation of knowledge about using data in research and evaluation. In this way, as the science of examining data, data analytics involves considering the characteristics of data you have or can access; its provenance and how it was collected; its availability for different uses and who can access it in unprocessed or analysed versions; understanding the ethical concerns, the consent given and obtained when data was created; the quality and what is missing in the data; and who data refers to or was collected from, to understand any in-built biases and data’s inclusivity. However, as well as drawing on traditional research and evaluation knowledge, data analytics also requires evolving thinking and skills as new forms of data and analytical techniques become available and new ethical principles and practices are developed in response (O’Neil & Schutt, 2013). Ultimately, good use of data includes careful attention to how it is generated, the widening range of data types that can be analysed, and the impact this may have on people’s privacy and other rights (see Chap. 3).

Exemplifying how using new types of data requires ‘old’ and ‘new’ thinking, we used a dataset of anonymised discussions on a national online peer support forum to evaluate services for rural mental health (Farmer et al., 2020; Kamstra et al., in press). Analysis was applied to identify themes in a large qualitative dataset of posts. Moving beyond traditional approaches to service evaluation, using the forum discussions as a rich qualitative dataset meant first agreeing on a rationale for the analysis conducted, and recognising the complexities inherent in the dataset as a sample. For instance, we had to address the potential for bias given that some people were over-represented in the data (i.e., posting far more often than others). With the focus on more isolated rural service users, we removed posts made by people living in large rural towns with hospitals to ensure only more isolated residents’ experiences were included. The data allowed us to access the geospatial locations of those using the online service, but when mapping quantities and themes of posts geospatially, we had to consider how to visualise the data at sufficient spatial scale and abstraction to remove any potential for identification. Thus, while computational techniques now allow analysis of much larger datasets, and new sources extend potential for social value extraction from data, many of the same basic research skills are required to intelligently conduct and interpret data analyses. Making good use of data involves navigating new possibilities, while translating traditional research skills to respond to new challenges.

Before progressing further, we now summarise the main types of external data sources and types of internal data content that we think non-profits might work with. Figure 1.1 illustrates characteristics of data we have used in our projects. It is not intended to be comprehensive of all data sources and content that could be used (for additional ideas, consult other relevant taxonomies, e.g., Susha et al., 2017).

Fig. 1.1
A circular wheel chart is spilt into 2 sections, internal data content and external data sources. The latter is further divided into 2 then 3 components and internal data content is further divided into 2 then 7 components.

Taxonomy of data that non-profits might use

Full size image

We divide the data that non-profits might use into two categories: internal data content (i.e., this indicates the broad types of dataset content generated by non-profits through their work) and external data sources where data with a range of characteristics may be accessed. In Fig. 1.1, we suggest non-profits’ internal data content can be divided into two types: operational data, where data is generated for and through an existing business purpose, including data about staffing, clients, services and funds; and what we term outcome data, referring to data collected specifically for assessing processes, outcomes or impacts of programmes. For the outcome data, what to collect is likely to be informed by a theory of change or programme logic showing links between non-profits’ programmes, how they are delivered, what they achieve and the ultimate fulfilment of social mission. Typically, outcomes data might be collected through surveys at intervals following provision of programmes. External data sources include all data that can be accessed external to the organisation and used, including open data generated by government statistical agencies and data made into open data by other organisations. An example we have used is the Infoxchange AskIzzy Open Data Platform (https://opendara.askizzy.org.au/), which provides anonymised geospatial location-based data from searches for community services across Australia. External data also includes government and other organisations’ data that can be made available under certain conditions and for particular purposes. Such data may be accessible subject to risk assessment or research protocol (e.g., sensitive government-collected health or crime data). Our understanding of ‘other organisations’ data extends to data from other non-profits, private sector organisations, academia and community groups. External data could include internal (private) datasets where data is only available to be shared within a limited collaborative group. This data will be available to the group under specific conditions through data sharing agreements as part of data sharing initiatives.

Data may be quantitative, for example, amount of time spent with clients, numbers of episodes of types of services delivered, distances travelled to deliver services and financial information; or qualitative, for example, discursive content of notes relating to clients, complaints and feedback, online forum post data. To be meaningful and relevant, analysis should also harness data that is temporal, for example, data capturing client needs and transactions on a daily or weekly basis over time and other forms of monitoring to enable longitudinal and even ‘real time’ analysis; and data that is locational, for example, giving a geospatial location of where services were provided or locations of clients and staff (Loukissas, 2019).

