Nudging donors towards high-impact charities (a request for funding for SoGive)

This post describes SoGive’s work to make it easier for donors to find high-impact charities. SoGive is also seeking funds at the moment.

The opportunity
Who plans to do this?
Is there a hidden agenda to this post?
How would SoGive encourage donors to make high-impact donation decisions?
How would SoGive reach its donors?
Why would these clients want to use SoGive?
What’s a realistic sense for the potential scale of SoGive?
How will SoGive determine which are the high-impact charities?
Will nudging actually work?
The team
The budget
What has SoGive done/learned so far?
What are some reasons why SoGive might not be a good thing to fund?

The opportunity

- - Many people consider all charities to be essentially about as good as each other

- - They are often happy to be reactive rather than proactive in charity selection

- - This suggests there may be an opportunity for someone to create a choice architecture where it is easy for people to choose high-impact charities

Who plans to do this?

SoGive. SoGive is run by me (Sanjay) and my cofounder Daniel, together with a small team of (mostly) developers who are working part time. SoGive has been running for about a year, during which time we have been exploring some ways to make an impact-oriented donation website have a positive effect on the world. More about the team below.

Is there a hidden agenda for this post?

Yes (I guess it’s not so hidden now!) We are seeking philanthropic support. Our budget for 2018 is for around £140k-£220k

How would SoGive encourage donors to make high-impact donation decisions?

We would provide donors with software which has recommendations built in. An example prototype can be found here http://sogive.org/ (although the exact software would be tailored for client needs)

How would SoGive reach its donors?

Our plan is to pitch SoGive to specific client groups, each of which represents lots of donors; in other words we find places where donors are already donating and we go to them. These are the three main client categories we’re interested in:

- Organisers of sponsored events (e.g. marathons)

- Corporates (who provide payroll giving for staff)

- Financial advisers (who look after clients’ financial affairs)

Individual users are also welcome, however we think individuals are harder to recruit than clients (i.e. D2C is harder than B2B) so this is not a main focus for our efforts.

Why would those clients want to use SoGive?

SoGive distinguishes itself by having a unique database of charities and the amount of impact each charity achieves, together with costings. This database can

- Tell users how much of charity’s impact is attributable to your donation (to help donors feel proud/positive about their donation)

- Serve as a value for money database

The ability to improve the user experience by making the selection of charities easier is also a positive.

In the sponsored event space, we also use a distinctive charging mechanism, which helps organisers avoid reputational risk.

For financial advisers, we are also developing an exciting inheritance tax product which will enable advisers to continue add value for their clients (and generate fees) beyond the death of the client.

What’s a realistic sense for the potential scale of SoGive’s influence?

It seems reasonable to believe that SoGive will influence c £1-10 million in 2018 and > £10 million in 2019.

To justify this, here are some of the most promising leads in the pipeline at the moment:

- A company has spoken with us about supporting them with charity selection: amount influenced >£4 million; likelihood not clear, will become more obvious over the course of 2018

- An event organiser is potentially going to use the SoGive platform: amount moved >£2 million; unfortunately it’s now clear that recommendations will not be possible for this client

- A company which provides payroll giving for other companies may ask SoGive to source the recommendations: amount influenced >£100 million ultimately, going up gradually; likelihood fairly good

Given that SoGive costs c £100k-£250k to run, a rough ballpark estimate seems to suggest a fundraising ratio of 10:1 for 2018. If SoGive becomes self-sustaining soon, which we think will happen in 2019 or 2020, then the ratio is much higher.

How will SoGive determine which charities are high-impact?

Currently we are leaning heavily on GiveWell’s recommendations. Over time we will take on more of that analysis from other sources than GiveWell, including our own in-house analysis. This is partly because some of our users have interests other than global health, and partly because (depending on our success) we may be at risk of testing the Room For More Funding of GiveWell-recommended charities.

Will nudging actually work?

