Last January, Effective Altruism Netherlands (EAN) became a registered charity in the Netherlands. The organisation consists of a three-person board and two full-time employees, Sjir Hoeijmakers and yours truly (Remmelt Ellen). Note: as of writing, we are still fundraising to cover our salaries.
On 28 May, we publicly launched with a large event. Since then, we have pivoted from supporting EA projects to collaborating with EA networks in the Netherlands. This post outlines our new strategy. We would appreciate your feedback in the comments, to enable us to improve it further.
Terminal Goal
Have as much positive impact as possible on the lives of others.
Instrumental goal
Engage impactful and potentially impactful individuals in the Netherlands with effective altruism and stimulate and facilitate them to bring effective altruism into practice.
Strategy
Work with selected ‘network builders’ to build networks of self-identified effective altruists by
helping define targets of the network
helping set up the network infrastructure
providing resources that differentiate effective altruism and the effective altruism community.
Do targeted EA outreach to fill these networks with qualified individuals
Provide a national infrastructure for these networks to connect with each other to share motivation, knowledge, skills and other resources
Be the point of contact for anything EA-related in the Netherlands to protect the correct use of the concept ‘effective altruism’
Before the pivot, we had a tendency to jump on an appealing opportunity, construct a concrete project framework and then look around for people who could carry it out. It was difficult to find people with a good personal fit and the project’s impact was limited by its duration. We now focus on selecting people first – those who we deem both EA-aligned and highly-capable of building EA networks within promising cause and talent areas, based on their track record and our experience in working with them. Note that with ‘self-identified effective altruists’ we mean network members who identify with and strive after CEA’s Guiding Principles.
Assumptions
We can have more impact by making others more effective than by doing direct work ourselves.
We can have more impact by getting people fully aligned with effective altruism than by spreading parts of the effective altruism idea (e.g. cause-specific optimisation).
Here are two arguments for this:
Impact being created by the multiplication of parameters. e.g. open-mindedness without scientific rigour leads to failure.
There being a log/power law distribution of impact across cause areas, just like there is across interventions within a particular cause area.
We can have more impact by focusing on individuals directly and on organisations indirectly rather than on organisations directly.
The underlying idea is that rather than focusing our efforts on a representative who has more say in the organisation’s direction but is less likely to be particularly aligned with EA values, we should instead focus on people within the organisation who are.
We can have more impact by focusing on (potentially) impactful individuals than on the general public.
With ‘impactful’, we mean people who we deem capable of having an outsized positive impact. The argument here is to spend most of our resources on identifying and targeting those people instead of undirected broad outreach in the hope that a tiny portion of people who hear from us work their way up the ladder of engagement to become EA-aligned.
We can have more impact by focusing on the Netherlands than on other geographic areas.
In the Netherlands, only a tiny minority know of the term ‘effective altruism’, meaning that there are still many low-hanging fruits to pick. We also have a geographic advantage because we speak Dutch and are able to connect in-person in this densely populated country. More tentatively, Dutch culture seems suitable for effective altruism, as the Dutch are seen to be pragmatic, cosmopolitan/ outward-looking, and consistently rank high at philanthropic giving.
Metrics
These are key metrics selected to track EAN’s underlying impact (i.e. wellbeing-adjusted life years) as best as we can to compare with the opportunity costs of the money spent by donors and the time spent by EAN. To test and adapt our strategy, we are also setting falsifiable hypotheses for each link in the chain of processes needed to build impact.
Inputs:
Time committed by EAN
Money spent by EAN
Outputs
Behavioural change: impact-adjusted significant plan changes (IASPC) attributable to EAN/ €1,000
As an ambitious benchmark for how well we’re doing, we intend to use 80,000 Hours’ trajectory after it officially started in Nov 2011. Their current output is around 4 IASPC/€1,000 spent (probably an overestimate). If we assume that this figure stood at 1 IASPC/€1,000 in its first year, that’s what we should be aiming for now for the networks we’re helping build (a more rigorous approach would also include estimated salary opportunity costs).
Belief change: increase in the number of self-identified effective altruists
This is about tracking whether people within the EA Networks have an applied understanding of CEA’s Guiding Principles, possibly through periodic surveys. This is to ensure that the behavioural change we see is not fragile. i.e. that people within the Dutch EA community are able to update with new evidence over the next decades.
Current collaborations
Here are three networks that we’re starting collaborations with (network builders in brackets):
Quote: “Effective Giving is a community of foundations and large philanthropists learning together how to apply our unique resources to make the maximum contribution to a better world.” In our view, this is an example of a well-functioning network with which we have had fruitful collaborations (Kellie and Robert have also supported the development of EAN from the start).
