> For the complete documentation index, see [llms.txt](https://research.ambitiousimpact.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://research.ambitiousimpact.com/explainers-and-house-views/our-charity-taxonomy/guidance-to-adjust-research-for-meta-interventions.md).

# Guidance to adjust research for meta interventions

Contributor: Juan Benzo

## Our evaluation approach should vary according to the type of mechanism being evaluated <a href="#our-evaluation-approach-should-vary-according-to-the-position-on-the-explore-exploit-spectrum" id="our-evaluation-approach-should-vary-according-to-the-position-on-the-explore-exploit-spectrum"></a>

* Ideas can be characterized by the [primary mechanism](https://ambitious-impact.gitbook.io/ambitious-impact-docs/explainers-and-house-views/our-charity-taxonomy) in their theory of change.
* Researchers should be mindful of this position when deciding what to front-load and how to approach the research.
* This guidance focuses on how to approach ideas that work on meta-charity interventions.

### Meta organizations&#x20;

Meta-charities have an indirect impact by enabling other charities to better pursue their missions — they don't deliver work intended for beneficiaries directly. Instead they raise the effectiveness of existing organizations or bring new ones into existence.

Meta charities channel three kinds of resources into the ecosystem — money (fundraising & capital allocation), people (talent, founders, community), and organizational capacity (services and infrastructure for orgs).

Note that our taxonomy splits this bucket into two mechanisms: ecosystem services (people and services) and fundraising and capital allocation (money).<br>

<table data-view="cards"><thead><tr><th>Mechanism</th><th>Definition</th><th>How to tell if it belongs</th><th>Examples</th><th data-hidden data-card-cover data-type="image">Cover image</th></tr></thead><tbody><tr><td><strong>Money</strong></td><td>Influencing where philanthropic, retail, HNWI, institutional, or grantmaking capital goes</td><td>Does the org primarily influence allocation of funds?</td><td>Giving What We Can, Effektiv Spenden, GiveWell, Coefficient Giving, Ark Philanthropy, Effective Giving Initiatives Launched by AlM</td><td><a href="https://2926888385-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F3fybH5C1XcqQVgndk58o%2Fuploads%2F6nIQvshTkU0cRPwSqvPn%2Fmoney.svg?alt=media&amp;token=0324ff62-5a2c-419a-9686-e82c3fa9c4ee">money.svg</a></td></tr><tr><td><strong>People</strong></td><td>Channeling talent into the<br>ecosystem: finding, training, and matching individuals, creating new founders and orgs, and building the communities they move through</td><td>Is the counterfactual claim about changed career trajectories, new organizations/founders, or connections — rather than money or services?</td><td>80,000 Hours, Animal Advocacy Careers, High Impact Professionals, AIM, School for Moral Ambition, CEA</td><td><a href="https://2926888385-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F3fybH5C1XcqQVgndk58o%2Fuploads%2FnsTdi2nKwLHetCH5U4hZ%2Fpeople.svg?alt=media&amp;token=4e989180-9331-42cd-9b90-976fbc59f895">people.svg</a></td></tr><tr><td><strong>Services</strong></td><td>Services and shared infrastructure — research, evidence, M&#x26;E, operations, fiscal sponsorship, software — that make charities more effective or cheaper to run. Includes meta-research that steers what the ecosystem works on (cause prioritization, exploratory research)</td><td>Is the output a service or platform that improves a charity's capability — or research that steers where the ecosystem spends effort?</td><td>IDinsight, IPA, Animal Ask, Impact Ops, The Mission Motor, CEARCH</td><td><a href="https://2926888385-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F3fybH5C1XcqQVgndk58o%2Fuploads%2FfGgOmTM2IOcRjqxR2Mci%2Fcapacity.svg?alt=media&amp;token=ffe12f52-2b9b-41dd-b558-4e0010af0ac5">capacity.svg</a></td></tr></tbody></table>

<details>

<summary><strong>Sub-mechanisms</strong></summary>

The General advice section below applies to meta charities in general. However, different sub-mechanisms exist within the three different areas (money, people, organizational capacity). Briefly, below we offer *potential* ways to divide the mechanisms sub-types:

Money:

* Effective giving and pledges (e.g. GWWC, Effective Giving Initiatives)
* Grantmaking, regranting, and donor advisory (e.g. Coefficient Giving, EA Funds)
* Charity evaluation and prioritization research: (e.g. ACE, Giving Green)

