Why did MyGiving need to be replaced? And why is the EffectiveAltruism.org replacement so bad?

post by Holly_Elmore · 2019-10-04T19:35:31.716Z · score: 31 (17 votes) · EA · GW · 9 comments

This is a question post.

Contents

  Answers
    43 Julia_Wise
    2 AlasdairGives
    0 Milan_Griffes
None
9 comments

I don't understand what the rationale was for leaving MyGiving and moving things over to EffectiveAltruism.org. Almost all of the function has been lost if you have to just keep track of the donations on your own throughout the year. I didn't expect that the transition would be completed all at once, but it's been quite a while now. Why was this transition even undertaken if there wasn't the capacity to rebuild the functionality of MyGiving on EffectiveAltruism.org? I'm embarrassed on behalf of EA that this is so bad. It makes us look like we're not that serious about donations and our tech skills are lacking, both of which are nearly hypocritical for EA.

Answers

answer by Julia_Wise · 2019-10-08T14:52:56.001Z · score: 43 (14 votes) · EA · GW

Hi Holly,

(posted as me but written as a collaboration between me, Michelle Hutchinson, and CEA's tech team)

I talked to Michelle (who ran Giving What We Can at the time the old system was put together) and she writes: “The old one was built linking to a CRM [i.e. a database of users] which was pretty difficult to use and which we chose at the time for being free and on a platform well-known to our then tech-lead, who long since moved on. It was built by a volunteer, who continued to help maintain it. He did a stellar job, but that’s not sustainable long-term: particularly because it contains such sensitive financial information and so keeping its various pieces up to date in order to prevent security breaches is incredibly important. It also didn't link in to any of the other systems - the main Giving What We Can website, the new CRM or EA Funds.”

In 2017 the old My Giving platform needed to be replaced because the old system was at the end of its life, and we didn’t have the in-house technical expertise to maintain it safely (e.g apply security patches) or extend its functionality (e.g. automate donation reporting through the Giving What We Can Trust, or its successor, EA Funds). As we were actively working on the EA Funds platform (for which we had in-house expertise, and runs on a significantly more up-to-date and maintainable tech stack), and because both EA Funds and Giving What We Can are about donating effectively, it made sense to roll them into the same system. The aim was to overcome some of the limitations of the old system, while striving for feature parity with it. Unfortunately, we didn’t allocate enough time to the project, and competing priorities meant that the final parts of the migration (along with more extensive user testing) didn’t happen when they should have.

Some functionality is decreased compared to the old version.

  • In order to add or edit recurring donations, you currently need to email us and we can make any changes you request to your recurring donations. In the old system, users were able to do this themselves. This is something we’re currently working towards fixing.
  • Only people who have made a Giving What We Can pledge (either the full Pledge or Try Giving) can use the system, while the old system could be used by anyone.
  • There are no longer graphs breaking down giving into categories, or estimating the real-world impact of donations (e.g. number of bednets distributed from donations to AMF). This is largely because members give to a much wider range of organizations than they previously did, and we’re not able to keep up-to-date with estimates of the impact of each organization (especially now that Giving What We Can doesn’t have an in-house research team).

Some functionality is added in the new version.

  • Donations from EA Funds automatically port in instead of needing to be recorded manually.
  • You can now view your pledge progress both by “overall pledge progress” (your lifetime pledge) and by income period (typically the financial year for which you reported income). This means you can see the total amount you’ve given, and also see that amount broken down over time.
  • You can set income periods to be anything you want (i.e. you don’t just have to report for calendar years).
    • For example, imagine you were a student from January to May, and then employed from June to December. You could enter those as separate income periods, making it easier to see what your pledge would be for each different income situation. The platform will even make sure that you’re only pledging 1% during times that you’re a student or unemployed.
  • You can record donations made in different currencies, and we’ll convert them to the currency your income was reported in. You can also now choose to show your overall pledge progress in any currency you’ve used to report income.

