Animal advocacy organizations use online advertisements, sometimes referred to as veg ads or just online ads, to inspire viewers to adopt more animal-friendly behaviors and attitudes. We will focus on ads that promote veg eating in this intervention report while considering other secondary outcomes, primarily raising support for an organization since this can easily be combined with a veg eating message. For example, organizations can include a Facebook Like button on the landing page.¹ We focus on the type of online ads run by Mercy For Animals and The Humane League.

We believe that online ads are probably successful, to some extent, in changing individual consumption and increasing social media presence. Even if the per-person consumption change is small, the low cost and easy testability² of online ads could make them highly cost-effective.

We think they have relatively weak secondary effects, such as building the animal advocacy movement and increasing the likelihood of a major social shift away from factory farming. Additionally, by framing veg eating as a personal choice, online ads might dampen the strength of animal-friendly social norms in society. For example, online ads might reduce the view of veg eating as a moral imperative. Overall, we think this potential indirect harm is outweighed by the benefit of the general promotion of animal-friendly ideas.

What Are Online Ads?

We use the term “online ads” to refer to any form of online advertising with an animal protection message. Usually, an organization or individual pays to have an image and/or text displayed to users who are not visiting the site specifically to view pro-animal materials, and users who click on the image and/or text are directed to a new website that provides persuasive materials encouraging the user to adopt more animal-friendly behaviors and attitudes. We focus on ads that are primarily aimed at encouraging users to reduce or eliminate their consumption of animal products. The ads usually take one of two approaches to achieve this outcome:

• Bring viewers to a page showing a video describing the cruel treatment of animals raised for food, often with a personal story encouraging users to take action against this cruelty by changing their diet.³
• Bring viewers to a page with a text description of the benefits of veg eating, as well as the large number of people participating, and highlights the offer for a free Vegetarian Starter Guide (or other materials).⁴

The latter approach (VSG ads) has been used only relatively recently, but due to its high efficiency of getting viewers to pledge to go vegetarian, it is likely to be used more in the future.⁵

The landing page can include a short sign-up form, where users can pledge to go vegetarian and/or receive more information via email. The emails the user receives usually cover topics such as (i) how to transition smoothly to a vegetarian diet, (ii) how to order meatless meals while eating out, and (iii) popular alternative meats, such as veggie burgers. The user is invited to reach out to the advertising organization with questions or comments.⁶

The landing page can have a social media widget for users to like or follow the advertising organization, which can lead to more engagement in the future.⁷

What Are Their Strengths?

Online ads reach a large number of people for a relatively low per-person cost. 1000 impressions (the ad is shown to users 1000 times) can be purchased for as low as $0.19 on a regular basis.⁸ There is also data showing that a significant number (roughly 0.3% to 3.8%⁹ ) of these users click on the ads, and a significant number (roughly 2.6% to 41.4%¹⁰ ) of those users then enter their email address and pledge to go vegetarian. Even if we think a small fraction of users actually change their diets because of the ads, the low costs could still make it a highly cost-effective way of changing individual diets. The staff costs are also low, given that one capable ads manager can run a very large number of ads.¹¹

Online ads also have relatively short feedback loops, meaning Program Managers can quickly see results of an ads campaign—measured in cost-per-click or cost-per-conversion¹² —and adjust the campaign to optimize these outcomes.¹³ For example, they can try using different images and text in the ad, or target different demographics. This can be done over a period or weeks or months with very large sample sizes.¹⁴ Taking this approach of experimentation and optimization seems essential to running a highly cost-effective ads program.¹⁵

By emailing users who took the pledge, and having them follow the social media presence of the advertising organization, online ads could lead to increased activism, at least in small ways like sharing animal advocacy content or signing petitions.

It’s not clear which factors make online ads more cost-effective, in terms of cost-per-conversion, because it varies so much based on demographics and when the ads are ran. One relatively clear trend is that the most receptive demographic is often young women.^{16, 17} Also, ads run in Spanish-speaking countries in Central and South America tend to have low cost-per-conversion.¹⁸

Another important finding has been the high cost-effectiveness of VSG ads, described above.¹⁹ It is possible that VSG ads could have lower rates of dietary change per conversion, which would make them less exciting, but we don’t see any compelling reasons to think that.

What Are Their Weaknesses?

We think online ads have relatively weak secondary effects, such as building the animal advocacy movement and increasing the likelihood of a major social shift away from factory farming. By framing veg eating as a personal choice, online ads might dampen the strength of animal-friendly social norms in society,²⁰ but we think this potential indirect harm is outweighed by the benefit of the general promotion of animal-friendly ideas.

It is also unclear what effect online ads have on their primary outcome, dietary change.²¹ The argument of, “Even if we think a small fraction of users actually change their diets because of the ads, the low costs could still make it a highly cost-effective way of changing individual diets,” isn’t particularly strong. We would favor an intervention that has a similar calculated cost-effectiveness based on a simplified quantitative perspective, but higher per-person costs and higher per-person impact.²² If you think social advocacy in general, especially that which is based in a brief interaction like an ad view, is generally ineffective, such that you would need robust evidence of a positive impact to see it as worthwhile, then the lack of robust evidence of the impact of online ads is a major downside.

What about Long Term Effects?