People:

* Talent allocation and careers (e.g. 80,000 hours, animal advocacy careers)
* Incubation and founder support (e.g. AIM, school for moral ambition)
* Community and field building / coordination (e.g. CEA)

Services:

* Research, evidence, and M\&E services (e.g. IPA, IDinsight, Mission Motor)
* Operations and Shared infrastructure (Impact Ops, Good impressions, Effective Ventures)
* Meta-research that steers the ecosystems efforts (CEARCH, Rethink Priorities)

Breaking down these sub-mechanisms can be useful across different stages of research. For example:&#x20;

* Sparks new charity ideas
* Helps narrow down central cruxes and crucial considerations
* Helps consider various theories of change
* Helps understand what charities to benchmark this idea against

Note: These categories may overlap and are not exhaustive nor mutually exclusive.

</details>

<details>

<summary><strong>Our Incubated Charities</strong></summary>

The table below highlights the different meta-charities we have incubated to date, including links to the reports. As of August 2026, we have incubated 14 meta charities (3 people, 3 services, 7 money)

<table data-search="false"><thead><tr><th width="127.17578125">Mechanism</th><th width="207.5234375">Report/Site</th><th>Charity Incubated</th></tr></thead><tbody><tr><td>People</td><td><a href="https://www.charityentrepreneurship.com/animal-advocacy-careers">AIM Page</a></td><td><a href="https://animaladvocacycareers.org/">Animal Advocacy Careers</a> (2019)</td></tr><tr><td>Services</td><td><a href="https://www.charityentrepreneurship.com/animal-ask">AIM site</a></td><td><a href="https://www.animalask.org/">Animal Ask</a> (2020)</td></tr><tr><td>Money</td><td>Own Idea (<a href="http://google.com/search?q=giving+green+charity+entrepreneurship&#x26;sca_esv=da0da0d6da6c68d3&#x26;rlz=1C5OZZY_enGB1214GB1214&#x26;sxsrf=APpeQnuUjpvg5MRhAGj4MJaBUqTnNW00dQ%3A1785840446631&#x26;ei=PsNxapOQJuCvhbIPn-DD-Ak&#x26;biw=1511&#x26;bih=779&#x26;ved=0ahUKEwiTp8mL5oaWAxXgV0EAHR_wEJ8Q4dUDCBA&#x26;uact=5&#x26;oq=giving+green+charity+entrepreneurship&#x26;gs_lp=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&#x26;sclient=gws-wiz-serp">AIM Page</a>)</td><td><a href="https://www.givinggreen.earth/">Giving Green</a> (2020)</td></tr><tr><td>Services</td><td><a href="https://3394c0c6-1f1a-4f86-a2db-df07ca1e24b2.filesusr.com/ugd/9475db_a215cf49667a40f7aca07db9c10bd246.pdf">Exploratory Altruism</a></td><td><a href="https://exploratory-altruism.org/">CEARCH</a> (2021)</td></tr><tr><td>People</td><td><a href="https://docs.google.com/document/d/14m1wvrup7whaRNHfLcjbvxyL5XvVgas7KIBrBbX1dZo/edit?tab=t.0#heading=h.1ibeovu64qqa">EA Training</a></td><td>Training for Good (2021), which became <a href="https://www.tarbellcenter.org/fellowship">Tarbell Fellowship</a></td></tr><tr><td>People</td><td><a href="https://docs.google.com/document/d/10xpN5IgB7cEz0yjeJ7ZzxqQ8NtrthmuvnfFkOIWTGRc/edit?usp=drive_link">Earning to Give</a></td><td><a href="https://www.highimpactprofessionals.org/">High Impact Professionals</a> (2021)</td></tr><tr><td>Money</td><td><a href="https://docs.google.com/document/d/1ir6p54x_OKIAxX82ImuNKmxkA0mQ5gbCc11oW2XyZSY/edit?tab=t.0">Fundraising for Effective Animal Advocacy</a></td><td><a href="https://www.farmkind.giving/">FarmKind</a> (2023)</td></tr><tr><td>Money</td><td>Own Idea</td><td><a href="https://www.arkphilanthropy.org/">Ark Philanthropy</a> (2023)</td></tr><tr><td>Services</td><td>Own Idea (<a href="https://www.ambitiousimpact.com/post/presenting-2023-incubated-charities-round-1-charity-entrepreneurship">AIM site</a>)</td><td><a href="https://www.themissionmotor.org/">Mission Motor</a> (2023). </td></tr><tr><td>Money</td><td><a href="https://9475dbf4-555e-4808-9886-5f8ee815cc82.usrfiles.com/ugd/9475db_d06fa170d29349318fe61cd87f1d5faa.pdf">Effective Giving Incubation</a></td><td>Benefficienza, Ellis Impact, Impactful Giving, Mieux Donner (2024)</td></tr></tbody></table>

