Thank you for this write-up! It sounds really intriguing. The calculation is a bit beyond my area of expertise, but here are some more general reflections:
If this is done through a fully vegan meal-kit service, it might be useful to also offer it as a long-term paid subscription option. For example, offering the first month for free and subsequent months at low/self-cost.
This may increase retention by making it more convenient for people to keep on using the meal kit service, automatically getting vegan meals afterward.
This would render the first month a form of philanthropy-funded advertisement for the meal kit service, giving it a great market edge over non-vegan competitors.
This might also enable more scale, pushing down per-unit costs.
A lot of people, including myself, are very averse to paid subscriptions. But I think that a full month's food budget is a big enough bait to overcome this.
I would be interested in gifting this type of free initial month to friends and family. Sometimes I've given away vegan cookbooks in the hope that this would make people more comfortable making vegan meals, both for themselves and for me when visiting, and this could have a similar function.
Importantly, to me, this would not trade against my donation budget, but against my gift budget, reducing the opportunity cost massively.
If this is given as a Christmas gift for someone's January food budget, there would be an obvious opportunity to pair up with Veganuary.
There may be a good opportunity to bundle meal kits with information campaigns, e.g., using the back side of the recipe to mention some fun (or not so fun) fact of the day, thereby increasing motivation.
Since people will already be prompted to think about their food choices, it might give them more food for thought (sorry). Anecdotally, for me personally, I remember taking a much bigger interest in these questions just after transitioning to a vegan diet, likely due to decreasing cognitive dissonance, making me much more receptive to the information.
Moreover, since the service is already gathering data on retention, this also creates a great opportunity for RCTs to test which form of information is most useful.
For interpretation of the results, I wonder if you could also mention what ranges of SADs/$ other promising interventions have. Is 19 a lot or a little, compared to our best options?
Thank you again for your work on this, and best of luck with it! I hope to be able to gift a subscription soon!
Somehow I didn't even realise there were fully-vegan services, thanks for pointing it out! There's definitely some good benefits to it, slight downside is that my initial scan puts them as ~1.5x base cost of Gousto, so there's a tradeoff there. I will consider this more.
The gift option might be cool even independently of this sort of trial, esp. with Christmas gift for Veganuary, as you mention.
Very good idea on the info campaign, and the further study. Would definitely require a closer collaboration with the kit service, for them to monitor which boxes are for which trial participants, this might be another point in favour of choosing a fully-vegan service.
For the final point, I've added comparisons to the 'Cost-Effectiveness Estimates' section. The midpoint of 19 SAD/$ is below these estimates, but the optimistic case of 49 SAD/$ is comparable with some of them.
I notice that your assessments make the intervention on par with "Securing Scale-up Funding for Alternative Proteins", which passed the bar for being recommended for AIM, which is encouraging. Given the uncertainty in the estimates, there seems to be significant Information Value from trying this at a smaller scale, and seeing what the data says.
At ~£4-6K it doesn't seem impossible to find a small-grant funder supporting it. Some ideas for alternative or additional funding sources include:
If you partner with an existing meal-kit service, they may be interested in co-financing the experiment, since it, as previously mentioned, basically provides them with free advertisements.
You could also ask people to refer participants in exchange for some co-payment, e.g. 25%. This would enable people to get their friends a month of food at a 75% discount, while also filtering for people who are a bit more invested and likely to be aligned.
Again, Veganuary might be helpful in finding funding sources or in marketing the project to find suitable referrals.
Faunalytics may be interested in helping with study design and implementation, and potentially helping find funding sources.
Since AIM recommended comparable interventions, they may be interested in helping or advising on this project.
Just getting in contact with these various organizations to bounce the idea would also give you an opportunity for additional feedback!
Another good point, I've added the below to the 'Cost-Effectiveness Estimates' section:
For comparison, the top potential interventions in AIM's 2025 and 2026 reports, are estimated at cost-effectivenesses of 153, 3-51, ~36-104, and ~30 SADs/$, respectively.