As to why improvements haven’t happened faster, our (usually) two-person tech team has been balancing a number of projects over the last few years:

  • Rebuilding and improving the Giving What We Can donation dashboard
  • Developing and maintaining the EA Funds
  • Developing and maintaining the new EA Forum
  • Developing and maintaining the donation lottery
  • Technical side of EA Global applications
  • Adding ability to donate cryptocurrency
  • Maintenance of givingwhatwecan.org, centreforeffectivealtruism.org, eaglobal.org, and effectivealtruism.org

I think it’s likely that CEA as an org has underinvested in tech capacity for various reasons, and I apologize for the slowness of improvements to the donation platform.

comment by Holly_Elmore · 2019-10-09T02:23:09.061Z · score: 18 (11 votes) · EA · GW

I really appreciate this! Thank you! And I feel lucky to get any free tools like this. I was just irked because I didn’t understand the need for the change. I feel much better about the loss of the recurring donations functionality now that I know the old platform was at the end of its life.

answer by AlasdairGives · 2019-10-07T15:46:38.676Z · score: 2 (6 votes) · EA · GW

There are two long term goals being pursued here by CEA, visible in the design of the site:

1. To increase donations to the EA funds from GWWC members by making the funds the "default" option" and thus increase the importance/power of CEA to guide donations through the funds. (The whole new site is setup to make the funds the default way to give, and to give prominence to the funds, other donation options or recording external donations are much less visible and hidden away in a way that seems deliberate)

2. To set defaults for donations through the site that nudge people towards the cause areas that CEA leadership largely favors and away from cause areas that GWWC was historically focused on. (By default - sliders on the new site allocate the majority of donations made from the pledge page to causes other than global poverty. And this resets as the default every time, with no option to change this or set persistent cause area preferences).

The site is designed with the interests of CEA and what it thinks is best in mind. The functionality for GWWC members is not the priority.

comment by Julia_Wise · 2019-10-08T14:28:31.933Z · score: 34 (11 votes) · EA · GW

[Edited: I missed some corrections that Michelle made to my paragraph about the history of the Giving What We Can Trust. Corrected now.]

I spoke with Michelle Hutchinson (former executive director of Giving What We Can) about this. She writes, “When we first set up the GWWC Trust, we assumed it wouldn't get much use (we set it up on an account designed for an annual turnover of £10k pa), and within a year it was getting up to £1mn. It turned out many GWWC members actually valued a low cost way of giving (in terms of decisions and of how easy it was to give) a bunch more than we expected. Making EA funds easy to use and prominent seems responding to that need.”

The Giving What We Can Trust was a separate legal entity to CEA and was legally restricted to only being able to direct funds within global poverty (since the Charities Commission prefers charities to have narrow focus areas and it wasn’t expected to get extremely wide take up). Given that it did seem to be widely used, and many members wanted to be able to donate to charities other than those tackling global poverty (particularly those joining after 2014 when the pledge became explicitly cause neutral) it became clear that it would be better to have a broader tool than the Giving What We Can Trust. EA Funds provides such a platform. I’d note that while CEA is the operator of EA Funds and does have final sign-off on grants for legal reasons, grant recommendations are made by each Fund’s respective management team, who are independent of CEA.

As to the breakdown of suggested donations, the default split I see on the EA Funds site is:

The default split is intended to provide an example that roughly reflects the donation patterns of both Giving What We Can members and the many other EAs who use the EA Funds. Of course users can adjust the sliders however they wish. You’re correct that the default does not reset based on users’ past donations.

Compare this split to the 2018 EA Survey analysis [EA · GW] of EA's responses to "this cause should be the top priority."

If anything, the ways the EA Fund options are unrepresentative of users’ preferences are

  1. in not having a fund for climate change, and
  2. in under-promoting the Long-Term Future Fund and EA Meta Fund compared to how many EAs believe those areas should be the top priority.
comment by SamDeere · 2019-10-08T15:07:04.169Z · score: 19 (7 votes) · EA · GW

To add to this, I re-analysed the EA Survey responses on cause areas, restricting to just Giving What We Can Members:

Obviously there's a selection effect where the members who take the survey are more likely to be more involved with EA, but I think it's still instructive that Giving What We Can members are a fairly broad church with respect to cause areas, and that it's reasonable to offer different cause areas to them as a default setting on EA Funds.

Disclosure: I work at CEA, and am the person primarily responsible for both EA Funds and the technical implementation of the new Pledge Dashboard.

comment by AlasdairGives · 2019-10-08T19:37:54.192Z · score: 3 (7 votes) · EA · GW

I don't feel either of these reply's address my points very well (as a member who signed the pledge prior to 2014).

As far as I can tell you accept the first point I made and don't address it. Ok, me. I think the funds are fine you just haven't done the work of showing they are better than other donation routes at all.

