Build Data Infrastructure to Bolster Scale
What systems do I need in place to manage and maximize my data?
As organizations begin to scale their impact, they often find that formalizing and bolstering their data infrastructure is required to power scale — particularly the systems that store and communicate data for operational and performance management, which can range from excel files to cloud-based, enterprise-wide data management systems.
IN ACTION: MiracleFeet, building a data platform to propel scale
MiracleFeet works with a network of over 500 local partners and treatment providers to supply the technical and financial resources needed to create and support high-quality clubfoot clinics. After working for six years with a legacy data management system, MiracleFeet knew what its data requirements were and that it needed to develop a more efficient and responsive system to support scale. MiracleFeet emphasized that having sufficient implementation and field experience before making such a major investment was critical.
“If we had developed our data system even a year or two earlier, it would have been a disaster,” said Chesca Colloredo-Mansfeld, Executive Director of MiracleFeet. “We needed to have enough diversity of experience in implementing our program — in different country contexts and facing different logistical challenges — to be able to design a system that would meet our needs.”
MiracleFeet was also fortunate to find a dedicated source of capital to create a system that could propel scale, when it won a Google Impact Challenge Award. With the $1 million award, MiracleFeet undertook a process to work with other key clubfoot treatment programs globally to ensure it designed a system that could meet all of their needs as well. After two years of development and testing, MiracleFeet rolled out its CAST mobile data application, integrated into Salesforce for real time monitoring and reporting, as shown below.
Top Tips for building data infrastructure
- Start scrappy and simple.
Enterprises are often tempted to build out complicated, best in class, data systems early in their journey. But as noted in the MiracleFeet story above, interviewees stressed the importance of taking the time — often several years implementing and refining a solution in different contexts — to truly understand data system requirements before making a major system investment. Interviewees reported using basic data collection and analysis tools (often Excel) or inherited legacy systems for the first five to six years of implementation and, once they did start to build a new system, warned not to overbuild or be seduced by flashy technology. Asif Akram, Living Goods’ Chief Technology Officer, advises, “start simple and make sure you are meeting immediate requirements. Don’t get tempted by what others are doing or try to meet future requirements too soon. But keep improving as you go, building on experience in more of a snowball effect, and do what you can to set yourself up for the future.” Similarly, Harambee shares its four non-negotiables when it comes to building out data-related technology.
Harambee’s Four Non-Negotiables for Data Technology
1. Simplicity
2. Data integrity
3. Enabling for the organization (i.e., helps us to take action)
4. Maintainability (i.e., we need to be able to fix broken things quickly)
Survey respondents echoed Harambee’s first non-negotiable with 55 percent reporting that they would have made their data systems less complex if they could do it again. - Spend time considering future data needs.
While Akram and others advise keeping a data system simple, they also acknowledge the importance of “setting yourself up for the future.” While you will never be able to anticipate all future needs, by spending time considering them you may be able to make your data infrastructure more flexible for the future. For example, building a data platform that could accommodate different alphabets in the future if you plan to expand globally, having storage capacity that could flex over time to accommodate larger amounts of data, or selecting widely-used technology platforms to help keep systems relevant over time. As Colloredo- Mansfeld of MiracleFeet explained, “Our choice to use established platforms [such as Salesforce and CommCare] has been key to our ability to grow and be sustainable. These companies already have a vested interest in critical areas such as data security and privacy. And choosing a system with a growing network of users and user forums has allowed us to access knowledge about the platform from so many others using it.” Forty-five percent of survey respondents reported that, had they spent more time considering future data needs during development, their current platform could have been more useful. - Seek out a dedicated source of capital…ideally.
Without a dedicated infusion of capital for data and technology, many interviewees reported that they would not have taken the resources away from programs to invest in data infrastructure and therefore would make incremental changes over time. Those with dedicated capital (e.g., Google Impact Challenge Award for MiracleFeet, and Salesforce Force for Change Grant for BOMA Project) have been able to create platforms that they report radically changed their ability to drive impact at scale. So, even if you do not have the resources now, think about what an investment in data infrastructure could help you do, articulate your requirements and costs, and be prepared to pitch current and new funders should the opportunity arise. According to survey respondents, if they were given a dedicated infusion of capital to develop and test a data system to meet their needs, 67 percent would spend USD$250,000 or less on a data system (broken down as 29 percent spending less than USD$100,000 and 38 percent spending USD$100,000–250,000). Twenty- four percent reported that they would spend USD$250,000–500,000, and nine percent would reportedly spend $750,000–1,000,000). - Build infrastructure that can be applied to the problem, not just the organization.
With the goal of driving impact at scale, especially as it relates to systems change (see here for more on systems change data), organizations should consider how others working on similar issues might be able to leverage the platform and data being collected. With its Google Impact Challenge Award, MiracleFeet was able to create a system that worked not just for itself but for anyone working in clubfoot. The process was beneficial as it forced MiracleFeet to identify common ground among countries and among organizations and distill the most important needs and requirements. While it was more complicated and time intensive, it resulted in a product with greater potential impact and partners with more faith in the platform and in MiracleFeet as a collaborator. (On the flip side, you should consider how you can leverage any systems that have already been developed by others, if they exist.)
What about data privacy and security?
As organizations scale and build out data infrastructure, data privacy and security must be central considerations. At a minimum, this means ensuring your organization understands and complies with the legal requirements around data privacy (standards such as the EU’s General Data Protection Regulation), has a privacy policy, and brings in the right expertise around data security. For useful information on legal and policy considerations, including standards, see USAID’s Considerations for Using Data Responsibly Guide. Responsible Data, a grassroots network of social impact organizations, implores organizations to go beyond these legal requirements and take a rights-based approach to the use of data — which entails centering equity considerations, including people’s right to consent, privacy, ownership, and security of data.
What does this look like in practice? Crisis Text Line chooses to scrub all of its data of any texter identifiers in order to protect its texters’ privacy. This leaves them with data that, from an academic standpoint isn’t wholly “pure” but Crisis Text Line’s mindset is that “scrubbed data give us 95 percent of the insights and any small improvements you might get from a ‘pure’ data set is not worth the risk of loss of privacy.” Crisis Text Line has an independent Data, Ethics & Research Advisory Board which helps guide its policies around data privacy and security. Responsible Data has also compiled a list of resources to support organizations seeking to follow this approach.
Read next: Engage the Right People to Support Data Goals, Ensure Data Efforts Drive Toward Equity and Inclusion, or return to see all articles in Data for Scale.
This article was written by Erin Worsham, Kimberly Langsam, and Ellen Martin, and released in June 2020.
