GiveCrypto Monthly Update — May 2020

Refocusing our product strategy and an update on the RCT

Number of people helped — 4,217 (received $34,900)

Operational efficiency — 52.3%

Fraud rate — 2.6%

Refocusing GiveCrypto

We have decided to focus GiveCrypto’s attention on a smaller set of priorities. More details are below, but the tl;dr is that we will work exclusively on direct crypto transfers for the foreseeable future.

We will continue to support the RCT (ending in early July) but will suspend all work on financial service experimentation. While some of these experiments showed promise, we were stretching ourselves too thin and were not innovating enough on direct transfers.

We will resume sharing key metrics in monthly updates. We will provide one north star metric (number of people helped) and two secondary metrics (operational efficiency & fraud rate) in all monthly updates. Here is the current definition of those metrics (note that these definitions may evolve over time):

• Number of people helped — the number of people that received money during the time period
• Operational efficiency — amount of money sent divided by total expenses during the time period
• Fraud rate — percent of money sent to people acting in bad faith during the time period

Above you can see the values of these metrics for the month of May. Note that these values apply to the funds sent via the RCT and that future metrics will only apply to new transfer activities.

GiveCrypto’s New Product Strategy

The changes mentioned above will have the most impact on our product strategy, which was previously focused on developing software that automates various aspects of the Ambassador Program. This includes the targeting, distribution and tracking of ambassador-directed crypto transfers.

Moving forward, we will focus on the development of a crypto transfer marketplace. The eventual goal is a service that allows anybody to apply for help and anybody to provide help. This will look like a typical marketplace; with people in need on one end and donors on the other.

This type of service raises a number of interesting product questions:

• How do we qualify recipients?
• How do we prevent and detect fraud?
• How much information about recipients do we share with the public?
• How do we measure the impact of the funds?
• What technologies/currencies should we use?

We are currently developing an MVP that will help us begin to answer some of these questions. It should be ready to launch towards the end of June.

Update on RCT

We are more than half way through phase 4 of the Ambassador Program. This phase involves a randomized controlled trial comparing direct cash to direct crypto transfers. We recently completed our first round of impact questions and have some preliminary results to share.

Both cash and crypto treatments seem to improve food security and psychological well being compared to the control group. We see cash having slightly more impact than crypto, although the difference is small and we have not yet run regressions on the treatment effect and thus cannot comment on whether these differences are statistically significant. At the end of the study, we are planning to evaluate the final impact of cash and crypto relative to the cost (including logistical difficulty) of sending both types of payments, which will give us a full picture of the impact per dollar of both programs.

Since there are many outcomes in each area (food security and psychological well being), we combined them into one index per area. The indices are always normalized such that the control group has a mean of 0; a higher value indicates better outcomes (i.e. better food security and better psychological well being). These indices tell us the overall, combined movements of the outcome variables in both food security and psychological well being, even if it may be difficult to detect movements from each single outcome variable. Below we show the figures for the food security index and the psychological well being indices.

This graph shows the average food security index in the past 30 days across different treatment groups, where the index is constructed from the four food security related outcome variables and normalized to 0 for the control group. Higher values correspond to better food security. The error bars show 95% confidence intervals of the means.

This graph shows the average psychological well being index in the past 30 days across different treatment groups, where the index is constructed from the seven psychological well being related outcome variables and normalized to 0 for the control group. Higher values correspond to better psychological well being. The error bars show 95% confidence intervals of the means.

We do not want to place too much emphasis on this interim data, but it is encouraging to see the cash and crypto groups showing positive impact on food security and psychological well being. We just started collecting the next round of impact data and should have additional results to share in June’s report.

Staying Connected

As always, we love to hear from our donors and supporters outside of our updates and events. Please follow along on Twitter and our blog for updates on our programs.