In regards to the second point you get very fixated on the default slider setting being representative of the most engaged members of the community. I don't want yet more peer pressure to donate to what the most engaged members of think.(And the fact that you unilaterally changed the pledge still shouldn't invalidate my reasons for signing it). But, even for post 2014 members it takes a lot of chutzpah to just set a default - at the very least even if you are recommending the EA funds start each bar at 0%. It is such a dark pattern when to donate to causes you by default get a recommendation you didn't choose then to find other choices you have to click a "back" arrow to a page you have never seen before, go past a statement saying we recommend these funds then manually unclick/click each choice then click forward again. It's so telling when you do the sliders don't start at some nudge level but at a level the user can choose!

comment by Julia_Wise · 2019-10-10T17:43:43.929Z · score: 17 (5 votes) · EA · GW

Sorry, I’ll try again.

  1. It’s true that we try to provide a default option for giving, because so many users seem to find that helpful. (See Michelle’s comment above on the surprising-to-us amount of use the Giving What We Can Trust got.) When we did charity research and recommendations, those recommended charities were also a suggested default. As a project with the mission of inspiring giving to the world’s most effective organizations, we do think it’s appropriate to provide a recommendation or default, with the knowledge that members have pledged to donate to wherever they believe will most effectively help others. (I acknowledge that those of us who pledged before the Giving What We Can became cause-neutral pledged with a different wording that was then specific to global poverty.) We understand and expect that members will make their own choices about where to donate.

  2. When I want to make a donation outside the EA Funds, I do so (for example at againstmalaria.com) and then report it on https://app.effectivealtruism.org/dashboard/pledge by clicking the “report a donation” button. This is the second of two buttons, and I agree that the “New EA Funds donation” button comes first and is more brightly colored, but I don’t think it’s any harder to select the “report a donation” button.

If I decide to donate using the EA Funds, I agree that the four funds are by far easier to donate to, and donating to other under organizations (under the “Choose Funds / Organizations” button as shown in the screenshot above) is more cumbersome. We want to provide this option for users who want to donate to multiple organizations in a single transaction, or who get a tax advantage by donating the Funds. But I agree that the setup of the EA Funds website is primarily designed around ease of donating to the four funds. If donating to individual organizations via the Funds is too cumbersome, I’d suggest donating to those organizations directly (as all members originally did).

As to the preset defaults on the sliders, I’m not sure of all the decisions that went into setting it up that way. My understanding of the intention is to demonstrate “You can move the sliders around and it will always add up to 100%” rather than trying to strongarm donors into donating in a way they don’t want to (although we did choose defaults that we thought would be broadly reflective of the values of the community). You’re right that currently allocations are not saved between subsequent donations, which seems like it would be an improvement to fix.

Again, we expect Giving What We Can members to make their donations based on their consciences and the basic parameters of the spirit of the Pledge. Thank you for explaining your view, but I think we might have a basic disagreement about whether it’s appropriate to suggest default options to members.

If you’d like to talk more, you can always schedule a call with me here: https://calendly.com/julia-d-wise

comment by AlasdairGives · 2019-10-10T19:42:14.609Z · score: 5 (4 votes) · EA · GW

Thanks, i do think we have a basic disagreement here about design patterns but i appreciate you taking the time to defend and explain your choices.

answer by Milan_Griffes · 2019-10-04T20:09:31.168Z · score: 0 (16 votes) · EA · GW

Naïve homo economicus answer is that the people who designed & carried out the change weren't incentivized to make it good.

comment by Stefan_Schubert · 2019-10-04T20:52:19.701Z · score: 24 (14 votes) · EA · GW

That's armchair conjecture - and you can say that about any supposed failure in an organisation. I don't find this comment helpful.

comment by Milan_Griffes · 2019-10-04T21:15:40.623Z · score: 5 (6 votes) · EA · GW

It's probably true for most organization failures? (See also Moral Mazes [LW · GW].)

It at least seems like "the incentives weren't good" should be the starting point for most analysis like this. (And for cases where it shouldn't be the starting point, there should be a clear story about why the case is exceptional.)

comment by Stefan_Schubert · 2019-10-04T23:23:50.722Z · score: 15 (7 votes) · EA · GW

It's a common hypothesis that a given supposed organisational failure is due to poor incentives, yes. That is well-known. There are also other common hypotheses. I don't think it's useful to merely state one particular well-known hypothesis for why organisations fail in a discussion about a specific case like this. It doesn't add anything.