Having summarised types of data that a non-profit might use, a further issue is how they might think about sources of internal data for use in analytics. Through our work, we observe two approaches to sourcing internal data that we term here the new data and re-use data perspectives. The new data perspective tends to align with growth of the outcomes measurement movement (Lalande & Cave, 2017; Social Ventures Australia, 2021), where non-profits want to substantiate their social impact. This is generally handled by collecting new data about outcomes, impacts and processes. Where organisations initially tended to generate data through bespoke programme evaluations, more recently there is a trend to collect generic outcomes data using frameworks and data models. Using standard tools means non-profits can save effort in generating their own indicators and measures, plus a standard framework allows comparison and benchmarking across different organisations. Theoretically, funders will be able to discover which non-profits most successfully address a social challenge such as social inclusion, employment or crime prevention. Examples of these are generated by governments (e.g., the New South Wales Government Human Services Outcomes Framework, see https://www.facs.nsw.gov.au/resources/human-services-outcomes-framework) and businesses or social enterprises (e.g., Australian Social Values Bank). Researchers have also developed frameworks, for example, the Community Services Outcomes Tree (Wilson et al., 2021) was designed to provide “a comprehensive outcomes framework to assist services to name and then measure their outcomes…[and]… a set of data collection questions so services can ask questions of service users and collect data” (p. 1).

While such frameworks might assist cash-strapped non-profits, they have potential downsides. They imply collecting yet more data and are potentially inflexible to the nuanced interests and missions of individual non-profits. Adhering to them could drive isomorphism where programmes tend to become increasingly alike as driven by addressing a standard set of performance measures. This could hinder innovation and lead to neglecting nuanced needs of different clients and consumers. Piff (2021) highlights that non-profits could waste valuable time trying to find the perfect framework and re-orienting their data collection to meet its new requirements.

Advocacy for the data re-use perspective comes from policy institutes, researchers and others that are interested in combining digital social innovation with growing community and civil society data capability (Dawson McGuinness & Schank, 2021). Analysing re-used data is something of a frontier space where data scientists may partner with social scientists, lawyers, community practitioners and citizens to formulate practices that are ethical and obtain added social value from data already collected (Williams, 2020). New rules, standards, models and tools are often emergent from practical data analytics ‘discovery’ projects and collaborations (van Zoonen, 2020). An example of generating novel transferable tools comes from our projects with non-profits (see Chap. 2) where data protocols and data-sharing agreements were formulated through iterative discussions with data scientists, practitioners at non-profits and lawyers, where necessary.

Ultimately, of course, data must have been collected in order for it to be re-used and so the new data perspective also could generate data with potential for added value from re-use. Sometimes there may be a need for new data, but given a lot of data is already collected and exists, we advocate for optimising data re-use (where ethical and feasible) and minimising collection of new data.

As mentioned above, non-profits might work with others (non-profits and other entities) and share or pool data for richer insights and to drive collective working. Sharing data in multi-organisation collaborations is notoriously challenging (Verhulst, 2021). Understanding the extent to which data can be re-used and for what purposes, including sharing across collaborations, involves knowing why and how data was collected originally—and crucially—the details of consent obtained from those contributing to data generation (Verhulst, 2021). In the case of non-profits with their propensity to collect personal data, it often involves knowing about the nature of consent from clients, citizens and staff. Issues around consent for re-using and sharing data are explored in Chap. 3.

Starting to Think About Data Capability

Moving non-profit data analytics out of an environment of research projects and experimental initiatives and into business as usual requires comfort with using data and understanding the roles of data across the organisation and beyond. As noted above, data capability can be understood as a holistic concept, and we explore this in more detail in Chap. 3. Building data capability is not just about buying software or employing data professionals. Rather, it involves deepening knowledge and expertise in connecting the goals and work of a non-profit—their mission—with resources enabling appropriate use of data to meet the goals. This includes proficiency about what, where, why and how data is significant and why and how to use different data analysis techniques (Tripp et al., 2020).

It takes effort and commitment to grow organisational data capability, and there is a temptation to turn to commercial platforms and tools, like Amazon Web Services or Microsoft Azure, for data management and analysis. The challenge with implementing such tools without an organisation having done the groundwork to gain data capability is that they apply advanced analysis techniques without transparency. An organisation that invests internal know-how into identifying and implementing tools and practices that match its needs will understand potential for bias and other data harms. While we do not explore use of artificial intelligence (AI) in this book, it is coming and indeed already present in some non-profit operations and social service work. Non-profits that build their data capability will be resourced with knowledge to understand this application of data and to advocate and advise on ethical and wise use of advanced techniques.