Possible Positive Effects

As mentioned in the previous section, we think the positive long term outcomes of online ads are likely quite limited. It is possible that simply by reaching a large number of individuals in the general public, ads can create a meaningful shift in the overall public attitudes towards animals. This shift could increase the chances of success for other advocacy efforts and policy change in the future. Online ads and similar forms of outreach could have contributed to the current perception in the U.S. that farm animals are frequently mistreated for commercial gain.^{23, 24} We would guess that other forces, such as undercover investigations, have been a stronger driver in the opposition to factory farming due to media coverage and a more society-focused message, but it is very difficult to tease out causation from observation of social trends.²⁵

Possible Negative Effects

As with possible positive long term effects, we have much uncertainty about the possible negative long term effects.

If online ads are successful in changing people’s attitudes, it is possible that the set of attitude changes they create could involve some negative attitudes. For example, if the ad frames the problems of animal agriculture as having an individual solution—going vegetarian—that could also increase people’s belief that society-wide action is inappropriate. This can be framed as seeing veg eating as a personal choice or as a social imperative. We speculate that online ads have a net positive impact on people’s views of veg eating as a social imperative, but note the opposite outcome as a possible negative long term effect.

In addition to a focus on individual dietary change, online ads tend to focus on vegetarianism or a reduction in consumption, rather than complete veganism. Some contest that this approach dilutes the message of animal advocates and makes it more difficult in the long run to convince people of the seriousness of animal suffering.²⁶ We also recognize this concern, but haven’t seen convincing evidence that it outweighs the potential benefits of reducetarian/vegetarian messaging, such as a lower barrier to entry for the animal advocacy movement that could increase total support and momentum.

Conclusions

Overall we think ads seem to have limited long-term and indirect benefits, which is likely where a majority of the potential impact lies in animal advocacy. We think the short, impersonal interactions of online ads programs have less of an effect on, e.g. creating new activists, than other forms of outreach. Also concerning are the potential negative impacts of online ads, at least in their current form, such as shifting attitudes towards veg eating as a personal choice rather than a social imperative.

Do We Recommend It? Why or Why Not?

We currently don’t recommend that organizations create new online ads programs or expand existing programs, at least when that funding could be used for more promising interventions such as corporate outreach and undercover investigations. We feel that the main upside of online ads, reaching a large number of people at very low costs, is outweighed by the concerns of low and uncertain per-viewer impact and the limited long term and indirect benefits. Even if we focus entirely on short-term impact, corporate outreach still seems more promising. However, if corporate outreach becomes more difficult in the future, e.g. once cage-free reforms are fully implemented, then online ads or other forms of dietary change outreach might be justified as the most cost-effective intervention for short term impact. We think that some ads programs, such as those that run VSG ads or have spent significant time minimizing their cost-per-pledge, tend to be more promising than others.

What Are Characteristics of a Strong Online Ads Campaign?

• Because the cost-effectiveness of online ads measured in cost-per-conversion varies greatly and unpredictably,²⁷ it seems that adaptiveness and experimentation are key in a successful online ads campaign.
• Some online ads have pro-veg messages in the ads themselves, instead of a message solely focused on creating clicks such as “Check out this crazy video!” We think the benefit of pro-veg messages, having a potential impact on all viewers instead of just those that click on the ads, is probably worth any potential cost to click rates.^{28, 29}
• Online ads managers tend to find that the most receptive demographic is young women,³⁰ and that ads run in Spanish-speaking countries in Central and South America tend to have the lowest cost-per-conversion.^{31, 32} However, it’s still important not to count on these trends, and for ads managers to continuously test different demographics.
• We also agree with ads managers that the recent use of VSG ads described above is likely a substantial improvement in cost-effectiveness. These ads show much lower cost-per-conversion, which likely corresponds to a similar increase in veg eating per dollar of funding, and we encourage more groups to adopt a similar strategy in their online ads programs.
• We think a large online ads program can save substantially on staff costs because most of the labor involved in ads, such as testing different campaigns and designing materials, is required whether running a small or large number of ads.

How Strong Is the Evidence about the Efficiency of Online Ads?

Online ads managers are able to directly track the rates at which people click on the ads and subsequently pledge to go vegetarian and/or enter their email address for more information.³³ We have little reason to doubt the validity of these measures.

There is limited evidence that these intermediate outcomes lead to reductions in the consumption of animal products. There are anecdotes, such as from people who email advertising organizations thanking them for running the ads and saying how they changed their diets.³⁴

We think the evidence is quite clear that, on average, dietary change leads to a reduction in suffering for farm animals. Although we cannot directly observe the process, we see no reason to believe that animal products are an exception to the basic economics of supply and demand.

Resources

Conversation with Jose Valle
Conversation with Alan Darer
Conversation with Cat Liguori and Andrea Gunn
Conversation with Josh Kalla

---

1. For example, Mercy For Animals includes a Facebook like button on this landing page. 

2. By “easy testability,” we mean that it’s relatively easy for organizations running online ads to test different images, text, etc. and optimize their approach for intermediate outcomes, such as the number of ad-viewers who pledge to go vegetarian. 

3. Examples: meatvideo.com, carnevideo.com, whosagainstanimalcruelty.org, descubrirlacomida.com 

4. Example: chooseveg.com/vsg-landing 

5. “One of these new approaches tried in 2015 was creating a landing page that allowed people to pledge to go vegetarian and download the [Vegetarian Starter Guide], but did not have the factory farming video. This also allowed us to make the text and picture of the Facebook ads reflect the offer of the free VSG instead of the video. Running this type of ad led to a massive drop in cost per conversion (initially to about ⅕ of the previous cost) and MFA’s cost per conversion has continued to decrease over time.” – Conversation with Alan Darer 

6. As part of this investigation, an ACE Researcher signed up for these emails (from each of the three organizations) to see what they received. 