I was just wondering whether you took existing academic literature into account on this topic. I am aware of at least 2 related studies. One looked at the effect of vegan meal boxes and found no effect on meat consumption.Yet they only delivered one meal box with 3 meals. Another one provided just plant-based meat alternatives for a month and did find a significant drop in meat consumption at least a month after intervention end. Based on this evidence I would be more optimistic about providing free alternatives only because of it’s effectiveness identified in this one study and because this would be cheaper than providing full meals. But the evidence is not conclusive at all. But might bring additional valuable input
Lotte, massive thanks for finding the articles! The month-long intervention in particular seems very similar to this idea, so it's very useful to know about :-)
This is an excellent idea. People respond to free, and there will be some non-zero percentage of people for whom the habit of eating plant-based will stick, and possibly fully embrace the ethics of veganism.
Some feedback:
1) free food for a month is expensive and may attract people simply looking for free food 2) it may be somewhat effective or as effective to offer a discount for a month (e.g. 25% or 50%) 3) any effort should have some kind of commitment to abstain from animal food for the duration of the month 4) any effort should have some video tutorials to help people know what their options are, and how to prepare simple meals 5) any effort should also include dining at local vegan restaurants 6) any effort should also include interaction with other vegans, both for social support as well as to help with the transition
Next steps: definitely DONT drop it! Do a pilot. Start with one household near you and self fund it if you can. Put a flyer up, and frame it as primarily a health or environmental thing. Stay in close contact with the household electronically, by phone, and in person. You'll learn so much about how it could play out and it will help you scale. If the household does not want to see the month through, go to the next household which has responded to your flyer.
Hi! I find your idea interesting, but I have one concern: how would you prevent people from abusing this system? For example, some people who are already vegan could just subscribe for some free food. You might think that ethical vegans would not commit this kind of misendeavor, but not all vegans are vegans for health reasons. I can also see that anti-vegan people could subscribe just to make the food go to waste.
Good point - unfortunately I don't have a good answer! At small scales I think getting most/all participants via referrals can limit this effect, but at large scales I'm unsure.
TL;DR: I've analysed temporarily subsidising plant-based meal kits as an animal welfare intervention. My preliminary estimate is that this averts 19 SADs/$ (90% CI: 5-49 SADs/$). I'm now looking for feedback and criticism, and suggestions on if/how to take this forward. Full doc is here. Cost-Effectiveness calculation is here.
Meta notes: The doc/spreadsheet/post took ~15hrs of effort. I wrote the doc and did the calculation. Claude did red-teaming and wrote ~all of this post (using the doc).
The Core Idea
If people experience a month of convenient, tasty plant-based dinners, some will continue reducing meat consumption afterward. We could pay for people to receive plant-based meal kits for ~1 month. This model would remove a lot of the initial friction associated with diet change. Then, we track whether this reduces their long-term meat consumption or catalyses dietary change.
Social spillover effects (harder to measure): Participants share positive experiences → Shift in social norms around plant-based eating → Broader dietary change
Market effects (speculative): Increased demand for quality vegan meal kits → Companies invest in better products → Plant-based options become more accessible and normalised
Practical Implementation
The intervention
Provide participants with ~30 days of plant-based meal kits for free (negotiating bulk discounts from retailers).
Two variants to test:
One meal: Participants replace only their main meal (easier compliance, more likely completion)
Fully vegan: Participants are expected to eat a fully plant-based diet (potentially more transformative, harder to sustain)
Provide some post-intervention support: meal planning apps (eg. Cherrypick), or a grocery stipend during transition
Track outcomes through surveys/interviews at exit, 6 months, and beyond.
For comparison, the top potential interventions in AIM's 2025 and 2026 reports, are estimated at cost-effectivenesses of 153, 3-51, ~36-104, and ~30 SADs/$, respectively.
Key Uncertainties
Intervention length (shorter is cheaper but may reduce long-term impact)
Discount negotiations with meal kit companies
Long-term dietary change rates (currently assumed based on similar studies from Faunalytics)
Whether "meat disgust" develops and matters for persistence
Spillover effect magnitude
Model Simplifications
Excludes overhead (staff time, coordination, survey costs, dropouts)
Treats all meals as equal in SADs (likely wrong—dinner probably matters most)
Uses Vasco Grilo's SADs/person-year estimates
Critical Assumptions
Core Assumption
One month of eating plant-based dinners causes sustained reduction in meat consumption.
This isn't immediately clear. Habit formation alone may not be sufficient without accompanying belief/value shifts.