In the context of non-profits’ data work, we favour using the term data capability. Data literacy and data maturity are other terms applied to try to capture the idea of being ‘ready’ for using data. The need for citizen ‘big data literacy’ is widely discussed (e.g., Grzymek & Puntschuh, 2019; Müller-Peters, 2020) in the context of ‘data citizenship’ (Carmi et al., 2020) as a response to expanded datafication and algorithmic decision making. Sander (2020), for example, suggests this “goes beyond the skills of… changing one’s social media settings, and rather constitutes …[being]… able to critically reflect upon big data collection practices, data uses and the possible risks and implications that come with these practices, as well as being capable of implementing this knowledge for a more empowered internet usage” (p. 2). One problem with using the term ‘data literacy’ in the context of non-profits is that it tends to target the competencies and critical awareness of individuals (D’Ignazio & Bhargava, 2015; Frank et al., 2016) and thus seems less suited to considering organisation-level attributes.

Similarly, we are not enthusiastic about the term ‘data maturity’, even though it suggests organisation-level qualities, because it conjures up the notion of an ultimate ‘finish line’ and doesn’t account for the wide variety of circumstances that shape the use of data. We opt to talk about data capability because what we envisage are plural and dynamic qualities, situated historically and culturally, that are fundamental to fostering change across new socio-technological milieux. While ideas of data literacy and maturity help by compiling skill and competency needs, our approach is to democratise data practices, open up data expertise to all parts of an organisation and push it beyond the IT department or the bounds of appointing specialist data professionals. Our holistic conceptualisation of data capability resonates with Williams’ (2020) depiction of ‘data action’ for public good—which is described as “a methodology, a call to action that asks us to rethink our methods of using data to improve or change policy” (p. xiii). Aligned with this call-to-action approach is a widening of data accountability, responsibility and ethics. In short, data capability involves more than ticking off attributes from a list but is about evolving understanding, resourcing, implementing and doing, involving people across organisations and in relevant communities, and interacting with changing contexts and missions.

In this book, we provide examples of how non-profits can use data and give practical strategies for non-profits to build data capability. The central approach we offer for building new capability is collaborative data action. Rather than consigning data solutions to individual projects or teams, we encourage collaborative processes within and across organisations. In Chap. 2, we give case studies of using collaborative data action with non-profits to generate new insights from using and re-using data. In Chap. 3, we delineate the collaborative data action methodology and highlight why it is particularly useful for non-profits. Based on our research with non-profits, we distil out key issues for non-profits to prioritise. Our mission is to put data analytics within the reach of all non-profits and to overcome isolationist and competitive data practices that concentrate capability with the well-resourced (large) few. That is, not replicating the logic of private enterprise, commercialism in data use and start-up culture exceptionalism.

Part of the ‘magic’ of collaborative data action is bringing together different knowledges, skills and experiences because data analytics for non-profits is a hybrid activity (Verhulst, 2021; Williams, 2020). It requires the skills of data scientists, but they tend to lack social science training. It requires social scientists with grounding in evidence and methods of social fields, and it needs practitioners because they know the practices and operating contexts of non-profit work. As non-profits’ capability is built, their data work increasingly must incorporate the voice and perspectives of clients, citizens and communities. To achieve this, it is necessary to navigate the problematic environment that has arisen due to some of the ways that social data analytics has been applied to date—that is, to address the (ab)use of data causing social harm.

Navigating Data Harms by Involving Citizens

Part of the rationale for growing non-profits’ data capability is to bridge the gap between desire to extract optimal social value from data, while addressing the risks from (re-)using this data. Much of the data non-profits generate and work with is likely to be personal data about clients and customers, perhaps sensitive and health-related data. Accountability to clients, customers and communities around use and re-use of data is paramount and challenging to execute well. At this point, as good data safety practices and technology are available, challenges are mainly due to a lack of established, evaluated models of good practice of how to work with people to formulate governance principles and processes for re-using data about them. And, building on this, how to engage citizens as empowered partners in data projects that engage with their data.

Constructing sound practices for using and re-using citizen data requires citizens at the table. In our experience of data projects with non-profits, they find it challenging even to think about holding discussions with clients and consumers about how to develop such practices. They appear afraid to mention ‘the d word’. This fear of engaging with clients and consumers regarding data is linked largely to perceptions of risk due to high-profile accounts of social data misuse. Critical accounts of datafication emphasise the way data has become a social and political issue “not only because it concerns anyone who is connected to the Internet but also because it reconfigures relationships between states, subjects, and citizens” (Bigo et al., 2019, p. 3). Accounts about the impact of datafication on society are multiple and sometimes depict grave consequences. They exemplify harms from use of data analytics in replicating and driving inequalities of race and ethnicity, gender and class, and concentrating power in the globally dominant technology corporations (e.g., Criado-Perez, 2019; Eubanks, 2018; Noble, 2018; O’Neil, 2016; Srnicek, 2016). High-profile failures to use data and technology in social welfare settings, for example in Australia, the notorious failed Federal Government ‘Robodebt’ automated debt recovery programme based on welfare services data (Henriques-Gomes, 2020), are mirrored internationally. Such cases have eroded public confidence in institutions that would traditionally be trusted to care for and about citizens and data.