7. For example, Mercy For Animals includes a Facebook like button on this landing page. 

8. “CPM (cost per 1,000 impressions) $0.19” (Spanish THL Ads (Dec-Feb 2016) ) – Conversation with Cat Liguori and Andrea Gunn 

9. These are the lowest and highest figures found based on calculations from the cost per 1,000 impressions and cost per click figures in these conversations: Conversation with Jose Valle, Conversation with Alan Darer, Conversation with Cat Liguori and Andrea Gunn 

10. These are the lowest and highest figures found based on calculations from the cost per 1,000 impressions and cost per click figures in these conversations: Conversation with Jose Valle, Conversation with Alan Darer, Conversation with Cat Liguori and Andrea Gunn 

11. For example, Mercy For Animals will spend about $550,000 in online ads in 2016, which will be managed by one ads manager working on it for 15-20 hours per week with some minor commitments from other staff.

    “Alan usually spends about 10-20 hours per week working on ads in total, and another MFA staffer, Krystal Caldwell, regularly tests different landing pages in order to maximize the percentage of viewers who pledge.

    …

    In 2016, they’ll spend more — approximately $550,000 — because MFA’s budget is increasing overall and online ads still appear to be very cost-effective.” – Conversation with Alan Darer 

12. A “conversion” in this context is usually a user who enters their email address and pledges to go vegetarian. 

13. “It’s important to have a mindset of constantly experimenting and improving the ads.” – Conversation with Cat Liguori and Andrea Gunn

    “An advertiser’s persistence at analyzing the changing relationship between these factors and successful ads, and constantly testing new approaches and ideas, directly relates to success in sales (or veg pledges in the case of animal advocacy).” – Conversation with Alan Darer 

14. For example, “They looked into targeting supporters of other causes, such as cancer research, especially after new findings were announced linking meat consumption and cancer. However, the results of these refinements did not justify paying the additional fee imposed by Facebook to narrow the target audience.” – Conversation with Jose Valle 

15. For example, “One of these new approaches tried in 2015 was creating a landing page that allowed people to pledge to go vegetarian and download the [Vegetarian Starter Guide], but did not have the factory farming video. This also allowed us to make the text and picture of the Facebook ads reflect the offer of the free VSG instead of the video. Running this type of ad led to a massive drop in cost per conversion (initially to about ⅕ of the previous cost) and MFA’s cost per conversion has continued to decrease over time.” – Conversation with Alan Darer 

16. “The ads have targeted young women between 24 and 34 years old who have animals or pets listed among their interests.” – Conversation with Jose Valle 

17. After arguing that the demographics with the most vegetarians are also those in which it is easiest to convince additional people to become vegetarian, Nick Cooney reviews the evidence about which demographics those are. “Female vegetarians also outnumber their male counterparts at a roughly even amount in all age groups. Since young people are more likely to not eat meat than older people, young women should be more likely to be vegetarian than any other age and gender group. Indeed, studies have consistently found that to be the case.” Cooney, N. (2013). Who’s ditching meat?: Gender. Veganomics. 

18. “MFA focuses their ads on Latin America because of the high cost-effectiveness, including cheap ad costs and high conversion rates, i.e. the proportion of people who pledge to go vegetarian and sign up with their email address after viewing an ad.

    …

    Cost per click in Latin America is just 1-4 cents, whereas in America it can range between 15-25 cents.” – Conversation with Alan Darer 

19. “One of these new approaches tried in 2015 was creating a landing page that allowed people to pledge to go vegetarian and download the [Vegetarian Starter Guide], but did not have the factory farming video. This also allowed us to make the text and picture of the Facebook ads reflect the offer of the free VSG instead of the video. Running this type of ad led to a massive drop in cost per conversion (initially to about ⅕ of the previous cost) and MFA’s cost per conversion has continued to decrease over time.” – Conversation with Alan Darer 

20. If you’ve ever argued with someone who eats or kills animals, you must have heard, “it’s my personal choice to eat meat!”. Why then, are we making this argument for them? By framing animal rights as an issue of dietary change – titling animal rights leaflets “Your Choice”, for example – rather than one of justice and equality, we set ourselves up for failure.” – Burns, B. “Why Beyonce Going Vegan is Bad for Animals.” 

21. We have indirect evidence from (i) intermediate outcomes, such as people pledging to go vegetarian, (ii) evidence from other outreach activities outside of animal advocacy (see, for example, our Conversation with Josh Kalla), (iii) anecdotes, such as people who tell advertising organizations that they went vegetarian because of the ads, and (iv) general intuitions and understanding of psychology and sociology. Although we make do with the best evidence we have available, none of this evidence is robust. 

22. Our reasons this include: (i) our Conversation with Josh Kalla indicated that “the more personal the communication the larger the effect size” was a general trend in political advocacy, (ii) deeper per-person communication seems to disproportionately increase the impact on secondary outcomes, such as activism and donations, (iii) we assign significant probability to a threshold model of advocacy, where people need a certain amount of interaction (e.g. 5 minutes of one-on-one conversation) to significantly change their attitudes and behavior, instead of a model where impact scales linearly with time. 

23. We think most people who work on farm animal advocacy would agree with this claim based on their personal experiences interacting with the public. In general, we haven’t found very convincing survey evidence for or against the claim, but there is one survey that found “A total of 64% of survey respondents agree with the statement, ‘farmers and food companies put their own profits ahead of treating farm animal humanely.’” Lusk, J, Norwood, B.F., Prickett, R. (August 17, 2007). Consumer Preferences for Farm Animal Welfare: Results of a Nationwide Telephone Survey. Oklahoma State University. 