Our findings so far indicate that most users are drawn towards individual specific charities that are put in front of them; looking past those default charity choices placed in front of them and seeking out a charity linked to their own interests seems to be a rare occurrence, based on our current experience. This experience is not yet at scale.

The research we have found seems to indicate that 70% of donations will be from donors happy to influenced, however this research was not intended to answer this question, and I’m not very confident that this interpretation of that research is correct.

Outside of our experience specifically within the SoGive context, there is plenty of evidence to suggest that people can be positively influenced by their choice architecture.

The team

SoGive is run by Sanjay (CEO) and his cofounder Daniel (CTO)

Prior to starting up SoGive, Sanjay has worked in finance and strategy consulting, whilst simultaneously supporting the management teams of several charities on a voluntary basis, largely as a trustee. In this process, he developed a lot of relevant experience:

· When working as a senior credit analyst at Standard & Poors, he used in-depth analysis of companies’ accounts and complex models to gain deep insight into companies, and then distilled this into highly simplified output; this is when he started thinking about applying this approach to charities

· He also has served on the boards of around half a dozen charities, including roles as treasurer, from which he gained insight into how charity accounting works, how charities use their funds and how they make decisions

· He has also become involved in a large number of groups for philanthropists, building up his network of potential donors/users of SoGive and his understanding of the donor perspective

Daniel Winterstein has a combination of a strong technical background and startup experience, having created Winterwell, the software company he now manages. He has developed successful products and worked as a consultant for companies such as the BBC, Harrods, and as a university researcher.

As a consultant CTO/developer, Daniel played a leading role in the launch of several successful Scottish start-ups: Buddhify, QikServe, AdAppTive, and Yavi.
He is an expert on the development and use of data analysis software to achieve business goals. He has 10 years industry experience – working for organisations including the BBC, Home Office, Ministry of Defence, Tesco Personal Finance, Harrods & Motorola.
A specialist in machine learning, probability theory, mathematical modelling, data science, profiling, text analysis and social networks, he studied Mathematics at Cambridge University and has a Ph.D. in Artificial Intelligence (AI) from Edinburgh University.

Sanjay and Daniel have worked together on SoGive for the last year, and have known each other for around 10 years before that.

The budget

Item	Plan A	Plan B
Charity data analysis+BizDev	£144,292	£89,270
Tech (developers)	£54,574	£32,739
Misc/overheads	£20,875	£18,375
Total	£219,741	£140,383

SoGive needs £140k-£220k to enable us to influence £millions of donations pa. Plan A provides us with a greater probability of growing faster.

What has SoGive done/learned so far?

SoGive considered using the app as a form of EA outreach – i.e. achieving greater scale but less depth than traditional EA outreach methods. This is described here: https://www.facebook.com/groups/effective.altruists/permalink/1211844465538575/

Our experience suggests that this is unlikely to work. We have tried presenting data about charities in a way that is strongly suggestive of value for money thinking, and yet found that users often don’t interpret the data in a value for money way.

Hence our move away from trying influence the mindset of the donor and towards making it easy for donors to find high-impact choices.

What are some reasons why SoGive might not be a good thing to fund?

I will start with the risk that I find to be the most credible concern, and then follow on with others that have been mentioned to me:

While there are potential clients with donation amounts well into the millions lined up, they are not yet confirmed; it is possible that all of them fall through and SoGive achieves only a minimal amount. SoGive response: While this is a risk, the existing pipeline does give grounds for hope.

Is it not better to persuade people to fund more impactful things like the far future (or whichever cause area you think is best)? SoGive response: it’s true that we are unlikely to lead lots of donors towards the most “weird”-sounding charities; our aim is to crowd out the most popular/”accessible” high-impact funding areas so that the community can focus on the more esoteric options.

The SoGive CTO has a lot of expertise in AI – is this project taking away talent that could be used on AI safety? SoGive response: Daniel is sanguine about the risks of AI, and unlikely to be persuaded that working for MIRI is the best use of his time.