Empower parents to improve the learning of children

India has about 1 million Government-run schools where over 40 million children at the primary level (Grades I - V)  have not attained the necessary foundational literacy and numeracy skills. According to the nationwide ASER survey, 60% of Grade 5 students in government schools cannot read Grade II text and 75% cannot do a simple division. In the National Capital, Delhi, about 1 million children are enrolled in primary grades of Government schools and struggle with learning. Lack of foundational literacy and numeracy in early years has been linked to higher crime incidence, lower income, and poor health indicators at a later stage.

The schools are the last resort for children to access quality education for a large section of underserved communities. The children enrolled in a Government school typically come from households of migrant labourers, often living in urban slums. The parents aspire for better lives for their children and quality education is their only chance to escape poverty. 

However, the children struggle with a critical gap in accessing appropriate learning solutions that can help them learn better. Their parents are often unable to help as they lack information about these solutions and capabilities to ensure their usage by the children. More than half of the parents we surveyed faced difficulties in accessing adequate information to support their children, while 62% expressed a lack of confidence in supporting their children's learning.

Our solution ensures that children can access appropriate learning solutions to achieve foundational literacy and numeracy. The learning solutions have been curated as per the "Teaching at the Right Level" approach which was found to have a significant impact on learning in independent studies. We enable the adoption of these solutions by providing information about children’s learning levels to the parents and enabling them to support their children in using these solutions.

Our support to the parents is delivered by a parent from a similar community, called Saajhedar, who are trained and onboarded into our team to support other parents. This makes the conversation between Saajhedar and the parent contextual and driven by empathy.

Our technology-led backend provides the necessary tools and data to complement the support that a parent receives from a Saajhedar. The tech comprises an Android-based calling application and a WhatsApp chatbot that enables a Saajhedar to reach out to the parent. It also captures relevant data related to the parent and the child support to ensure that Saajhedar’s support is effective.

The child’s journey with our program can be divided into two stages - :

Stage 1- Identification and onboarding of the parent: We work in partnership with the State Government and reach out to the parents through the parent-teacher meetings regularly organized at the school. Parents in need of support can give us a missed call to get onboarded into our network. Once a parent is onboarded, the parent is mapped to a Saajhedar from our team. 

Stage 2- Learning support: 

The onboarded parents receive one support instance every month. Each support instance comprises three steps 

I) A child's reading and math levels are identified through a standardized assessment over a phone call. We use a standardized tool (ASER tool) to assess a child's learning level.

ii) assessment results and potential learning solutions are explained to the parent over the call by our Saajhedar

iii) we resolve parents' challenges with contextual advice and motivation, completing a support instance. 

This cycle continues every month till the child achieves the highest levels of foundational literacy and numeracy. In addition, parents can reach out to our WhatsApp chatbot, and our incoming call helpline to seek support whenever they need it.

All the calls to parents are made through our Android-based calling application, Saajha for Saajhedars. The application is used by Saajhedars and allows them to collect all information about the parents being supported. The information includes child's name, learning level, parent information, interaction history, call notes etc. 

Our content includes level-specific worksheets for Maths shared daily, while worksheets for Hindi are sent on alternate days. Our WhatsApp bot built on the Turn.io backend allows us to send automated messages to the parents.

We are working towards enabling quality learning for children studying in government schools in India.  The government schools generally cater to low-income families, employed in the informal economy. A significant proportion of parents with whom we work are migrants working as daily wage earners, domestic support staff, community service workers, security guards etc. Most of the parents are educated up to Grade VIII and have an average income of less than $100 per month. 

The economic and social inequality, coupled with a lack of information about child’s learning difficulties and potential solutions results in a significant lack of confidence among the parents to reach out to the teachers to seek help for their child. Further, the existing channels for seeking information about learning are often unreliable. 

The parents aspire to have better conditions for their children and are forced to explore costlier alternative options like private tuition. These private tuition costs ~$30 per child annually, in cities like Delhi and further add to the financial burden for the families. Further, several families are not able to afford the costs associated with private tuition and hence are unable to support their child.

Our solution addresses this key challenge and offers necessary support to the parent at no cost. Our Saajhedars are trained to build trust and empathy with the parents to ensure that we can understand the problem and offer relevant solutions.

Our Team Lead and the core team have been working in the communities for more than 10 years since 2014. Over the years, we have gained an in-depth understanding of the critical challenges faced by marginalized student populations. Further, through our interventions, we have been able to build trust within the communities enabling us to provide more effective nudges to the parent for the child’s learning. Further, we have identified and onboarded ~50 mothers from the communities into our team. This has enabled us to build a primary perspective on parent’s and child’s challenges. This also enables easier sharing of information between the parents we support and our team. Further, our support is entirely free of cost and hence remains accessible for all parents.