24. Although we haven’t performed a full analysis, it appears that public awareness of poor conditions for farmed animals has increased over the last 10–20 years. For instance, we think there has been more discussion of these issues in the media and through documentaries, and there has been more public reaction through seeking out “humane” alternatives and supporting legal and corporate policies that restrict certain practices. 

25. These changes correspond to the growth of undercover investigations as a tactic, and the reactions of farmers, corporations, and the media indicate that these investigations have been one of the key drivers of changing attitudes. 

26. “We have our greatest success in helping others go vegan if we discuss the implications of what we do to animals; the ethical argument is by far our strongest one. Large advocacy groups with access to considerable resources and large followings tell others that we should be asking people to reduce—rather than end—their exploitation, diluting our collective vegan message of social justice and undermining the ethical argument. Too many people are being taught that animal exploitation is okay in moderation, and that the best approach to having a vegan world is to not talk about veganism at all. Be aware of the source of this wrong-headed advocacy approach: pandering by large animal advocacy groups to those engaging in the exploitation.” Taft, C. (November 19, 2015). Mainstream Animal Advocacy Messages Framed By Those Doing The Harm. Vegan Publishers. 

27. “One of these new approaches tried in 2015 was creating a landing page that allowed people to pledge to go vegetarian and download the [Vegetarian Starter Guide], but did not have the factory farming video. This also allowed us to make the text and picture of the Facebook ads reflect the offer of the free VSG instead of the video. Running this type of ad led to a massive drop in cost per conversion (initially to about ⅕ of the previous cost) and MFA’s cost per conversion has continued to decrease over time.” – Conversation with Alan Darer 

28. One study, done in the context of retail advertising, showed that “78% of the increase [in advertising profitability] derives from consumers who never click the ads.” Lewis, R., Reiley, D. (May 28, 2014). Online ads and offline sales: measuring the effect of retail advertising via a controlled experiment on Yahoo!. Quantitative Marketing and Economics. 

29. We do not have data for the difference in click rate for these ads, but Mercy For Animals doesn’t recall a drop in click rates after they adopted pro-veg or “anti-meat” messages. In another example, Mercy For Animals has seen an increase in cost-per-click with their Vegetarian Starter Guide ads (from about $0.026 to $0.041), but this has been outweighed by an increase in conversion rate once on the landing page (from about 3.3% to 40.0%). – Conversation with Alan Darer 

30. “The ads have targeted young women between 24 and 34 years old who have animals or pets listed among their interests.” – Conversation with Jose Valle 

31. “MFA focuses their ads on Latin America because of the high cost-effectiveness, including cheap ad costs and high conversion rates, i.e. the proportion of people who pledge to go vegetarian and sign up with their email address after viewing an ad.…Cost per click in Latin America is just 1–4 cents, whereas in America it can range between 15–25 cents.” – Conversation with Alan Darer 

32. THL gave us the following information for the cost-per-conversion in different languages/regions where they run ads:

    • Spanish (video ads): $0.55
    • Spanish (VSG ads): $0.09
    • Hindi: $7.08
    • Mandarin: $7.68
    • Russian: $3.79 (note that this was data is from a different time period due to technical issues, and historically the Russian campaign has actually been the most expensive)

     

33. Each of the organizations we interviewed for this review about their online ads program gave us this information: Conversation with Jose Valle, Conversation with Alan Darer, Conversation with Cat Liguori and Andrea Gunn 

34. “Mercy For Animals currently receives about 800 emails a week from people replying to emails in our email series who have specific questions about eating vegetarian or vegan and we provide personalized support and answers. Many people also thank us for the meat-free tips and recipes.” – Conversation with Alan Darer 

Online Ads Detailed Review

The following document contains a template for evaluating virtually any intervention designed to help animals. We use this template to guide our intervention research and make it available for others to use in comparing interventions on their own. Observations about the intervention being evaluated can be inserted in the Evaluation and Error-tracking subsections for each Area.

There are many ways for an animal activism organization to pursue its goals, and choosing from these options requires understanding a number of quantitative and qualitative factors. Animal Charity Evaluators (ACE) finds it highly useful to standardize the methods by which proposed interventions are evaluated. This is why we use the following evaluation criteria. The criteria are divided into five areas of focus. For each criterion, we propose a method for summarizing everything that can be known about a given intervention. We hope that this will allow us to compare various campaigns across multiple dimensions and thereby maximize the reliability of our recommendations.

At the moment, there are many unknowns regarding the science, economics, and sociology surrounding animal activism interventions. Much of what is now known about how campaigns translate into reductions in animal suffering comes from small samples, anecdotes, or estimation. For this reason, this evaluation tries to provide as much information as possible, while still recognizing the limits of this young field of research. We do this by providing room for significant qualitative evaluation, while also attempting to give rough quantitative summaries of our work.

The evaluation is organized into five areas of focus. (I) type of intervention, (II) certainty of success, (III) barriers to entry, (IV) expected indirect effects, and (V) expected direct effects and overall efficiency analysis.Areas II through III might be of particular interest to organizations deciding what sort of intervention to pursue. Areas IV and V estimate the expected results from staging the intervention, both direct and indirect, and conclude with a calculation sheet that attempts a cost-benefit analysis.

