Varsha Venugopal

And that's the kind of thing you don't get in an RCT.

So that's something we want to go there and interrogate further.

That's a good question.

And honestly, we do not know.

In the study itself, there was some most enthusiastic people in the J-PAL research, which put up posters and they deliberately removed them from their study to ensure they could focus on the intrinsic gossip element.

Exactly.

So it seems something about this intrinsic motivation ambassadors by just being influencers

are able to take on that role.

How they actually share that information is not something we have figured out yet, except that they seem to be able to do it.

Somewhere between 3% and 7%.

Yeah.

Yeah, so we know SMS reminders, you're absolutely right, are easier to scale up.

In fact, in one of the two states we work in, Maharashtra, we already have a memorandum of understanding with the state government, which allows us to rapidly scale up across the state.

Having said that, we are quite aware that it's the Ambassador Program, which seems to be having some of these large effect sizes, and we're really excited to overlay it on top of the SMS Reminders Program.

With that in mind, we are now introducing Ambassador Program in Maharashtra, we hope.

And in the other state, Bihar, we have already completed this pilot and we just this week have also started exploring an additional district, Patna.

I think for us, the fact that we've already moved to a remote version makes it much cheaper than the original version.

So that's one.

So the fact that we are now calling up these people makes the overall cost effectiveness higher, we would think, depending on what the actual impact is.

So the big focus for us is to make the ambassadors as scalable as possible.