</details>

## General considerations

**Judgment calls play a larger role** than in other research topics: there's rarely evidence on "what works," and nearly all the evidence available will be far from RCT-grade. Determining need relies on conversations with actors and their subjective judgments.

* We should therefore: consult widely, and state the theoretical judgments an idea rests on.

**Attribution is a recurring problem.** How much of the "impact" can a meta charity claim? Multiple organizations often touch the same donor or career change, and counterfactuals are harder to establish here than elsewhere.

* We should therefore: be explicit about uncertainty and the inferences and assumptions being made; and focus on benefit-cost ratios (BCRs), BOTECs, and Monte Carlo-style analysis with reported sensitivity rather than point estimates.

**Ultimate impact flows through the impact of other organizations.** Final impact is inherited from the enabled orgs — a meta org's impact claim always chains through someone else's work, often with long lags.

* We should therefore: set falsifiable benchmarks on what meta orgs do observe (money moved, placements, paying customers); and be explicit that impact-level claims rest on modeled assumptions rather than observed feedback.&#x20;

**Many meta ideas lean explorative and show follow our** [**guidance for 'explore' ideas**](/explainers-and-house-views/our-charity-taxonomy/guidance-to-adjust-research-based-on-explore-exploit-spectrum.md)**.-**

## What criteria should we focus on?

Our evaluative criteria (see [here](https://docs.google.com/document/d/1_CbD5IzXKeExRRitwRWXvrmAlcv6GTkLLs8c9ilCTO4/edit?tab=t.lx24p4gfhcw4#bookmark=id.sf203yywl8gx)) describe how to investigate ideas. We should use similar evaluative criteria for meta-organizations, adapting them to this specific area in a few ways. Table 1 discusses some of the changes in focus we think are needed.&#x20;

&#x20;**Table 1: Alternative focus in meta-organization evaluations**&#x20;

<table><thead><tr><th width="155.671875">Criteria</th><th width="186.30078125">Traditional focus</th><th>Approach in “meta”</th></tr></thead><tbody><tr><td>Counterfactual impact</td><td>Cost-effectiveness above a particular bar, and potential for scale</td><td><ul><li>Focus on the benefit-cost ratio (BCR) instead of cost-effectiveness </li><li>Back-of-the-envelope estimation, instead of a more detailed analysis <a href="https://paperpile.com/c/UR4cOj/bQa3/?prefix=e.g.%2C">(e.g., Open Philanthropy GHW Staff, 2025)</a></li><li>Scale is most often measured in money moved</li></ul></td></tr><tr><td>Tractability or Potential for Success</td><td>Rigorous evidence, forecasted chances of success, and strong ToC</td><td><ul><li>Prioritize Theory of Change thinking. </li><li>Focus on crucial considerations for tractability (e.g., limiting factors) and implementation over rigorous evidence (in the absence of it).</li><li>Favor “many weak arguments” <a href="https://paperpile.com/c/UR4cOj/SG90">(Karnofsky, 2014)</a></li></ul></td></tr><tr><td>Neglectedness</td><td>Actors, funding, attention</td><td><ul><li>Determine need, especially by directly consulting/surveying potential service users <a href="https://paperpile.com/c/UR4cOj/GAPT/?prefix=e.g.%2C">(e.g., Laughlin, 2025a)</a>. </li></ul></td></tr><tr><td>Relevance </td><td>Risk, portfolio, etc. </td><td><ul><li>Understanding the tradeoffs between incubating an organization to provide services to the AIM community, relative to contracting external organizations.</li></ul></td></tr></tbody></table>