Theoretically, this final cost-benefit consideration, which estimates the expected return on an organization’s investment, is the only important piece of this evaluation template. However, it is important to supplement this with additional reasoning regarding the likely effectiveness of interventions because we expect these calculations to include significant amounts of error and uncertainty. Such additions are especially valuable where they use different evidence than that which is primarily responsible for our cost-effectiveness estimate, since when only weak evidence is available, many independent lines of evidence can be analyzed, and if all agree, a strong conclusion can still be made. Additionally, given the diversity of organizations, with different resources, goals, and tolerance for risk, we thought it would be useful to give a more holistic picture of what each of our examined interventions entails, while still providing our best guess as to how efficient each one is at reducing animal suffering. As an illustration of the importance of including these auxiliary measures,an intervention with high barriers to entry (e.g. very large startup costs) might not be possible for some organizations, while an intervention with high uncertainties of success (i.e. higher risks) is likely to also require more estimations in our evaluations, and thus be prone to more error in our calculations.

Each area of interest includes an evaluation section with specific questions as well as areas for free-response. Additional forms allowing for summary of the evaluation overall are provided at the end of the document. We suggest using the provided forms during the evaluation process to ensure that all areas have been addressed.

This evaluation process is not without subjectivity. It also requires imperfect estimates of quantitative values. We hope that by publishing the materials used in conducting evaluations, these subjective judgements and estimates will be transparent to our audience. In the case of specific interventions, difficulties in evaluation may arise from:

• limited data regarding certain aspects of past campaigns,
• limited information about the connection between the immediate goals of the campaign and help for animals (e.g. neuroscientific or economic questions that have not been answered),
• biases in the information sources we find,
• limited staff of evaluators (limited ability to cross-check judgements),
• limited time in which to conduct the evaluation,
• other areas we have not accounted for.

Even when it is imperfect, we believe that formal analysis is often a useful supplement to decision-making. We strive to make clear where there are gaps in our knowledge so that our conclusions can be integrated appropriately with other sources of information.

Table of Contents

• Suggested method for the summary of all evaluative information
• Areas of evaluation:
  • Area I: Type
    • I.1–I.9 Classifications of campaign type
    • I.10 Evaluation
  • Area II: Certainty of Success
    • II.1 Precedents
    • II.2 Flexibility of Goals
    • II.3 Scientific Certainty
    • II.4 Other Uncertainties
    • II.5 Evaluation and error tracking
  • Area III: Barriers to Entry
    • III.1 Skill Required
    • III.2 Work Required
    • III.3 Resources Required
    • III.4 Evaluation and error tracking
  • Area IV: Expected Indirect Effects
    • IV.1 Positive influence
    • IV.2 Negative influence
    • IV.3 Evaluation and error tracking
  • Area V: Expected Direct Effects and Overall Efficiency Analysis
    • V.1 Expenditures
    • V.2 Results (“Profits”)
    • V.3 Evaluation
    • V.4 Calculation Sheet Discussion
    • V.5 Error tracking

Summary of the Evaluation

Activism Philosophy	Activism Approach
Online ads are usually an incremental, consumer change approach. They fit into a model of social change driven by consumption, but don’t specify an abolitionist or welfarist position except insofar as they assume reducing the number of farmed animals is a good outcome.	Internet users are shown ads, usually in their Facebook feed, that invite users to click and be directed to a landing page. On that page, pro-veg information is displayed, often in the form of a video showing footage from undercover investigations.

Target Demographics	Target Animal(s)	Size of Campaign	Time Delay between Intervention and Results
Targeted at consumers, usually Facebook users. Most frequent demographic is young and female.	Usually not targeted at a specific species, but the ad or video may include specific animals.	Can be done with very small amounts of time and money, although some organizations spend large parts of their budget on it.	Dietary change resulting from the ads could be immediate or might happen over time if viewers are more likely to change behavior based on future exposure to veg outreach.

General Summary of Area I (Type).

Online ads are typically done on Facebook by showing images with text in the news feed. Viewers who click on the ads are directed to a landing page (for example, www.meatvideo.com) that typically shows an auto-play video (for example, What Cody Saw, where an undercover investigator narrates his experience on animal farms) and invites visitors to pledge to go vegetarian and enter their email addresses. Those who sign up will receive several weekly emails giving information on how to eat a more plant-based diet. These ads could benefit humans by reducing the size of animal agriculture, an industry that many believe is harmful to public health, the environment, and the global food supply. It is relatively easy and low-risk to run online ads without much organizational capacity, especially if they use the materials of other organizations, such as the video or even the landing page itself. However, with more experience and resources, online ads can be optimized to greatly increase their impact, as measured by conversions from dollars spent to pledges to go vegetarian (or take some other action on the landing page). Online ads can bring in support from other organizations, most commonly in the form of advice or materials to help create the ads. The intervention has been growing in prevalence in the past few years as organizations have begun to see it as a cost-effective form of veg outreach and social media has become a more popular advocacy platform. In theory, online ads can also inspire new activists and donors that can benefit the movement, but we’re not sure if this happens much in practice. Finally, insofar as online ads increase demand for animal-free foods, this benefits organizations that sell and promote these products.

Area II: Certainty of Success

Campaigns with identical efficiency ratings can still be differentiated. One way to do this is to focus on the riskiness of an intervention. If efficiency is held constant, riskier campaigns offer a relatively low chance of achieving a relatively high level result. If our estimates of expected value were perfect, we might not care about certainty of success except as it affects organizations’ willingness to attempt a campaign. However, we will not be able to provide perfect estimates of expected value, because in the real world we can never have perfect information. Our estimates of expected value are particularly likely to deviate from reality in the case of risky campaigns, since they succeed only some of the time, and, based on a finite and often small number of attempts, we cannot know the actual success rate with great precision.