This is a network that develops and gives career workshops in the Netherlands and would help connect potentially impactful people to other Dutch EA networks. The biggest bottleneck right now to Alje and Yaël working on this at least part-time is building a sustainable revenue model.
Quote: “RAISE… is an initiative to improve the pipeline for AI safety researchers. The road to research-level understanding is hard and uncertain. We want it to be easy and clear. To this end, we are creating a MOOC.” This network has just started recording its first lecture and is still looking for volunteers.
Here are potential network collaborations that we’re exploring right now:
Local/student groups These would clearly define their target groups and offer newcomers a path to learn about EA principles and build up their capacity to do good (established groups like EA London and EA Berkeley are inspirations to us here).
Animal Welfare Research network This network would compile and produce rigorous animal welfare intervention research that can be applied by Dutch and international charities.
Rationality hub This network would enable (potentially) impactful individuals in the Netherlands to improve their epistemic and instrumental rationality.
Startup Founders network This network would support the creation of effective for-profit and nonprofit startups.
Corporate Intrapreneurs network Similar to the work of EA Workplace Activism, these would be entrepreneurial employees who spread EA thinking within companies through EA projects, charity drives, changes in pension funds investments and the like.
Policy network This network would search for the most high-impact lobby opportunities for improving Dutch government or EU-wide policy and start projects to implement them.
Examples of actions
Concretely, here are examples of actions that EAN would and would not take based on this strategy:
Would do:
Sit down with a high-potential and highly committed person working in a corporation to set specific targets (e.g. moving money of corporations’ foundations) and plan out when and how their Corporate EA network would communicate (e.g. arrange meeting spot, online platform) to reach these targets.
Suggest to a Startup Founders network to organise a cause prioritisation training and connect them with an experienced trainer or online self-learning material for this.
Connect an AI Safety Research network with a Policy network through a shared event where the former informs the latter about promising policies to pursue.
Would not do (although someone in a network could):
Run a campaign to promote GiveWell-recommended charities.
Do research into the most effective ways to lobby the Dutch government or the European Union.
Incubate an effective NGO startup.
Work with existing organisations or networks to improve their impact
To avoid becoming long-winded in this post, we left out tactical details. Feel free though to write out your questions and feedback below to improve this strategy.
Testing an EA network-building strategy in the Netherlands
Last January, Effective Altruism Netherlands (EAN) became a registered charity in the Netherlands. The organisation consists of a three-person board and two full-time employees, Sjir Hoeijmakers and yours truly (Remmelt Ellen). Note: as of writing, we are still fundraising to cover our salaries.
On 28 May, we publicly launched with a large event. Since then, we have pivoted from supporting EA projects to collaborating with EA networks in the Netherlands.
This post outlines our new strategy. We would appreciate your feedback in the comments, to enable us to improve it further.
Terminal Goal
Have as much positive impact as possible on the lives of others.
Instrumental goal
Engage impactful and potentially impactful individuals in the Netherlands with effective altruism and stimulate and facilitate them to bring effective altruism into practice.
Strategy
Work with selected ‘network builders’ to build networks of self-identified effective altruists by
helping define targets of the network
helping set up the network infrastructure
providing resources that differentiate effective altruism and the effective altruism community.
Do targeted EA outreach to fill these networks with qualified individuals
Provide a national infrastructure for these networks to connect with each other to share motivation, knowledge, skills and other resources
Be the point of contact for anything EA-related in the Netherlands to protect the correct use of the concept ‘effective altruism’
Before the pivot, we had a tendency to jump on an appealing opportunity, construct a concrete project framework and then look around for people who could carry it out. It was difficult to find people with a good personal fit and the project’s impact was limited by its duration.
We now focus on selecting people first – those who we deem both EA-aligned and highly-capable of building EA networks within promising cause and talent areas, based on their track record and our experience in working with them. Note that with ‘self-identified effective altruists’ we mean network members who identify with and strive after CEA’s Guiding Principles.
Assumptions
We can have more impact by making others more effective than by doing direct work ourselves.
The main argument for this is the multiplier effect.
We can have more impact by getting people fully aligned with effective altruism than by spreading parts of the effective altruism idea (e.g. cause-specific optimisation).
Here are two arguments for this:
Impact being created by the multiplication of parameters.
e.g. open-mindedness without scientific rigour leads to failure.
There being a log/power law distribution of impact across cause areas, just like there is across interventions within a particular cause area.
We can have more impact by focusing on individuals directly and on organisations indirectly rather than on organisations directly.
The underlying idea is that rather than focusing our efforts on a representative who has more say in the organisation’s direction but is less likely to be particularly aligned with EA values, we should instead focus on people within the organisation who are.
We can have more impact by focusing on (potentially) impactful individuals than on the general public.