Different countries and regions are beginning to clarify data rights and heighten the accountable, responsible production and use of personal and social data through high-level legislation, such as the European Union’s General Data Protection Regulation (GDPR) (European Parliament and the Council of the European Union, 2016), and proposed Bills to regulate Artificial Intelligence (AI). However, there is still ongoing uncertainty about what rules pertain in different contexts—and even how to find out. Data security and privacy law and responsible data governance are core elements of the context of non-profit data analytics, but we also note that risk aversion around working with data can be the immediate, and apparently easiest, response. Among non-profits highly sensitive to social injustice, vulnerabilities and systemic inequality, the idea of doing more with client and citizen data can be met with considerable anxiety, resulting in waiting until things get clearer (i.e., not re-using data). We suggest a key reason why non-profits should grow their data capability is so they can confidently and competently engage with clients, citizens and communities around responsible data use. While there are risks, and a need to proceed with caution, using citizen data for insights could bring benefits to clients, customers and the wider community. Data is already generated, so it is responsible re-use that is the central issue to be resolved. There are, arguably, three key issues to be considered in non-profits working with citizens and data: (1) developing sound data governance practices, (2) working with citizens to gain insights from data and (3) raising citizen data literacy and community data capability.

Some researchers have begun to explore how to involve ‘lay’ participants in discussions around responsible use of data. For example, the Data Justice Lab (Warne et al., 2021) produced a civic participation guidebook outlining participatory methods including citizens’ juries and mini-publics (deliberative conversations) to discuss data use. Living labs and hackathons are other methods discussed (e.g., Flowing Data, 2013). These methods, though, tend to engage citizens in discussing large administrative or government datasets, rather than making direct links between citizens and re-use of data about them. There are some cases of active engagement of citizens with deciding about uses of their own data; for example, the Salus health data co-op in Barcelona involves people making decisions about selective use of their data (e.g., for health research), as opposed to making it entirely open or private and unavailable for re-use (Calzada, 2021). Open Humans (https://www.openhumans.org) is a non-profit dedicated to supporting individuals and communities to explore use of their data for social purposes. We found a few examples of engaging more marginalised groups about their data, and these are the citizens with which non-profits are most likely to work.

Here, perhaps, work on Indigenous data sovereignty indicates a useful way ahead (Kukutai & Taylor, 2016). Data sovereignty is a way of understanding the importance of establishing consent and respecting the rights of, and ensuring benefits for, those who are the subjects of data (Carroll et al., 2020). In many parts of the world, Indigenous data sovereignty working groups and scholars are defining and addressing data inequalities and exploitation among those who have had least control and benefit over data collected about them. Carroll et al. (2020) discuss the process and rationale for developing the CARE Principles for Indigenous Data Governance. CARE stands for: Collective benefit, Authority to control, Responsibility, Ethics; and the principles are intended as a guide for stewardship and processes to enable self-determining citizens to make decisions relating to collection, storage, analysis, use and re-use of data. The CARE principles were developed by Indigenous people due to widespread abuse of data about them involving issues of over-surveillance, use of data for policing, lack of transparency and control, and under-counting (thus under-representation). Data is as important to the sovereignty of a people as language, artefacts, landmarks, beliefs and cultural knowledge, and natural resources. As Tahu Kukutai and John Taylor eloquently argue: “missing from those conversations have been the inherent and inalienable rights and interests of indigenous peoples relating to the collection, ownership and application of data about their people, lifeways and territories” (Kukutai & Taylor, 2016, p. 2). Indigenous ways of knowing can offer new models for data governance that are built on collaborative, rather than individual or proprietary responsibility, and more respectful forms of consent. Work on Indigenous data sovereignty can offer principles for wider application to engage with citizens represented in data and who have experienced power inequities.

Moving beyond citizen engagement in designing data governance, clients and consumers should be engaged where non-profits re-use data about them. This could involve data analyses relating to, for example, situational awareness, impact assessment or for community insights. This goes beyond acknowledging people’s representation in the data, but also acknowledges their vital ‘lived experience’ roles in ground-truthing and interpreting ‘what is going on’ in data analyses. Most contemporary non-profits have established relationships and ways of engaging lived-experience clients and customers in informing and enabling services so engagement with data analytics would represent an extension of such work. Partnering with citizens about data is important for informing the work of non-profits, and, as such, should be appropriately recompensed. This acknowledges the expertise of citizen clients and customers as key stakeholders in use, visualisation and interpretation of data that is about them. As Williams (2020) notes, involving citizens is integral because “data are people” (p. 220).