We have also been able to build robust systems to capture the information and derive insights from it. Our tech backend is capable of recording all interactions with the parents through voice recordings, responses, assessment results, and other details such as the child’s reading and math ability. Our WhatsApp bot captures relevant parent interaction data including the content of the messages exchanged, links clicked, and items on the chatbot explored. 

The communities’ input, ideas, and agendas play a key role in designing our intervention. We have adopted a three-step approach to incorporate them into our intervention.

i) Need identification - We have a team of enumerators who listen to call recordings with the parents, as well as, identify areas where we could improve. We also float regular surveys and questionnaires on our chatbot seeking parents’ feedback about key elements of our support.

ii) A/B Testing - Once we can identify a need or a gap, we design and launch rapid A/B tests with a sample of parents. This is often followed by a qualitative survey to understand the findings

iii) Scale - Post the test, the findings are then scaled.

During the past 12 months, we have launched A/B tests for understanding a WhatsApp-based reward program, a dynamic support instance scheduler on WhatsApp, and a reminder feature for support.

We have set up a partnership with the Centre for Data Science and AI at IIM Ahmedabad, Platform Commons, and Wadhwani Foundation to improve the parent experience. Further, we have set up an Advisory Council comprising of professors from IIM Ahmedabad and Data scientists to help in improving our program. The partners have significant technical expertise in Data analysis, tech development, and AI-enabled assessments for reading. These would ensure that we build a solution that remains closely aligned with communities’ needs.

Our solution has been implemented and has been found to show positive outcomes in the learning improvement of children in our internal assessments. The development of technology started in April 2020 and we were able to undertake a well-structured program starting April 2022.  Since 2023, we have enabled ~200,000 parents to join our network and are being supported through WhatsApp and call-based support currently. As of March 2024, about 20,000  children were supported regularly out of which 9,000 children showed improvement of a level in Maths on our assessments. We have exchanged 20 Million messages with parents and ~45 thousand hours of call-based support till March 2024. 

Further, our program has enabled a parent-to-parent support mechanism, creating a ripple effect in the communities we work with. Many parents who have been supported previously have joined us as Saajhedars to support other parents through the solution. This has also ensured that the support being delivered is inclusive and locally relevant for the parents in the communities. 

In 2022, we partnered with the state government to work with ~1,000 Government schools. This has also enabled greater trust and reliability for our solutions for the parents. We have been able to find a balance between ensuring scale through technology while keeping the support personalized through regular interactions with the Saajhedars. This model has given us the confidence to further refine it and ensure that it can be replicated across other states with their respective governments.

We seek support in three major areas through Solve.

1. Data and actionable insights -  We have a large amount of data that we have collected through interaction with the parents. However, we are not able to use it entirely to generate insights. Through Solve, we are hoping to generate insights to improve our program further.

2. Technology support - A key goal for the current year is to build new channels for support - particularly AI-enabled assessments of reading levels. Through Solve, we are hoping to understand drivers for the adoption of AI-enabled assessments. 

3. Building team capabilities for a pivot - Over the next two years, our model will undergo a significant change, particularly as our tech capabilities grow and we can analyze data more meaningfully. Through Solve, we hope to understand how to make the team ready for the significant changes in our program delivery over the next couple of years.

Further, we are setting up an independent study for an external study to understand our impact on children more deeply. The grant support will partly cover the cost of the study. Additionally, through the mentorship and network, we hope to get our data and tech backend audited and plug any gaps.

Interventions that focus on supporting families and parents have often struggled with reaching out to them. With the increasing penetration of smartphones, message-based services to parents is a cost-effective mode to reach parents. However, these modes often face high churn rates and limited adoption. Further, experiences in other places suggest simple text messages are not effective in improving a child’s learning. 

Our approach is novel in three ways - 

1. It focuses on parents to ensure their child’s foundational literacy and numeracy. We go a step further to ensure that the parents are supported by trained parents from similar communities. This helps in building greater trust and facilitating empathy-based conversations with the parent. 

2. Phone-based conversations, coupled with WhatsApp-based messages allow us to provide specific instructions and bridge information gaps for the parents about the learning content to be used. 

3. Community-based networks - We now have ~50 parents in our team whom we have hired from the communities and who are supporting ~20k parents. This model, implemented at scale would evolve into a peer-to-peer learning model which can be set up in other parts. 

Several community-based organizations are working in the same geographies. Further, several Government bodies work closely with the parents of a child enrolled in a school. Our solution could be easily adopted by these organizations, transforming how we look at parent’s role in a child’s education, particularly in a Government school. Overall, our solution stands for its simplicity, empathy, and scalability. By meeting parents where they are, understanding their challenges, and fostering a culture of peer support, we not only empower individual families but also catalyze broader positive impacts within communities.