With ‘impactful’, we mean people who we deem capable of having an outsized positive impact. The argument here is to spend most of our resources on identifying and targeting those people instead of undirected broad outreach in the hope that a tiny portion of people who hear from us work their way up the ladder of engagement to become EA-aligned.
We can have more impact by focusing on the Netherlands than on other geographic areas.
In the Netherlands, only a tiny minority know of the term ‘effective altruism’, meaning that there are still many low-hanging fruits to pick. We also have a geographic advantage because we speak Dutch and are able to connect in-person in this densely populated country. More tentatively, Dutch culture seems suitable for effective altruism, as the Dutch are seen to be pragmatic, cosmopolitan/ outward-looking, and consistently rank high at philanthropic giving.
Metrics
These are key metrics selected to track EAN’s underlying impact (i.e. wellbeing-adjusted life years) as best as we can to compare with the opportunity costs of the money spent by donors and the time spent by EAN. To test and adapt our strategy, we are also setting falsifiable hypotheses for each link in the chain of processes needed to build impact.
Inputs:
Time committed by EAN
Money spent by EAN
Outputs
Behavioural change: impact-adjusted significant plan changes (IASPC) attributable to EAN/ €1,000
As an ambitious benchmark for how well we’re doing, we intend to use 80,000 Hours’ trajectory after it officially started in Nov 2011. Their current output is around 4 IASPC/€1,000 spent (probably an overestimate). If we assume that this figure stood at 1 IASPC/€1,000 in its first year, that’s what we should be aiming for now for the networks we’re helping build (a more rigorous approach would also include estimated salary opportunity costs).
Belief change: increase in the number of self-identified effective altruists
This is about tracking whether people within the EA Networks have an applied understanding of CEA’s Guiding Principles, possibly through periodic surveys. This is to ensure that the behavioural change we see is not fragile. i.e. that people within the Dutch EA community are able to update with new evidence over the next decades.
Current collaborations
Here are three networks that we’re starting collaborations with (network builders in brackets):
Effective Giving (Kellie Liket & Robert Boogaard)
Quote: “Effective Giving is a community of foundations and large philanthropists learning together how to apply our unique resources to make the maximum contribution to a better world.”
In our view, this is an example of a well-functioning network with which we have had fruitful collaborations (Kellie and Robert have also supported the development of EAN from the start).
Career Workshops (Alje van den Bosch & Yaël Duindam)
This is a network that develops and gives career workshops in the Netherlands and would help connect potentially impactful people to other Dutch EA networks. The biggest bottleneck right now to Alje and Yaël working on this at least part-time is building a sustainable revenue model.
Road to AI Safety Excellence [RAISE] (Toon Alfrink)
Quote: “RAISE… is an initiative to improve the pipeline for AI safety researchers. The road to research-level understanding is hard and uncertain. We want it to be easy and clear. To this end, we are creating a MOOC.”
This network has just started recording its first lecture and is still looking for volunteers.
Here are potential network collaborations that we’re exploring right now:
Local/student groups
These would clearly define their target groups and offer newcomers a path to learn about EA principles and build up their capacity to do good (established groups like EA London and EA Berkeley are inspirations to us here).
Animal Welfare Research network
This network would compile and produce rigorous animal welfare intervention research that can be applied by Dutch and international charities.
Rationality hub
This network would enable (potentially) impactful individuals in the Netherlands to improve their epistemic and instrumental rationality.
Startup Founders network
This network would support the creation of effective for-profit and nonprofit startups.
Corporate Intrapreneurs network
Similar to the work of EA Workplace Activism, these would be entrepreneurial employees who spread EA thinking within companies through EA projects, charity drives, changes in pension funds investments and the like.
Policy network
This network would search for the most high-impact lobby opportunities for improving Dutch government or EU-wide policy and start projects to implement them.
Examples of actions
Concretely, here are examples of actions that EAN would and would not take based on this strategy:
Would do:
Sit down with a high-potential and highly committed person working in a corporation to set specific targets (e.g. moving money of corporations’ foundations) and plan out when and how their Corporate EA network would communicate (e.g. arrange meeting spot, online platform) to reach these targets.
Suggest to a Startup Founders network to organise a cause prioritisation training and connect them with an experienced trainer or online self-learning material for this.
Connect an AI Safety Research network with a Policy network through a shared event where the former informs the latter about promising policies to pursue.
Would not do (although someone in a network could):
Run a campaign to promote GiveWell-recommended charities.
Do research into the most effective ways to lobby the Dutch government or the European Union.
Incubate an effective NGO startup.
Work with existing organisations or networks to improve their impact
To avoid becoming long-winded in this post, we left out tactical details.
Feel free though to write out your questions and feedback below to improve this strategy.