For social impact organizations, finding the time and resources to prioritize learning and innovation can be a struggle. This can lead to stagnation and a difficulty in adapting to new opportunities and evolving needs. 

In the past decade, as Noora Health has grown to 350+ people across four countries, we’ve tried to tackle these challenges head-on. We’re working on fostering a culture that encourages learning and creative problem-solving — something we believe is essential for addressing complex social issues.

Participating in a workshop on learning, data, and technology organized by The Agency Fund* in Naivasha, Kenya, last year, was one meaningful stop in this journey of continuous learning. Attended by over 15 peer organizations, it was a powerful opportunity to exchange raw insights, reflect on challenges, and deepen our collective understanding of learning, growth, and innovation. 

Here are four key lessons we took away from the workshop that have shaped the way we think and work. In this new year, we hope to turn more of these insights into action, and continue learning together to build a stronger social impact learning ecosystem. 

Lesson #1: Learning is both a marathon and a sprint

In Naivasha, we met with organizations of all sizes experimenting with innovative technologies, from Jacaranda Health’s UlizaLama, a bespoke English-Swahili language model, to SameSame, which is testing new WhatsApp interventions. We learned about the various learning projects organizations were using to gain insights from new programs and experiments on an on-going basis. These efforts ranged from a few weeks to a few months long and aimed at quickly gathering feedback to confirm initial assumptions, identify challenges in implementation, or measure the claimed benefits. 

From these conversations, one overarching takeaway emerged: Learning is a continuous — and often uncomfortable — process for organizations, regardless of size.

As a rapidly growing organization that wants to foster a culture of learning, experimentation, and innovation across all teams, this resonated strongly with us. Working across four different countries has made it increasingly critical for us to systematize learning so that we aren’t reinventing the wheel or repeating the same mistakes. Without it, we cannot hope to achieve our ambitious impact goal of reaching 70 million caregivers by 2027.

To do this, over the past year, we’ve started putting tools and processes in place to enable long-term learning. One example is our knowledge management system which allows teams to input updates, documentation, reports, and other valuable information in a standardized format. Previously, this information was fragmented and underutilized, making it difficult to synthesize learnings across verticals. 

The new system addresses two key challenges simultaneously: reducing the effort required to collate information across the organization and making it easier for teams to access and navigate this information. 

Simultaneously, we’ve worked on strengthening short-term feedback loops focused on specific projects or outcomes. One example of this was conducting rapid A/B experiments within our digital interventions to maximize engagement and caregiver satisfaction. With this momentum, we plan to continue doing more A/B experiments with the help of modern tools like Large Language Models to improve onboarding and engagement with our WhatsApp service. 

Taken together, the hope is for these approaches to facilitate a more holistic evolution of our programs as we scale. Changing processes and mindsets is hard, but the potential dividends of getting this right are immense.

Lesson #2: Learning, knowledge management, and data go hand in hand

If learning is the star of the show, data and knowledge management are the backstage crew ensuring everything runs seamlessly. As organizations grow, they need to intentionally build a data-driven culture, particularly when it comes to identifying the key metrics that will inform program decisions and changes.

Some ideas from Naivasha that stood out to us included:

• Creating data champions: Team members who actively integrate data into everyday conversations, meetings, and reviews.
• Sharing data stories: Highlighting meaningful insights with the larger organization to foster transparency and increase comfort with using data in daily work.

These practices help weave data into the fabric of organizational processes, ensuring it becomes a trusted partner in decision-making.

Leveraging data to inform programmatic decisions, however, is only one piece of the puzzle. Equally critical is documenting this data within a standardized framework to ensure it is accessible and actionable. Every day, we generate a wealth of information — from needs assessments and design research to feedback, monitoring data, and impact evaluations. We also operate across multiple modalities (in-person, WhatsApp, IVR etc.) and across multiple levels of the health system. The result is data that lacks standardization, which ultimately impacts categorization, searchability, synthesis, and analysis. 

Here too, our knowledge management system becomes indispensable. By providing pre-determined formats for different kinds of data, such as field visits, curated experiments, meeting notes, etc., we’re making information more digestible for our internal teams (and also potential future AI systems, which require structured data to function at their best).

Alongside this, we’re focusing on strengthening our knowledge management practices, starting with a more deliberate approach to planning the information and data we collect for each intervention. Some efforts include:

• Planning ahead for the kind of information and data we collect for any intervention. For example, in 2024 we started to systematically document assumptions and metrics at the start while designing our programs, to then inform the larger learning plan after program implementation. 
• Prioritizing critical information pathways including daily and weekly facility visit reports, insights from needs assessments and user research visits, and patient and caregiver stories collected by our team of medical support executives and teletrainers.