## What to front-load

* **Consultation with experts.** We suggest a mixture of survey consultation and in-depth interviews, an approach we followed in our report for a livelihoods non-profit evaluator [(Laughlin 2025a)](https://paperpile.com/c/UR4cOj/VQ8h/?noauthor=1). We think we should consult more people than in our usual reports because:
  * A meta organization's customers are other orgs, donors, or talent, so "would anyone actually use this?" is a crucial consideration.
  * These consultations will feed and shape the theory of change.
  * Tips: Focus on revealed interest rather than stated interest. One approach might be to interview like a customer-discovery call, not a pitch: ask what the org currently does, pays for, and struggles with — not "would you use X?", which invites polite yeses ([the "Mom Test" approach; Fitzpatrick, 2013](https://www.momtestbook.com/)). Also, if existing players saying they already plan to cover the gap, ask probing questions to assess if another actor is still needed - in our experience follow-through is rare ([Cox 2023, fn. 5](https://docs.google.com/document/d/1ir6p54x_OKIAxX82ImuNKmxkA0mQ5gbCc11oW2XyZSY/edit?tab=t.0)).
* **Map out a theory of change** and review evidence underpinning assumptions. Unlike our typical evaluations, mix evidence review into the ToC: evaluate evidence narrowly against the listed assumptions, not as a standalone review.
* **Crucial considerations.** Draw on the ToC and the consultations; evaluate them transparently.

**What to do Later**

* **BCR model and implementation details.** This is the part that changes most often. In implementation details, focus on how to run the organization and how plausible it is that it generates the outputs as intended.

**What might be less informative for meta-ideas**

* Burden or DALY-style needs assessment — need is established through consultation instead.
* Deep evidence review of the end-line intervention literature — that evidence belongs to the recipient orgs, and the meta org inherits it.
* Standard geographic prioritization — for donor-facing ideas geography follows the founders' network, or the work is online and global.

**Suggested structure**

* Exec summary (same as usual)
* Introduction
* A detailed theory of change, including evidence considered
* Crucial consideration analysis&#x20;
* Consultation findings not otherwise reported
* BCR model&#x20;
* Implementation details&#x20;

## Pointers for different mechanisms

This section is intended to break down the different mechanisms in this category, provide examples of prior AIM reports that fit into these mechanisms, and give remaining cruxes to investigate.

<details>

<summary><strong>Money ideas (Fundraising, capital allocation)</strong></summary>

**Topics to investigate and notes**

* Market saturation: how crowded is the target donor segment/geography?
* Counterfactual of the money moved: new money vs. reallocated giving vs. funged donations. Where does it come from; where would it have gone?
* Attribution across orgs touching the same donor (pledge org + evaluator + platform might all claim the same dollar). Consider what discounting should occur for this.

**Resources**

* **Past AIM Work:**&#x20;
  * **Effective giving initiatives (2024)**
    * Resources for reference: (Report by Baas, Speziali & Moore ([2023](https://9475dbf4-555e-4808-9886-5f8ee815cc82.usrfiles.com/ugd/9475db_d06fa170d29349318fe61cd87f1d5faa.pdf#page=4.08)), [EGI country short list WFM](https://docs.google.com/spreadsheets/d/1yN4iwtsXqPP4xZ4f-gJ-FiY6cvOami3i-PCCKNle18E/edit?gid=1396733280#gid=1396733280), example research notes drafts for [India](https://docs.google.com/document/d/1TV89b0hFykn0PsALXgtOHGrwXOB4y7f9Hn57NNBUC-k/edit?usp=drive_web\&ouid=113318567093078694329), GWWC impact evaluations ([2022](https://www.givingwhatwecan.org/impact/2020-2022-impact-evaluation), [2024](https://www.givingwhatwecan.org/impact/2023-2024-impact-evaluation)) (methodology sections) .&#x20;
    * Incubated charitie&#x73;**:** [Benefficienza](https://google.com/search?gs_lcrp=EgZjaHJvbWUyBggAEEUYOTIHCAEQABjvBTIHCAIQABjvBTIHCAMQABjvBTIHCAQQABjvBTIHCAUQABjvBTIGCAYQRRg70gEIMzY2MWowajSoAgCwAgA\&ie=UTF-8\&oq=benefficienza\&q=benefficienza\&rlz=1C5OZZY_enGB1214GB1214\&source=chrome.ob\&sourceid=chrome) (Italy), [Ellis Impact](https://www.ellisimpact.org/) (US), [Impactful Giving](https://www.impactfulgiving.in/) (India), Mieux Donner ([France](https://mieuxdonner.org/home/))
  * **Fundraising for Effective Animal Welfare** ([Report](https://docs.google.com/document/d/1ir6p54x_OKIAxX82ImuNKmxkA0mQ5gbCc11oW2XyZSY/edit?tab=t.0)), which became [**FarmKind**](https://www.farmkind.giving/)**.**
    * Note: No CEA — judged not decision-relevant; leaned on expert interviews and case studies of existing fundraising orgs, with a 'what do you need to believe' exercise instead
  * **Earning to Give (2021)** ([Report](https://docs.google.com/document/d/10xpN5IgB7cEz0yjeJ7ZzxqQ8NtrthmuvnfFkOIWTGRc/edit?tab=t.0))**,** which became High Impact Professionals (HIP).
    * Note: This report started as a money idea charity, while the organization (HIP) became a 'people' charity.
* **Guidance resources on giving multipliers**: GWWC impact evaluations ([2022](https://www.givingwhatwecan.org/impact/2020-2022-impact-evaluation), [2024](https://www.givingwhatwecan.org/impact/2023-2024-impact-evaluation)) (methodology sections); [Founders Pledge giving-multipliers report](https://www.founderspledge.com/research/giving-multipliers).