Practical Assumptions
Meal kits are high-quality and easy to use (so people actually complete the trial)
Can negotiate extended discounts with retailers (40-50% off is standard for first 2 weeks, aim to match that for a month)
Can effectively account for social desirability bias in surveys
Target demographic won't be saturated quickly
Why This Might Be Better Than the Numbers Suggest
Quality control: Unlike asking people to "go vegan," this provides instructions and ingredients, reducing risk of nutritionally inadequate or unpalatable attempts that create negative associations.
Higher compliance: Free food is a strong incentive! People are much more likely to actually try this vs. a purely motivation-based intervention.
Measurable: Can confirm delivery and track consumption more reliably than self-directed diet change.
Potential for higher conversion: If spillover effects are real and substantial, impact could be much larger than individual diet change alone.
Why This Might Be Worse Than the Numbers Suggest
Scaling challenges: May only work cost-effectively for specific demographics. At scale, people might claim interest just for free food.
Measurement problems: Surveys suffer from desirability bias. Actual long-term behaviour change could be much lower than reported.
Rebound effects: Once meal kits stop, friction returns. People might revert entirely or even overcorrect (eating more meat after feeling deprived).
Negative spillover: Bad experiences could make participants less likely to try plant-based eating in the future.
Please Give Feedback!
I'd really appreciate feedback from anyone on this. Here's some questions that I think anyone could weigh in on:
Is this an appealing intervention? Would you participate?
You can copy the model and try your own numbers if you want!
Am I missing anything really important in the model?
Am I missing any good improvements in the intervention design?
You can leave a comment below, or comment directly on the doc or spreadsheet.
Potential Next Steps
I'm pretty uncertain what (if anything) to do next. I'd appreciate thoughts, especially from those with relevant background. Some options I've considered:
Do more investigation.
Theoretical, eg. refining the plan.
Practical, eg. contacting meal-kit companies.
Try and run a pilot, and apply for funding to do so.
I expect a decent pilot needs n=20-30 and would cost ~£4-6K.
I don't know who (if anyone) would fund this.
Drop it, and focus on something different and more effective.
This is a crosspost from the new Animal Welfare Alignment Newsletter by Anima International. You can subscribe on Substack if you are interested in following these efforts. Audio reading also available on Substack.
The goals of this post are to:
1. Raise a question I see as crucially important to the goal of aligning AI to animal welfare...
Hello! I'm Justin Portela. I got hired by GWWC to make YouTube videos after AI in Context did such a kickass job.
My channel is using that same cinematic, high-production value beauty to talk about everything in the EA universe that isn't AI.
...
“How long have you been v*g*n?”
This is one of the most common icebreakers at animal protection events. It’s a baseline assumption, and it mostly holds true: if you’re out advocating for animals not to be tortured or abused, realistically these days you are v**n, or close. And it makes for good conversation. It seems fairly safe to assume when you meet strangers.
But this assumption is hurting the movement in a way which we don’t always notice: someone new comes into the sp...
Thank you for this write-up! It sounds really intriguing. The calculation is a bit beyond my area of expertise, but here are some more general reflections:
Thank you again for your work on this, and best of luck with it! I hope to be able to gift a subscription soon!
Thanks for engaging!
Somehow I didn't even realise there were fully-vegan services, thanks for pointing it out! There's definitely some good benefits to it, slight downside is that my initial scan puts them as ~1.5x base cost of Gousto, so there's a tradeoff there. I will consider this more.
The gift option might be cool even independently of this sort of trial, esp. with Christmas gift for Veganuary, as you mention.
Very good idea on the info campaign, and the further study. Would definitely require a closer collaboration with the kit service, for them to monitor which boxes are for which trial participants, this might be another point in favour of choosing a fully-vegan service.
For the final point, I've added comparisons to the 'Cost-Effectiveness Estimates' section. The midpoint of 19 SAD/$ is below these estimates, but the optimistic case of 49 SAD/$ is comparable with some of them.
Thanks, the comparisons are very helpful.
I notice that your assessments make the intervention on par with "Securing Scale-up Funding for Alternative Proteins", which passed the bar for being recommended for AIM, which is encouraging. Given the uncertainty in the estimates, there seems to be significant Information Value from trying this at a smaller scale, and seeing what the data says.
At ~£4-6K it doesn't seem impossible to find a small-grant funder supporting it. Some ideas for alternative or additional funding sources include:
Just getting in contact with these various organizations to bounce the idea would also give you an opportunity for additional feedback!
Best of luck!