Poorly structured or fragmented data can create bottlenecks for downstream teams, complicate data system workflows, delay analysis, and ultimately hinders the generation of actionable insights. Although improving knowledge management is a multi-year journey, we are committed to building systems that make our data more structured, accessible, and actionable for the long term. 

Lesson #3: Trade-offs are inevitable

In the words of American economist Thomas Sowell, “There are no solutions. There are only trade-offs” — a sentiment that resonates deeply with social impact organizations like ours. The Naivasha workshop offered us a glimpse into how other nonprofits navigate the critical trade-offs they encounter. Seeing how these play out across various stages of growth gave us much to reflect on. Here are a few examples that stayed with us — issues we continue to navigate, without a one-size-fits all answer:

• Experimentation vs standardization: Government partners often prefer to tread the tried and tested path with interventions that have already been proven or ones that do not require additional testing or prototyping. As an organization that values iterative learning and constant experimentation, we often find ourselves navigating a delicate balance. Whether it’s testing new modalities and behavior change techniques or refining content and presentation, our need for iteration is sometimes at odds with the government’s preference for well-researched, replicable solutions that can deliver immediate impact at scale.
• Data collection vs stakeholder fatigue: We rely heavily on data to monitor our programs, evaluate their impact, and iterate on the learnings to improve them further. However, much of this data comes from frontline healthcare workers already overburdened with their government-mandated responsibilities. They are strapped for time, yet their cooperation is essential. This often leaves us in a conundrum: How do we gather the data we need without exacerbating the workload of these key stakeholders?
• Building vs buying: As a social impact organization with limited resources, we’re always grappling with the question of building competency in-house as opposed to buying technology off the shelf or outsourcing it entirely. The answer, as always, is somewhere in the middle. At the Naivasha workshop we observed that most organizations that operated digital interventions at scale, like us, had invested in building in-house tech teams — which can be an expensive activity. However, it also gives us greater control, agility, and room for innovation. That said, we don’t always take the ‘build’ route. We frequently leverage open-source tools and technologies and outsource experimental projects until they’re ready to be scaled internally. And we realized we’re not alone in this approach. For instance, IDInsight, another Naivasha attendee, adapted an open-source tool like d-answer for their internal knowledge management system. This allowed them to harness the power of modern LLMs without heavy development investments, demonstrating a strategic middle ground.

Navigating trade-offs will always be an integral part of any organization’s growth journey. For us, the key lies in approaching these trade-offs thoughtfully, ensuring that decisions remain aligned with the organization’s vision, mission, and goals.

Lesson #4: Collaboration is key, both inside and outside the organization 

On the last day of the workshop, as each of us shared reflections and ‘aha!’ moments, we recognized the importance of leaning on each other and working together, given the similarities in our missions.

History doesn’t repeat itself, but it often rhymes — perhaps particularly in the social impact sector. Problem statements and solutions tend to be similar, and organizations sometimes use the same tools to achieve slightly different, but related, outcomes. WhatsApp, for example, is a common modality used by a variety of organizations like ours to drive better health outcomes, such as Rocket Learning to improve educational outcomes and Digital Green to enhance vocational outcomes. 

When social impact organizations collaborate with the intention of maintaining transparency and openness, they can learn from and with each other.

However, the reality of the social impact space can sometimes be an impediment to collaboration, as organizations often have to compete for limited resources, such as funding and talent. This only highlights the need for events, convenings, and workshops like the one we attended at Naivasha, where we can come together and collaborate effectively despite these constraints.

Collaboration is vital for growth — not just across organizations, but within them as well. This seems intuitive but becomes critical to prioritize in growing organizations where silos are easy to create and difficult to break. At Noora Health, to systematize collaboration, we’ve established rituals such as learning circles which are structured sessions where different team members present their work and insights. These sessions serve as inspiration for the broader organization, creating spaces and opportunities to learn from each other. For example, we conducted a learning circle on the integration of our community-focused Care Companion Program with the Jagananna Arogya Suraksha program in Andhra Pradesh, India. This was aimed at inspiring other implementation teams to explore similar integration opportunities within the health system.

As we move forward, these lessons from Naivasha continue to shape our approach to learning, collaboration, and technology at Noora Health. We definitely don’t have all the answers, but are committed to staying curious and keep improving. 

*The Agency Fund collaborates with a community of donors representing a wide range of world views, but unified in their interest in expanding human agency. They are one of Noora Health’s funders, and collaborate with us on data, research, and learning activities. 

We’d like to thank Anubhav Arora and Suparna Kalghatgi for their support in writing this blog post.