Our interventions ensure that children can use learning content appropriate to their learning levels and that parents can support their children in using the content. This is based on a proven methodology  "Teaching at the Right Level", for children studying at the elementary and primary level.

We start by identifying the child’s learning level through a simple assessment of the child conducted over a phone call in the presence of the parent. The results of the assessments are explained to the parents, which helps them understand the difficulties that their child faces in reading and numeracy. Level-appropriate learning content is then shared with the parent over WhatsApp on a daily/ weekly basis to be used by children. This ensures that parents are equipped with the right tools to support their child’s learning journey. Parents also reach out to us to resolve any challenges they face in using the content through our call helpline and WhatsApp chatbot. The cycle of support is repeated every month.

Research and our interactions with parents have revealed that school-based learning alone is insufficient for a child's holistic development. A supportive ecosystem is crucial, with the parent being the central figure and primary point of contact for the child. Studies, including those by Emeagwali (2009), and Epstein (2009), emphasize that effective parental involvement is best viewed as a partnership between educators and parents. A study conducted in Botswana and published by the IMF indicates that combining weekly text messages and phone call tutorials to parents significantly improved learning outcomes for primary school children. The program focused on foundational numeracy concepts and proved to be cost-effective, delivering results equivalent to over one year of high-quality instruction for every $100 spent. The study emphasized the necessity of live, direct instruction via the phone, underscoring the importance of human interactions in providing effective support. In 2019, we together with other organisations, tested the impact of parental engagement on children's learning. Using the ReadAlong application (formerly Google Bolo), which supports children in reading through an online persona, the project aimed to understand how parental participation enhances children's engagement with the application. The project was assessed by a third party. The findings revealed that the support to parents doubled the likelihood of improving children's reading levels. The engagement also led to increased application usage, with children whose parents were supported spending 122 minutes per week, compared to 72 minutes per week for the control group.

Over the next five years, our goal is to impact a million families, focusing on ensuring children achieve foundational literacy and numeracy skills. We use modified ASER tools to measure the child’s foundational literacy and numeracy. The tool is a standard questionnaire that evaluates a child’s reading and math abilities and classifies them into various levels. The child is expected to reach the highest level in both reading and numerical assessments to achieve foundational literacy and numeracy. The data from the ASER tool is widely used in India, Latin America, and Africa. 

There are two ways that we track the data from assessments - i) Internally, the data generated during the support provided by the team members is recorded and internally checked for accuracy. This data is collated at the organization level to track progress. ii) through our partners, we are now planning an external independent evaluation of our impact with our data partners, CDSA. This could be a Randomized Control Trial or other quasi-experimental methods, appropriate for our context.

In addition, we also float surveys on WhatsApp to collect parent feedback around the support we provide - this includes ratings for key elements of our support including the content we provide, Saajhedar’s conversation, and assessments we do.

To achieve our goals, we are focusing on three intermediate outcomes and tracking three key metrics.

i) Parent engagement and retention- We collect the percentage of parents who complete three support instances after completing the first support instance to gauge parent engagement. This is defined as the retention rate. Further, we also track click rates for the content shared to assess the content we share.

ii) Improvement by one level - We track the number of children who improve at least one level in reading and/or math abilities on a real-time basis to track our progress.

iii) Cost per child improved - To ensure that our model is scalable, we are also working on reducing the unit costs. This is defined as the marginal cost incurred to help one child improve in reading/ math skills.

Elements in our core technology can be divided into two categories - 

1. Integrated - these elements have been already integrated and are being iteratively improved through A/B testing and user feedback

WhatsApp chatbot:- Our chatbot serves as a platform for parents to opt-in to our network, receive learning content for their child, and seek responses to frequent questions related to the schools etc. The chatbot is built over a turn.io backend and collects all data related to the parent interaction 

Android Application: We have built and deployed a mobile-based application called Saajha for Saajhedar. The application serves two purposes - it enables our Saajhedars, community champions to place a phone call to the parent for a conversation around support. Secondly, it also helps them capture all relevant data including assessment results, any questions asked or responded to, and call recordings to facilitate a better conversation. The app has been built by Platform Commons, a leading tech-development firm in India. 

Our chatbot and the Android app are integrated at the backend to ensure a seamless experience for the parent.

2. Development - We are also building capabilities in three key areas - 

AI-powered reading assessment - We have partnered with Wadhwani AI to integrate an AI-enabled assessment feature into our WhatsApp chatbot. This tool could analyze a voice recording and predict a level. The tool is currently in the testing phase.

Experimentation engine- Over the past 12 months, we have significantly enhanced our capability to undertake rapid A/B tests. We are now designing an experiment backend to run multiple A/B tests, create organizational memory, and generate evidence for building new features.

Auto-predictor - We have access to 20M WhatsApp messages and more than 45,000 hours of call recordings in addition to several other data points related to parents’ journeys. We are now building a feature to predict the need for support for a parent to improve the journey.