</details>

<details>

<summary><strong>People ideas (Talent, incubation, community)</strong></summary>

**Topics to investigate and notes**

* As with most meta-ideas, we focus more on the ToC, need, neglectedness, and tractability, and less on quantifying cost-effectiveness.
* **Value of the shift**: what would the person have done otherwise, and how much better is the new path (e.g., medical doctor → AIM researcher)? This is the counterfactual *delta*, not the gross impact of the new role.
* **Quantifying the good**:&#x20;
  * We focus on BCR.
  * We think precisely quantifying the good is very challenging and unresolved.
  * Sometimes the lag can be very long — a university group may take 5–10 years to reveal whether it changed anyone's long-term trajectory.
  * Attribution is also very hard and unresolved (e.g. how much of a career transition can be claimed by a person who took part of a university EA group, 80,000 hours, and high impact professionals. Still, these are plausible steps in the 'domino-effect', where each step might be needed for the person to enter a high impact position.

**Resources**

* **AIM Reports**
  * 2021 AIM report on 'EA training' ([Hilton 2021](https://docs.google.com/document/d/14m1wvrup7whaRNHfLcjbvxyL5XvVgas7KIBrBbX1dZo/edit?tab=t.0#heading=h.1ibeovu64qqa)). This report recommended a new org to identify a talent gap and address these with training and mentorship programs. This did not have a quantified BCR or CEA. This report led to incubating [Training for Good](https://www.charityentrepreneurship.com/training-for-good).&#x20;
  * **Other incubated meta-charities without reports:** Animal Advocacy Careers.

**Resources on quantifying impact**

* **AIM Incubated Example - High Impact Professionals 2021:**
  * **The 2025 and 2024 Annual Reports (**[**HIP - our impact**](https://www.highimpactprofessionals.org/our-impact)**)** provide a reference for how an AIM incubated charity have evaluated their own impact. In 2025, they estimated 46 "career transitions" which accounted for 90% of their impact, and a total counterfactual impact of 2.2 Million USD. They estimate a 16.2x impact multiplier.&#x20;
* **External Example - 80,000 Hours:**
  * 80K's history on units to measure career changes which we think should provide caution about being overly optimistic about measuring career changes: they moved from impact-adjusted significant plan changes (IASPCs) ([2015](https://80000hours.org/2016/07/update-on-number-of-significant-plan-changes/#impact-adjustment-of-significant-plan-changes), [2018](https://80000hours.org/2019/05/annual-review-dec-2018/)) to full career-value units ([DIPYs - 2019](https://docs.google.com/document/d/1PFSKzGV_fVcBJgV77dtiNTi9d1kFE1_skODtPXXqGec/edit?tab=t.0)) and to simple verifiable counts called criteria-based plan changes ([CBPCs - 2020](https://docs.google.com/document/d/1rWfQ3Lja2kYoUm_t9uNqBgEn5nz6KL8fmNP5db8cZRU/edit?tab=t.0)).&#x20;
  * They speak about these changes in their mistakes page ([here](https://80000hours.org/about/credibility/reviews/mistakes/#we-were-overly-credulous-about-how-easy-it-is-to-cause-career-changes-and-our-investigations-into-these-claims-were-insufficiently-skeptical-and-thorough)), where they say with regards to IASPCs that "We now think that on average we were too optimistic about how easy it is to cause a ‘trajectory change’... since the EA community often helps people shift careers separately from our work"

