The Watsi Experiment


Along with Wikipedia, Watsi is probably one of the best uses of the Internet I’ve seen.

Disclaimer: I have no association with Watsi, except as a user.

Watsi specialize in low-cost, high-impact healthcare funding. 100% of donation money goes to a specific patient of your choice, and they are ridiculously transparent with their finances (there’s a google doc linked from their website containing all their operations and financials, including screenshots of bank transfers for each operation). In short, it’s hard to imagine a better way to fund healthcare around the world.

Last Christmas, I decided to conduct an experiment. I wanted to give people an opportunity to use Watsi, but instead of giving a large amount to a few people, or a medium amount to some people, I decided to give a small amount to many people.


I did not make a clear hypothesis back then. I was generally curious about how many people would give Watsi a try, and how many of those would find it compelling enough to use again. As one person put it,

My inner engineer can’t help but appreciate that it’s likely to amplify your original donation by roping in some more people too!

This was my intution too.


Around Christmas time, 16 friends, family and acquaintances got a small dollar dollar gift card ($5 for most, and $20 for family), along with an email explaining Watsi and that this was part of an experiment.

This was not completely thought out. I had opted people into something without their permission, and several things were unclear (what is the experiment exactly? what if I don’t want to participate?). After some feedback I sent a second email to clarify that if they didn’t use their gift certificate within three months, the money would go to Watsi’s Universal Fund.

Three months later, I sent a second email asking people to fill out a short anonymous 30 second survey. I got a response rate of 11/16. The questions I asked were the following:

  1. Had you used Watsi before?
  2. Did you use your gift card?
  3. On a scale of 1-10 (10 being certain), how likely is it that you will donate to Watsi again within the coming year?
  4. Comments/suggestions/critiques (optional)

The responses were as follows. Only 1 person had heard of Watsi before, 2 people did not use the gift certificate, and the answers to the third question was: 2-3-3-4-4-6-6-6-7-8-10.

That only 1 person had heard of Watsi before came as a surprise to me, and indicates that Watsi is still very small - there’s already a sample bias in the people I chose, as a lot of them are involved in tech and the startup world.

Assuming that those who didn’t respond to the survey didn’t use it, implies that more than half used their certificate, which is still pretty good. There were however some problems. One person tried to use the gift card two or three times, but wasn’t able to claim it without being logged in. Setting up an account did not help the matter either. This is unfortunate and seems to be a, hopefully temporary, bug in their system.

The third question is a trick I learned from my marketing class many years ago. The idea is that in such questions, only 9 and 10 constitutes a real yes. If we are being generous, we could call a 7 or 8 a maybe, but the rest are best treated as a no. This means we have 1 yes and 2 no, out of 16 people.

This last result was surprising to me - I would’ve assumed this number would be higher. I’ll talk more about this in the next section.

Here are some other comments that people made.

I donated in my mom’s name

Tried to set up my own regular donation but I was not allowed to give a lump sum and too much was lost in currency exchange when transfers were made monthly. I contacted them for alternatives such as pay pal but no joy.

I’m impressed by the follow up with information sent to me about the receiver of my donation.

Not sure I’d do it this year – I think it would be more successful if the campaign was more connected to something that meant to me. Like educating inner city kids etc.

Yeah – mostly I’m just super disconnected from it.

I’ve done a lot with Kiva, and honestly I’m more likely to continue with that.


This was a successful experiment in that we could test the hypothesis, even if it was ill-specified. It turns out that people are quite willing to use a gift certificate, but not as interested in using Watsi again.

Given my own high opinion of Watsi, and their amazing growth, which indicates that they are doing something that people really want, what went wrong?

The biggest error source is the way the experiment was set up. There was a lot of unclear and bad communication on my part, and there was absolutely zero intent among the people surveyed. This was something I didn’t expect, but it makes a lot of sense. People have different priorities, are at various stages in their lives, and have a limited amount of attention. Consider this example: people want cars. It’s a huge industry, people need to get from point A to B. But if you were to semi-randomly ask people if they want to buy a car from a certain brand within the next year, most people would say no. Even adjusting for the different magnitude in purchase decision, that alone is enough to explain how few people would use it again.

In fact, despite my initial disappointment, that I got ~10 more people to know about the existence of Watsi, and that 1 out of 16 will try Watsi again (a number that would be 2 if they fixed their European payment options), should be seen as a success, albeit a small one.

Future work

I see two interesting directions. One is to set up a better experiment with a bigger sample, and make it opt-in, as opposed to opt-out like my experiment was. The other is to explore the concept of intent. For example, what would happen if whenever someone expresses intent of helping people with healthcare, or talks about wanting to help people in poor areas, they would get invited to try Watsi for $5? Of course we have to be careful getting too spammy, but the fact that (a) 100% goes to the person and (b) it’s real money, would mean that the risk of that would be almost negligible.

There are many opportunities that open up with the Internet, and Watsi is a great model for charity in the 21st century.

If you want to fund healthcare for people all around the world, go to