II.1 Precedents

The success of future campaigns can be estimated in part by the success of past similar campaigns.

II.1.a Existence of previous attempts

To base predictions on past performances, it must first be established whether or not the proposed intervention has ever been attempted before. If so, it will be useful to know how many times it (or something similar) has been attempted, and whether those attempt(s) were staged in the same setting and context as the intervention currently being considered. The setting and context include such considerations as target demographic, geographical location of campaign, etc. It is generally assumed that unprecedented campaigns are, all things being equal, riskier than campaigns with solid precedents.

II.1.b Variation in efficiency of previous attempts

If a similar intervention has previously been staged, it is important to understand the variation in both costs and results. Campaign types with a history of unpredictable costs and outcomes are riskier than those with a less chaotic performance record. Wild fluctuations in run costs or in successful outcomes achieved will translate directly to wildly fluctuating measures of efficiency.

II.2 Flexibility of Goals

Campaigns that have lofty, all-or-nothing goals are inherently riskier than campaigns with continuous spectrums of potential positive outcomes. Campaigns that do not have fallback victories will be riskier than campaigns that do. For example, a low-visibility campaign to convince a corporation to stop supporting a certain type of animal testing may fail to produce any positive result if it is not successful. It is thus important to find out if there is only one goal set for the campaign, or if there are a variety of possible outcomes that will satisfy the purpose of the campaign. Campaigns that cast wider nets are less risky than ones that focus on narrow areas of improvement. For example, a gestation crate awareness campaign may have multiple potential goals, including getting its audience to adopt vegan diets, getting its audience to purchase pork from more humane sources, etc.

II.3 Scientific Certainty

There are a variety of scientific disciplines that study issues relevant to the success of animal activist interventions. The extent to which scientists have succeeded at understanding factors relevant to a particular intervention affects our certainty that the intervention will have a positive effect.

II.3.a Nonhuman neuroscience

There is not always clear scientific consensus regarding the neural complexity of nonhuman animals. A greater understanding of an animal’s neural networks allows us to be more certain that it is affected by allegedly harmful practices. For example, a campaign focusing on fishing may have to temper its scientific certainty more than a campaign focusing on pigs (given what is known about the latter’s neuroanatomy, and what is not known about the former’s). In making this determination, it may be necessary to read into the neuroscientific and/or psychological literature on the target animal population in question. We expect this to be a particularly controversial area of research, which will likely evolve along with scientific understanding of nonhuman animal consciousness.

II.3.b Sociology, psychology, and economics

Many campaigns rely upon assumptions about how humans will respond to an intervention, how societies will adapt to proposed change, and/or how economic forces will react to a campaign. It is thus necessary to uncover any ambiguities in the relevant literatures supporting the social science assumptions of an intervention. For example, the reliability of data on recidivism rates amongst new vegans will be important in estimating the certainty of success of a diet-focused campaign.

II.3.c Ecology

The expected effects of an intervention may rest on ecologically-based predictions. These interventions also must be rated based on reliable ecological data. For instance, it may be important to understand how factory farms impact their surrounding environment before being able to reliably quantify the effects of an anti-factory farming campaign.

II.4 Other Uncertainties

There may be other miscellaneous risk factors not yet addressed.

II.4.a Known unknowns regarding an intervention

Some risks are unpredictable, but others may be anticipated with proper planning. Before beginning an intervention, it is advisable to go through all components of a plan of action in order to determine what might go wrong, however unlikely. This is purposefully vague, as this method can be thought of as a “catch-all” for types of uncertainties thus far not addressed. For example, a certain legislative campaign may have to take into consideration the possibility of a cooperating political partner losing a reelection bid. Another campaign may have to take into consideration unfamiliarities in the culture of their target demographic, or a lack of certainty that all members of the campaign team can stay on for the duration of the campaign.

II.4.b Distance between intervention and improvement in animal welfare

The distance between a campaign and the animals it intends to benefit can be defined as the degree of separation or number of steps between an intervention’s activities and the end result of actually helping animals. For example, a campaign to care for rescue animals has a direct impact on animal welfare, while a campaign to train new animal activists would expect to achieve far-removed, indirect results (training → new activists → new campaigns → improvement in animal welfare). Similarly, a campaign to pass a new animal welfare law would also be further removed than caring for animals directly. A greater distance between intervention and result leads to a greater uncertainty of success.

II.5 Evaluation and Error Tracking for “Certainty of Success”

Evaluation for II.1.a

How many times has a similar intervention been attempted before?

1	2	3	4	5	6	7
1= No past attempts 7= Multiple organizations have started this type of campaign each year for at least the past decade.
Notes: Online ads are usually a continuous intervention and not easily divisible into campaigns. We give the rank of “5” here because at least one organization, Mercy For Animals, has been running these ads for several years, and several others are now also running online ads.

Has a similar intervention ever been attempted in the same context as the one being evaluated?

1	2	3	4	5	6	7
1= Few similarities, and no exact parallels, exist between past campaigns and one being evaluated, in terms of target demographic, geographic location, organizational similarity, etc. 7= Multiple campaigns have been previously attempted that have identical or near identical contexts as the one being evaluated, in terms of the factors listed above.
Notes: Most online ads follow the same general format described above. Some changes in delivery, like having ad-clickers go to a website sharing “how to” information as opposed to a video showing the “why” of reducing/eliminating consumption, could make substantial differences in impact. Changes in demographics could also make a substantial difference, particularly with campaigns that target countries and languages that the advertising organization isn’t very familiar with.

Evaluation for II.1.b

How widely have previous campaigns varied in total results?