</details>

<details>

<summary><strong>'Ecosystem services ideas (ops, fiscal sponsorship, platforms, meta research)</strong></summary>

**Topics to investigate and notes**

* Focus on stakeholder mapping and expert interviews with 'superconnectors' or potential recipients organizations of services to address this.
* Utilization: will enough orgs actually adopt the platform to justify fixed costs? What's the realistic addressable ecosystem?&#x20;
* Counterfactual providers: for infrastructure ideas commercial alternatives often exist (payroll, legal, software) — the case must rest on ecosystem-specific value, not generic services.
* Single-point-of-failure and lock-in risk: concentrating ecosystem functions in one provider creates fragility (what happens if it shuts down?).
* Sustainability and funding model: standing platforms need standing funding — can it move to fees, or is it a permanent philanthropic cost?
* For meta-research ideas where the goal is to steer the ecosystems efforts, we expect a greater degree of uncertainty.&#x20;
  * More emphasis on scoping out various theories of changes, their underlying assumptions of the theory of change, and expert views.

**Resources**

* **Past AIM work - Exploratory Altruism (**[2021](https://3394c0c6-1f1a-4f86-a2db-df07ca1e24b2.filesusr.com/ugd/9475db_a215cf49667a40f7aca07db9c10bd246.pdf)**)**
  * We recommended a charity focused on Exploratory Altruism ([2021](https://3394c0c6-1f1a-4f86-a2db-df07ca1e24b2.filesusr.com/ugd/9475db_a215cf49667a40f7aca07db9c10bd246.pdf)), focused on finding new highly impactful causes. This could influence time, funds, and other resources. This lead to incubating [CEARCH](https://exploratory-altruism.org/).&#x20;
    * *Exploratory Altruism caveat: The CEA was quick (2–10 hours) and, per the authors, only comparable to other CE meta CEAs. The recommendation rested mainly on neglectedness, community demand, and expert views; the CEA's role was to show the numbers clear the bar under deliberately conservative counterfactuals. A sensitivity analysis or Monte Carlo on the impact multiplier would strengthen it.*
  * **Other incubated meta-charities without reports:** [Giving Green](https://www.givinggreen.earth/), Animal Ask, and Mission Motor.
* Outside AIM resources examples:&#x20;
  * [GWWC "Evaluating the evaluators" series](https://forum.effectivealtruism.org/posts/pp2jmWHyDK9sfC4Rh/evaluating-the-evaluators-gwwc-s-research-direction) (for evaluator-type ideas).

</details>

## Quantifying impacts&#x20;

Carry out BCR analysis by quantifying ultimate outcomes in monetary terms. In some cases, outcomes will be straightforward, like money saved. In others, quantifying outcomes in financial terms will not be clear. For instance, some models will require us to value improvements in service delivery or higher-quality staff.&#x20;

In cases where there is significant estimation uncertainty, we suggest the following approaches:&#x20;

* Guesstimate values through different approaches and report different results based on different assumptions, as well as an average result based on all estimations
* Task independent researchers to come up with their estimates individually&#x20;
* Conduct willingness-to-pay assessments by
  * Directly asking stakeholders how much they currently pay for similar services and how much would they be willing to pay&#x20;
  * Researching how much existing service providers cost&#x20;
* Value the counterfactual delta, not the gross flow. Shifted money counts only for the effectiveness difference between where it lands and where it would have gone; a career shift counts only for the difference between the new and old path, not the full value of the new role.

