A chatbot for an NGO can be a godsend for keeping in touch with beneficiaries on a large scale, especially when one-on-one talks are impossible to continue in the midst of a global pandemic.
NGOs can use chatbots to digitize their programs and distribute them to a wider audience through various channels such as IVR calls, text messaging, WhatsApp, and so on.
To make your chatbot a success, you must constantly enhance it by identifying user trends and adjusting the way the chatbot responds, which might include things like changing the time of the message, giving regular nudges, and sending feedback forms.
In this article, we’ll look at how STiR Education has been utilizing Glific to deliver messages via Whatsapp, as well as some of the trends that they captured in their custom Datastudio dashboard.
STiR Education is a non-governmental organization that works with governments to help them put in place practical methods to foster positive behavior change in teachers and government workers. As a result, the groundwork for children’s lifelong learning is being laid. They’ve been using Glific to send WhatsApp messages to these teachers and government officials, and they’ve started their pilot in two cities: Delhi and Karnataka, with Tamil Nadu to follow.
As illustrated in the Glific Architecture, Analytics/Reports are an important aspect of the Glific ecosystem, where the main application database is multi-tenant. NGOs, on the other hand, can use BigQuery integration to store and access their chatbot data. As soon as an NGO enters their GCP credentials in Glific, data such as Messages, Contacts, and Flows begin syncing to their BigQuery, which acts as a data source to build a dashboard using DataStudio for analytics purposes and capture the user trend.
Some of the graphs which STiR Education used in their dashboard to keep track of their chatbot were:
1. User status graphs: They have divided users into many groups in order to gain a better understanding of all of the users who have interacted with the bot and how well the bot has been welcomed.
- Opted in but not registered: Users who have joined the bot by sending the first message but have not completed the registration process.
- Total registered users: Users who have completed the registration procedure by providing pertinent information such as their name, state of residence, and role
- Active users: Users who have messaged the bot in the last 7 days
- Inactive users: Users that last interacted with the bot at least a week ago
2. Engagement time: Another important metric for a chatbot is the time of the day beneficiaries spent interacting with it the most. This most active period can then be utilized to deliver reminder/nudge messages that are more likely to be received during this period.
3. Invalid response graph: Because conversations in Glific are flow-driven, it’s possible that the user will get stuck if they don’t grasp the message or enter an incorrect response. It’s an important metric because it helps determine which flows people get stuck in the most, indicating which ones need to be improved.
4. Reminder message effect: They also send out nudge messages to beneficiaries on a regular basis to convey critical timely messages and keep them engaged, in addition to flow-driven messaging where the next message is sent based on the user’s prior answer.
Three different reminder messages are being sent, namely:
Inactive reminder messages: Sent to users who are inactive for the previous seven days.
Pending registration reminder messages: Sent to users who have not registered yet with all relevant details
Survey not filled messages: Sent to users who are haven’t completed any survey in the previous 15 days
On the right, it also indicates how many reminder messages were successful, which implies how many people registered after the not registered reminder was sent over the next two days, and how many people became inactive after the inactive reminder was sent.
5. Survey-filled users: Filling out surveys is a vital element of their program, but it’s also crucial to keep track of the trends. As a result, the number of surveys completed across various roles is an essential indicator.
Custom filters, in addition to numerous graphs for capturing various metrics, are an inbuilt component of DataStudio that aids in obtaining more precise data.
state, district, role, gender, and time were some of the filters employed by STiR Education.
These filters aid in pinpointing specific data at a specific point in the bot’s lifecycle.
As a result, with numerous metrics and custom filters, it’s easy to keep track of how well your chatbot is functioning and how well it’s welcomed by end-users. These data also aid in determining which places users interact with the most and where they get stuck, pointing out areas where the chatbot can be improved in order to reach more end users.
And we at Glific will be Happy to help you create it so that your program may reach a wider audience and have a stronger impact.
Leave a Reply