1	2	3	4	5	6	7
1= Very wide variation; some campaigns have achieved remarkable results, while others have been completely unsuccessful or counterproductive 7= Very little variation; it is difficult to detect any meaningful difference in results
Notes: Ads are usually not done in campaigns, but impact (as measured in veg pledges per dollar) does vary with time, organization, demographics targeted, and the content of the ads themselves. Figures range from around 10 cents per conversion to several dollars.

How widely have previous campaigns varied in efficiency?

1	2	3	4	5	6	7
1= Very wide variation; some campaigns have had large results with few resources, while others have been completely unsuccessful or counterproductive 7= Very little variation; all variation in results can be explained by variation in resources put in
Notes: Because online ads aren’t clearly separable into campaigns, this answer is the same as the previous one.

Comment on what is known about the reason(s) for the above mentioned variations between campaigns. Specifically, what variables were external and which were specific to a particular campaign/organization.

It’s not very clear why some ads are more effective than others, but organizations like Mercy For Animals (MFA) are able to get very high conversion rates through continuous and exhaustive A/B testing. Some trends have emerged. Often young women end up being the most cost-effective demographic. The video What Cody Saw seems to be quite effective. And the new ad strategy, pioneered by MFA, that focuses on the “how to” of vegetarianism rather than the “why,” has shown promise. Any group can utilize these advantages.

How well have previous predictions of results matched actual results?

1	2	3	4	5	6	7
1= Very poorly; advance predictions do not appear to correlate with campaigns’ eventual success 7= Very well; successful campaigns were in general widely expected to succeed and unsuccessful ones were clearly longshots
Notes: The trends are somewhat reliable, but there is still a large amount of variability and experienced ads campaigners can still be surprised by which ads have the highest conversion rates.

How well have previous predictions of costs matched actual costs?

1	2	3	4	5	6	7
1= Very poorly; most campaigns have either raised funds they were unable to use or had unexpected expenses leading them to need more funding than projected 7= Very well; campaign costs have generally matched the planned budget closely
Notes: It is quite easy to monitor the costs of your ads and ensure you’re only spending as much as you planned. On Facebook, you’re able to set your budget prior to running the ads.

Evaluation for II.2

How many different avenues of success does the campaign have available?

1	2	3	4	5	6	7
1= The campaign has only one major goal or path towards success 7= The campaign has several distinct goals that could each be reached independently
Notes: Online ads are primarily focused on individual dietary change. Other goals are possible, of course, such as influencing policy leaders who happen to see the ads, or creating activists who then share pro-animal messages with their social circles.

How many intermediate positive outcomes does the campaign have?

1	2	3	4	5	6	7
1= Partial success on any of the campaign’s goals has no clear benefits 7= Any step towards any of the campaign’s goals has proportional benefits
Notes: Changing diets has benefits at any scale as it reduces the expected number of farm animals. Of course, major spending on online ads can allow for more optimization and creation of better materials, which could increase the marginal benefits.

Evaluation for II.3.a

How well understood are the neurological abilities of the primary group of animals targeted by the campaign?

1	2	3	4	5	6	7
1= Understood mostly on an intuitive level 7= Well understood behaviorally and on a neurological level
Notes: Online ads tend to focus on a general reduction in the consumption of animal products. We don’t understand the neurological abilities of fish and birds as well as we do mammals, but we understand them enough to feel confident that helping them is very important.

How well understood are the neurological abilities of any other groups of animals targeted by the campaign?

1	2	3	4	5	6	7
1= Understood mostly on an intuitive level 7= Well understood behaviorally and on a neurological level
Notes: For the purposes of this report, we’re only considering the impact on farm animals. However, it is possible that changes in animal product could affect other animals such as wildlife, perhaps due to pollution or habitat destruction.

Comment on the choice of target group(s) of animals by the campaign. Address how well the goals of the campaign align with the ability of the recipient animals to meaningfully benefit from those goals, based on what is known about their neurobiology.
The welfare of farmed animals seems very poor on average, and based on our understanding of their neurobiology, it seems to be a very good thing to cause fewer animals to endure this suffering.

Evaluation for II.3.b

How well understood are individual and social group responses to this type of campaign?

1	2	3	4	5	6	7
1= Understood mostly on an intuitive or theoretical level 7= Many rigorous studies show very similar responses
Notes: We have some decent evidence that some individuals respond to online ads by reducing their consumption of animal products by some period of time. This is mostly in the form of intermediate outcomes, such as signing the pledge to go vegetarian or emailing the advertising organization with questions about changing their diets. However, we are very uncertain about how this relates to dietary change, and even more uncertain about its effects in a wider social context. For example, the emphasis on helping animals through individual changes in consumption could reduce the public impetus for institutional or society-wide changes.

How well understood are economic factors affecting the outcome of this campaign?

1	2	3	4	5	6	7
1= Understood mostly on an intuitive or theoretical level 7= Many rigorous studies show very similar responses
Notes: We feel fairly confident that dietary change has a substantial impact on the number of farm animals bred and slaughtered.

Evaluation for II.3.c

How well understood is the primary ecological impact of this intervention?

1	2	3	4	5	6	7
1= Understood mostly on an intuitive or theoretical level 7= Many rigorous studies show very similar responses
Notes: Reduced demand for animal products could reduce the pollution and environmental damage that results from most animal agriculture practices. It could also increase habitat as less farmland is used for feed, housing, and other agricultural activities. There is substantial research on the environmental impact of animal agriculture, but little on how it affects wild animal welfare.

How well understood are any other ecological impacts?

