Lessons for NGOs exploring a group-based model for beneficiary engagement
The Setup
Key Education Foundation (KEF) runs CLAP — a structured early childhood education program delivered to parents of LKG and UKG children. It covers worksheets, play-based activities, and parenting tips.
KEF’s original model: a 1-on-1 WhatsApp chatbot. Each parent receives personalised content directly.
The new pilot: replace individual chatbot conversations with class-wise WhatsApp groups. Same content, different delivery channel. Groups are automated — messages go out at 4 PM and 7 PM daily.
The pilot ran across 6 schools in Anekal, reaching 601 parents across LKG and UKG.
The Numbers
Engagement held steady in the 28–33% range across all three content types. Here’s how it broke down:
| Activity | Metric | LKG | UKG |
| Worksheet | Fully completed | 29% | 32% |
| Let’s Play | Fully completed | 27% | 30% |
| Parenting Tips | Engaged (watched) | 28% | 32% |
| Parenting Tips | Understood (poll correct) | 19% | 21% |
Two things to note: UKG parents engaged slightly more than LKG. And while engagement on Parenting Tips looks decent in raw numbers, only about 1 in 5 parents actually understood the content (i.e., answered the follow-up poll correctly).
What Worked
1. The group creates social momentum
Parents reported that seeing other parents’ activity in the group motivated them to participate. The visible peer behaviour — someone sharing a completed worksheet, another responding to a poll — created a sense of accountability that the private chatbot never could.
2. Worksheets drove strong, consistent engagement
Parents were highly aware of the Friday worksheet and its support video. Many checked the group daily for this specific content. It became a weekly anchor. This is the stickiest behaviour to build on.
3. Teacher workload dropped sharply
Adding parents to a group is far simpler than registering each one on the chatbot. Teachers also gained visibility — they could now see who was engaging, making follow-up targeted rather than guesswork.
4. Videos bridged literacy gaps
One grandmother who couldn’t read the worksheet watched the support video on the group and understood what to do. The visual format made the content accessible to parents with lower literacy.
5. Engagement was stable over time
Across 14+ weeks, engagement didn’t collapse. The 28–33% range held steady. That’s a reasonable baseline for a new, automated program with no in-person nudge.
What Didn’t Work
1. Non-worksheet content was largely ignored
Despite similar raw engagement numbers, qualitative interviews revealed that Let’s Play activities and Parenting Tips were frequently missed or skipped. Parents knew about worksheets. They often didn’t know the other content even existed.
2. Language was a real barrier
All content was in Kannada. Hindi-speaking parents in Sompura found it inaccessible. Parents also requested English content — not just for comprehension but because they wanted their children exposed to English vocabulary.
3. Content volume felt overwhelming
Two messages per day, across multiple content types, created noise. Parents tuned in for what mattered to them (worksheets) and filtered out the rest. The schedule needs simplification.
4. Smartphone access was a constraint
In some households, only one person owns a smartphone — often the spouse. When that person isn’t the primary caregiver, content doesn’t reach the parent doing the actual childcare.
5. Parenting Tips comprehension was low
Only 19–21% of parents answered the post-video poll correctly. Watching a video is not the same as understanding it. The format or framing of parenting content needs rethinking.
Group Model vs. 1-on-1 Chatbot: A Quick Comparison
| 1-on-1 Chatbot | WhatsApp Group Model | |
| Onboarding effort | High (each parent registered individually) | Low (add to group) |
| Teacher monitoring | Difficult | Easy — visible in the group |
| Social motivation | None | High — peer visibility drives participation |
| Content personalisation | Possible | None — same content to all |
| Language flexibility | Possible | Same barrier for all members |
| Privacy | High | Lower — activity visible to group |
| Scale-up feasibility | Moderate | High |
If You’re Thinking of Running a Similar Pilot
Based on KEF’s experience, here are the decisions you need to make before you start — and how to make them well:
- Pick one anchor activity and build around it. KEF’s worksheet is a weekly ritual parents look forward to. Design your content calendar around one predictable, high-value touchpoint. Don’t try to push everything at once.
- Map your language landscape before you begin. Survey caregivers in your target schools for their primary language. If your cohort is multilingual, either localise content or segment groups by language.
- Limit daily messages to one. Two messages per day across a week creates a high-volume feed. Start with one high-quality message per day and expand only if engagement holds.
- Brief teachers on their role. Teachers are the most trusted adults in the parent’s school relationship. Train them to send one personal nudge per week — this dramatically improves engagement over automated messages alone.
- Design comprehension checks for every content type, not just parenting tips. A poll after every video (not just parenting ones) gives you real data on what’s landing.
- Address the smartphone access gap. Ask at enrollment who holds the family phone. Consider whether a family member (elder sibling, spouse) needs to be added instead of or in addition to the parent.
- Run a 4-week baseline before drawing conclusions. Engagement in week 1 is inflated by novelty. Weeks 4–8 give you a more honest picture of sustained behaviour.
Bottom Line
The group model is not a replacement for personalised programming. But it is a strong, low-cost entry point — especially for organisations that can’t yet afford the infrastructure of a 1-on-1 chatbot at scale.
The KEF pilot suggests that ~30% engaged parents — consistently, over 14 weeks — is achievable with a basic automated group setup. With better language access and a simplified content schedule, that number should be meaningfully higher.
The group is not just a delivery channel. It is a community. Design for that from day one.
Big shout out and thanks to Swarupa Manjunath from KEF team, for running the pilot, providing the insights and allowing us to share this with the community of NGOs.
Please write to info@glific.org to understand more leveraging whatsapp groups engagement model.