It's like saying "maybe it's due to self-serving signalling" [edited for clarity: that would be another hypothesis]. Self-serving signalling may indeed explain many phenomena, but it's not very useful to state that as an explanation of a specific phenomenon without giving any evidence pertaining to that specific phenomenon.

Furthermore, under a reasonable interpretation saying that a supposed failure is due to poor incentives entails that the individuals who acted on those incentives did so for self-interested reasons (naïve homo economicus is of course associated with self-interested behaviour). Thus it means that they did what served their own interests rather than the common good. I wouldn't make such sensitive conjectures unless I had evidence that that is indeed true of the case at hand.

comment by Holly_Elmore · 2019-10-05T01:00:24.687Z · score: 23 (6 votes) · EA · GW

Overgeneral though this comment is, it does seem to me like GWWC and donations are really getting the shaft from EA Uber-orgs, and that giving simply not being a priority is probably part of the problem.

What I still don’t understand is why they abandoned a perfectly good platform with MyGiving (imo) in order to make an incomplete move to EA.org.

comment by DavidNash · 2019-10-05T14:27:00.141Z · score: 16 (7 votes) · EA · GW

I thought it was so that it tied in with EA Funds, which is something that made me think CEA was paying much more attention to donations, making a unified system that also allowed people to donate from one platform and automatically record donations.

Although I agree that repeated donations not being an option is quite annoying.


9 comments

Comments sorted by top scores.

comment by Julia_Wise · 2019-10-05T08:20:45.664Z · score: 17 (11 votes) · EA · GW

Just a quick note that I'm working on a proper answer but need to check a few details with the tech team once it's a workday.

comment by aarongertler · 2019-10-04T22:37:07.052Z · score: 9 (8 votes) · EA · GW

I'm a former MyGiving user, and I now use the EA.org dashboard. I remember having more problems with MyGiving than I do with the new dashboard, but I'm not really a "power user" in either case, so I may not notice some of the changes. What specific functionality do you miss?

comment by WilliamKiely · 2019-10-05T22:10:08.151Z · score: 27 (11 votes) · EA · GW

Functionality I would like added to the Pledge Dashboard (note: I didn't use MyGiving):

  • A comment field next to each donation. Currently I use the "Recipient" field to write the organization name plus extra notes I want to record (e.g. whether the donation was counter-factually matched).
  • The ability to see total amount I've given to each organization I've given to.
  • A way to label each donation as being associated with a certain cause area.
  • Bar chart of my donations over time, and chart of my donations per organization, and chart of my donations per cause area (by my labeling).
  • The ability to share one's donation page with others.
comment by Cullen_OKeefe · 2019-10-05T23:07:53.219Z · score: 5 (4 votes) · EA · GW

Agreed! CSV exportability would also be good. So would receipt storage/linking for tax reasons.

comment by SamDeere · 2019-10-09T20:39:19.812Z · score: 2 (2 votes) · EA · GW

These are all great suggestions William, thanks for providing them. I'll take them into account as we're making future updates to the platform – no promises on a timeframe re current tech capacity constraints unfortunately, but I think they're all very sensible ideas and would constitute significant improvements.

comment by Holly_Elmore · 2019-10-05T00:55:41.492Z · score: 16 (5 votes) · EA · GW

I’m no power user either. I just want to be able to add and modify recurring reservations, which you can’t do with EA.org. (I just learned you can email them with the details of a recurring reservation to have them add it for you, but come on.) You could do this easily in MyGiving. I also find the EA.org interface very bare, unlike MyGiving. I just don’t understand why they needed to make this move when they weren’t prepared to finish it.

comment by Cullen_OKeefe · 2019-10-05T03:18:22.347Z · score: 3 (3 votes) · EA · GW

Agreed that recurring donation support would be good. But I also like the current interface better aesthetically and, on function terms, equally.

comment by Holly_Elmore · 2019-10-05T13:18:40.847Z · score: 3 (3 votes) · EA · GW

I find it much less intuitive and the aesthetic very cold. I liked the pie chart on my MyGiving dashboard... although I understand how diversifying causes made it easy to break that feature.

comment by Julia_Wise · 2019-10-08T14:34:41.554Z · score: 13 (6 votes) · EA · GW

Thanks, this kind of specific feedback on features you'd like is more helpful than vaguer comments about it being "so bad."