Some excellent examples of resources for involving citizens with lived experience in data projects have been generated in recent years through work of Elsa Falkenburger, Kathryn Pettit and others at the Urban Institute and specifically its National Neighborhood Indicators Partnership (NNIP; https://www.neighborhoodindicators.org/). These community data advocates devised a ‘data walk’ methodology to engage citizens with analysed and visualised datasets to help make decisions about their communities (Murray et al., 2015). More recently, a short Guide to Data Chats resource has been produced for practitioners, giving really practical advice and tools for involving citizens with data (Cohen et al., 2022). As part of the NNIP’s projects, citizens are often trained to collect new, granular ‘citizen science’ data about aspects of living in the locale.

The NNIP sees building community data capability as a key outcome of engaging citizens in data projects. In their role as engaged with clients and customers, non-profits could be significant in developing citizen and community understanding around ethical data collection and use. As digital inclusion becomes central to social equity agendas, non-profits’ data work with clients, customers and citizens could move beyond service delivery and contribute to a wider social mission of building client data literacy. This could be done by engaging people with their data, discussing issues such as sovereignty and potential to re-use data and generating co-designed data governance. Such activities would centre clients and consumers in non-profits’ data practices and contribute to building data capability at community level.

Initiatives around the world are working to provide examples of ways to engage citizens, for example Our Data Bodies (https://odbproject.org) is a project working with low-income people in the US and data rights, and Amnesty International is engaging with data volunteers to help organise crowd-sourced datasets (Acton, 2020). However, specifically considering the range of large and small non-profit organisations, our experience of current practice is that non-profits’ engagement of consumers and clients in re-use of their data does seem to present quite a leap. Most non-profits we have worked with are still at the stage of building their own internal data capability. As Sander (2020) concludes—with regard to citizen engagement—there is, as yet, “too little knowledge on what kind of literacy efforts work best and a lack of constructive or comprehensive research on how to address people’s lack of knowledge” (p. 1). We argue that non-profits’ management, boards and staff require their own data knowledge, awareness and experience as a precursor to engaging clients appropriately in conversations about data and involvement in codesign of data use practices. This is not ideal but realistic based on our experiences. Until this time, it is imperative that non-profits understand the consent they have to gather, using this knowledge to work within general ethical parameters (Williams, 2020).

Key Takeaways from This Chapter

In this chapter, we set the scene and introduce some key ideas about why and how non-profits need to engage with data analytics. The key points we’d like readers to take away are listed below.

Key Takeaways

  • Non-profits should have the same access to data capability as commercial businesses. They should build data capability so they can inform data strategy for their organisation and the sector.

  • Non-profits should resist generating new data if possible, rather they should explore ways to re-use data they already generate and use open social data instead.

  • Once non-profits build their organisational data capability, they are well-placed to work with clients and citizens to help build wider digital inclusion and community data capability.

  • Non-profit data analytics is a hybrid space that, at its best, draws on multiple areas of knowledge, expertise and lived experience.

In the next chapter, we present case studies that illustrate our journey of working with non-profits and data, from an earlier example of working largely with social media data and government consultation submissions, to working with non-profits exploring their own data, to generating a data collaborative with non-profits and other organisations taking a place-based approach (Chap. 2). We present our case studies in Chap. 2, to give a picture of the different kinds of data projects we are talking about in this book, but also because it was working on these projects that led to the understanding of data capability we suggest here and our appreciation of the benefits of working collaboratively. In Chap. 3, we build out from those learnings from the case studies. We more fully describe what data capability for non-profits looks like and outline the collaborative data action methodology that we generated and refined while working on the case study projects and reflecting on similar work elsewhere. In Chap. 4, we look to the future—discussing the way ahead for non-profits and data analytics for social good and suggesting research and practice priorities. Data practices and regulation are dynamic and rapidly changing so there will be new work that constantly refreshes and extends what we say here. Our focus in this book is on what we gleaned from very practical projects with practitioner partners. We note the book does not provide a comprehensive international scoping of all uses of data for social good or initiatives. Rather, here we tend to highlight the initiatives and resources that we have drawn on most in developing our work (see appendix for specific detail of these). We hope this book gives help and inspiration to non-profits seeking data analytics for social good and researchers working alongside them.