There is a large non-profit church frequented by many visitors. Donations are currently done by visitors who donate money to collection box. You need to launch an existing paypal like collection product for church donation instead of the current mode of donation box. Design an A/B experiment to test results and complete the launch within three months.

Interesting A/B test question. Here, I believe the real question is once we have implemented a Paypal method to enable people to donate, how are we going to motivate them to donate more. Essentially, we want to test out ways motivation could be increased, or retention basically. (you could say that yes).

And to gain further undersatnding, this is basically not just creating a paypal address to send money to but more so the church having it's own website on which you can click Donate button and it connects with your paypal for financial transaction. (correct).

ok. having understood this, I think my hypothesis is that we need to touch upon a human aspect to motivate people. Just having a donate option online won't get people to donate since 1) they forget 2)there is nothing really motivating them. (fair enough).

I think we should gamify this by enabling dashboard that tracks top 10 doanners for instance. I didn't say top 3 since that's super competitive and doesn't account for different level of financial status amongst community quite well. Yes, 10 is an arbitrary # but let's go with that. (ok sounds fine).

So the dashboard would present the leader board and that could motivate people to donate more. #2 thing we could do is notifications (leadership status / ranking changed, reminding to donate if not donated within x amount of time, a brand new donner has been identified, etc.). What this would do is silent reminders and also the reminder of new donner provides a message that more and more people are donating, maybe I should as well donate more. #3 thing is campaing where 50% of $1 of donation can be donated to chosen charity.

Since we don't have as much time, if you are ok let's focus on just one of these. (sure. But which one do you want to focus on?)

#3 is a great idea and while we have 3 months, it may not be enough time to establish partnership, all the legality that may go behind it, etc. So let's skip that. I think we should start with #1, leadership board / dashboard. #2 is easy to implement but I think the motivation is driven by gamification mainly not so much reminders. Reminders many times could be annoying and that could turn off people. So let's focus on #1. (ok)

I think we should choose a control group and an experimental group (a group that is to be presented with new feature) but for 2 sets of users. 1 set would be users who are leaders and another set who don't donate as much and are small contributors. Actually, I am sorry but I am thinking a bit differently. Can Ichange course? (sure. What do you have in mind?)

Well, I did say 3 months may not be enough time for establishing enough partnerships but I think we can start small. We can prob spend a month on establishing needed partnerships. We don't have to have 1000s of them to start with. + since these are visitors of the church, not neccessarily members (that's the part I overlooked), I doubt leaderhip board really matters here. Am I correct in my assumption here? (yes more so now then before but I am glad you recognized).

I think this is imp since visitors may wonder how their money is being used. While there may need more to be done to provide 100% trust, this is step in the right direction on establishing some sense of trust of the use donations. Here, I would setup a control group and an experimental group. Control group can only donate while experimental group can will be shown that 50% of their contribution will be donated to a charity from the list. Peerson can even choose upto 3 charities let's say (25% to 1, 50% to another, 25% to another, etc.). Testing this for 1.5 months should give us enough data to establish our conclusions assuiming this chruch is frequented quite often. We can tehn measure which group we obtained more contributions from. If we see that experimental group donated more then we know that this is a good feature to implement. I predict we won't see a ton of increase but we will likely see some increase. This is because there are set of users who will question "how do i know that 50% of my donation is actually being contributed to the cause of my choice?"

The downside is definitely that aspect and for that we would have to think of how we can provide transparency when their money goes towards that organization of choice.

Any questions? (no this is good. Thank you).