Generally, we expect modeling to be rougher for these evaluations. The following advice applies to all our evaluations, but is of particular importance in these:&#x20;

* Be transparent about all modeling choices&#x20;
* Carry out and report sensitivity analysis. Suggest how the BCR outcomes would vary depending on how you adjust your assumptions and estimations.&#x20;

## What should be our “bar”&#x20;

We do not expect to incubate meta-organizations with a BCR lower than 5:1 at scale. We are unsure what our bar should be, and we expect this to change over time as we refine our approach.&#x20;

We think the counterfactuals of funding are very important to the ultimate BCR bar and will matter greatly in decision-making. All else equal, if a service-providing organization can be funded by the services it provides instead of philanthropic dollars and, in doing so, deliver highly beneficial services to effective charities, then the argument for recommendation is stronger (and may justify recommending something with a marginal BCR).&#x20;

Our most recent historical example of a BCR bar used at AIM for decision-making purposes relates to effective giving organizations in Baas et al. [(2023)](https://paperpile.com/c/UR4cOj/Cgo0/?noauthor=1). There, we suggest that the impact multipliers—essentially a BCR—achieved by several organizations provide good reason to incubate more effective giving organizations. These include:&#x20;

* Effektiv Spenden (Germany): 13&#x20;
* Giving What We Can (Global, registered in the UK and USA): 30&#x20;
* RC Forward (Canada): 4–10&#x20;

In earlier research rounds, AIM recommended ideas with the following BCRs:

* Exploratory Altruism: \~8 [(Hilton & Agarwalla, 2021)](https://paperpile.com/c/UR4cOj/etXo)
* Earning to Give: \~6 [(Savoie & Agarwalla, 2021)](https://paperpile.com/c/UR4cOj/abyr)

We did not comprehensively examine BCR thresholds in other fields. However, our rough search found that the most cost-effective programs in the development field (related to the SDGs) can reach BCRs between 20 and 125 [(Lomborg, 2023)](https://paperpile.com/c/UR4cOj/UYY0).

## References

[Baas, J., Speziali, F., & Moore, L. (2023). Effective Giving Incubation. Ambitious Impact; Giving What We Can. ](http://paperpile.com/b/UR4cOj/Cgo0)<https://9475dbf4-555e-4808-9886-5f8ee815cc82.usrfiles.com/ugd/9475db_d06fa170d29349318fe61cd87f1d5faa.pdf>

[Hilton, S., & Agarwalla, V. (2021). Exploratory Altruism. Ambitious Impact. ](http://paperpile.com/b/UR4cOj/etXo)<https://3394c0c6-1f1a-4f86-a2db-df07ca1e24b2.filesusr.com/ugd/9475db_a215cf49667a40f7aca07db9c10bd246.pdf>

[Karnofsky, H. (2014, June 10). Sequence thinking vs. cluster thinking. The GiveWell Blog. ](http://paperpile.com/b/UR4cOj/SG90)<https://blog.givewell.org/2014/06/10/sequence-thinking-vs-cluster-thinking/>

[Laughlin, M. (2025). Livelihoods-focused Non-profit Evaluator. Ambitious Impact. https://doi.org/](http://paperpile.com/b/UR4cOj/GAPT)[10.5281/ZENODO.16411832](http://dx.doi.org/10.5281/ZENODO.16411832)

[Lomborg, B. (2023). Save 4.2 million lives and generate 1.1 trillion in economic benefits for only 41 billion: Introduction to the Special Issue on the most efficient policies for the sustainable development goals. Journal of Benefit-Cost Analysis, 14(S1), 1–15. https://doi.org/](http://paperpile.com/b/UR4cOj/UYY0)[10.1017/bca.2023.32](http://dx.doi.org/10.1017/bca.2023.32)

[Open Philanthropy GHW Staff. (2025, May 28). How we use back-of-the-envelope calculations in our grantmaking. Open Philanthropy. ](http://paperpile.com/b/UR4cOj/bQa3)<https://www.openphilanthropy.org/research/how-we-use-back-of-the-envelope-calculations-in-our-grantmaking/>

[Savoie, J., & Agarwalla, V. (2021). Earning to Give +. Ambitious Impact. ](http://paperpile.com/b/UR4cOj/abyr)<https://3394c0c6-1f1a-4f86-a2db-df07ca1e24b2.filesusr.com/ugd/9475db_5bb71c0641bf445c8a51be0a2ea1df8e.pdf><br>


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://research.ambitiousimpact.com/explainers-and-house-views/our-charity-taxonomy/guidance-to-adjust-research-for-meta-interventions.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.