These features are expected to be fully integrated within the next 6-12 months.

114 total team members

~7 volunteers

Our model has been initiated for the past 10 years, and the tech-based intervention was launched in 2021. The app was fully deployed in 2022.

Our team is diverse and has people from different socioeconomic backgrounds, qualifications, religions, and genders. Women constitute an overwhelming majority of our team members, with our Leadership Team and Advisory Board having equal gender representation as well. We were recognised as the Top 50 Workplaces for Women in India (Mid-size organization category) by Great Place to Work in 2023. This has been a result of the consistent and deliberate effort to ensure an ecosystem to enable and support diversity in the team. About 50% of our team members of mothers from similar communities. To ensure that they can integrate, we have policies that allow people to take 24 weeks of paternity/maternity leave (irrespective of gender). Our flexible leave and WFH policy have made it easy for parents, especially mothers to transition back to work. Further, we organize regular sessions with the team to build leadership, plan and implement ideas, and boost self-confidence in the team members. Our policies and processes are documented multilingually to make them accessible to all. To make the selection process transparent and accessible, we make it simple and easy for candidates to reach out to us. Once people are at the organizations, we create spaces for listening to them and create changes in our policies to allow everyone to feel safe and grow within the organization.  

Key beneficiaries and value proposition: Our ultimate beneficiaries are children from low-income households, enrolled in Government school. Parents are our secondary beneficiaries. Our intervention helps the parents in taking necessary actions to improve learning of children.

Customers and value propositions: However, the parents with whom we work have limited paying potential. Our costs are currently covered by corporate philanthropies (50%), which are mandated by law to spend 2% of their profits on social welfare programs. Further, individual philanthropy (20%) and private foundations (30%) invested in the cause are other major customers. Government funds is a potential revenue source that we would leverage in the next two years.

Impact measurement: We track the percentage of children achieving highest levels in the ASER tool, a standard assessment tool.

Partners, and Key Stakeholders - There are three major partners and stakeholders

1. Government - we work with Government schools which are the last resort for quality education. In partnership with them, we reach out to and onboard parents into our network.

2. Technical experts - we have also partnered with Platform Commons and Centre for Data Science and AI, IIM Ahmedabad to strengthen tech and data capabilities. 

3. School staff - we closely work with teachers and principals to reach out to parents.

Key activities: 

There are four types of activities - 

1. Onboarding - We reach out to parents through Parent-teacher meetings. Our team then helps interested parents in completing a simple WhatsApp-based onboarding process.

2. Assess - We reach out to the onboarded parents to assess the learning levels of the child.

3. Advice - Results are then explained to the parents over a call. In addition, necessary information about the relevant content is shared with the parent. Further, advice about other challenges is provided to the parent as well. 

4. Share - Learning content is then shared with the parent via WhatsApp messages. 

For any other challenges, parents reach out to our calling helpline to resolve concerns.

Type of intervention and channels

Our intervention is a personalized advisory service for the parent focused on the child’s learning. This is delivered by a hybrid channel comprising a trained parent from a similar community and a tech backend which includes a calling solution and Whatsapp chatbot.

Key resources:

1. Saajhedars - We have ~50 parents from the communities where we intervene to help deliver interventions effectively.

2. Tech and data expertise - Our tech and data solutions enable easier outreach to the parents cost-effectively and provide appropriate support.

3. Regulatory buy-in - Buy-in from the local government bodies is essential for outreach

4. Funds - We raise funds to cover the cost of the Saajhedars, program management, and tech/ data expertise.

Cost: ~50% of the costs is parent support cost and includes the cost of calling WhatsApp, and acquisition. 15% cost is for the program design, while ~25% is for tech and data. The remaining 10% is for org building. Tech and data costs would not escalate at scale, while we are attempting to reduce the parent support cost by integrating AI.

We rely on three big bets to ensure we can sustain our model as we scale the solution:-

• Strengthen the existing network of supporters:- We have built strong partnerships with our corporate funders and foundations that have supported our work. We would hope to grow these partnerships. These partnerships enabled us to raise ~$500k during the last financial year.

• Reduce unit costs - Our current unit cost of intervention is ~$100 per child per year. Through experiments and tech development, we are hoping to bring this down to $12 per child per year. This will allow us to scale much more rapidly without incurring a proportional increase in the cost. Enabling greater integration of AI, onboarding volunteers for support, and other initiatives for engagement are some of the key levers for reducing costs.

• Government adoption - Once our model is validated, we are hoping to integrate our intervention with the Government schooling system. This would allow us to build systemic capabilities to support parents, thus, ensuring more sustainable solutions for the parents. We are already partnering with the Delhi Government and are hoping to replicate it in other states using Delhi as a proof-of-concept.