1	2	3	4	5	6	7
1= Understood mostly on an intuitive or theoretical level 7= Many rigorous studies show very similar responses
Notes: There are no other apparent ecological impacts.

Comment on any perceived oversights or uncertain assumptions in the scientific, sociological, economical, and/or ecological rationale behind this campaign.

In addition to the uncertainties listed above, we have substantial uncertainty about the impact of online ads on dietary change. We have evidence available from intuition, intermediate outcomes like veg pledges, other applications of online ads like increasing product sales, and our general understanding of human decision-making. The most robust study to date on the effects of online ads was published by Mercy For Animals in February 2016. We see it as an important piece of evidence in assessing the impact of online ads and elaborate on our thoughts here.

Evaluation for II.4.a

How many known uncertainties affect the outcome of the campaign?

1	2	3	4	5	6	7
1= None 7= Several different sources of uncertainty not addressed above
Notes: The main uncertainties affecting the primary outcome, individual dietary change, are the likelihood of dietary change after an ad view, the amount of dietary change made by the average viewer who changes their diet, and the amount of good done for each portion of animal products that aren’t consumed. Because we put less emphasis on other potential outcomes, this is a relatively small number of uncertainties.

How strongly do known uncertainties affect the outcome of the campaign?

1	2	3	4	5	6	7
1= No known uncertainties affect this campaign 7= The success of the campaign hinges upon one or more uncertain factors not explicitly addressed by the campaign itself
Notes: The success of the campaign does hinge on each of these uncertainties. We think the amount and likelihood of dietary change resulting from the ads could be low enough to make the campaigns unsuccessful, but it is unlikely a substantial reduction in consumption of animal products, if it occurs, would not have a substantial benefit for the animals.

What outside actors and external events are expected to be able to affect this campaign’s success?

Facebook (or other advertising service) – Their prices and services, e.g. what other content the user is seeing when they see the ad, can affect its effectiveness. Ad blockers – If more or less people block ads on their browsers, it can affect how many people actually see the ads. Animal-free food technology – An increased availability of these foods can make the target audience more likely to change their diets based on the ad’s suggestions. Other activism and social change – An increased public interest in veg eating and helping farm animals can lead to increasing returns as people take more interest in the ads. It could also potentially saturate the market and target audience, but that seems like it won’t be the case in the near future. Other ads – Other ads received by the target audience (as well as non-ad influences) can affect receptiveness. For example, a large number of ads asking for behavior change could make someone less likely to change based on veg ads. Other ad campaigns can also affect prices.

What unanticipated problems have befallen past campaigns? How were they handled?

People fail to adopt veg eating for a number of reasons, such as perceived reduction in quality of food or a lack of friends who are vegetarian or vegan. It’s unclear which roadblocks have most affected the effectiveness of online ads, but people running online ads have noted substantial differences in cost-per-conversion for different campaigns at different times. One example of a difference in effectiveness has been between standard online ads that show a video intended to provide motivation to change one’s diet, often including footage from undercover investigations, and ads that focus on how one can adopt a vegetarian diet. The latter tend to be much more effective in terms of vegetarian pledges per dollar. The “problem” that these ads solved might have been that a large number of people clicking on the standard ads weren’t actually interested in dietary change, while the new ads target a more narrow audience that is already considering veg eating.

Evaluation for II.4.b

List the chain of events connecting this intervention to a change in animal welfare, in as much detail as possible.

An organization sets up the ads On Facebook, this includes setting up an account and attaching a payment method. The ads are designed, usually with a compelling image and short text that have a pro-veg eating message. A landing page is set up for those who click on the ads, usually with a video or text-based information and a short form to make a pledge to go vegetarian or reduce one’s meat consumption. Other features, such as an e-mail series sent to pledgers providing more information on veg eating, are created. The ads are shown to viewers This usually involves making purchases from a service like Facebook, often for a certain number of clicks or impressions, and usually delivered to a specific demographic, at least specified by age and gender. Viewers engage with the ads On Facebook, ads are usually shown in the Facebook news feed as a “Sponsored” item. Some viewers click on the ads, leading them to the landing page. Some of those viewers then sign the pledge. Each of these three segments might cause behavior change that helps animals, including the reduction in consumption of animal products, and attitude change that can affect future behavior. Fewer animals are consumed, leading to fewer animals produced As consumers demand fewer animal products from grocery stores, restaurants, and other sellers, those companies will, in expectation, demand fewer animal products from producers who will, in expectation, reduce their production. This means fewer animals will endure the suffering of animal agriculture. Ad viewers could be impacted in other ways Although the ads are primarily intended to create dietary change, viewers could also help animals in other ways, such as by voting and pushing for humane reforms that improve farm animal living conditions, becoming activists themselves, or donating to animal advocacy organizations. Some landing pages include a button to like the Facebook page of the advertising organization. Being engaged with these pages could lead to decreased recidivism from dietary change, sharing social media content, or other positive outcomes.

How many steps separate the actions of the campaign from a change in animal welfare?

1	2	3	4	5	6	7
1= None; the campaign helps animals directly 7= Very many; the chain is so long that it is difficult to think about in detail
Notes: The most uncertain step in the chain is when viewers actually make meaningful behavior changes. Outside of this, the route to impact is relatively straightforward, although it can become more complex when considering other impacts and long-term effects.

How many different actors are involved in the chain of events between the success of the campaign and the change in welfare?

1	2	3	4	5	6	7
1= None; the campaign helps animals directly. 7= Very many; the chain is so long or involves such complicated events that it is difficult to identify all actors or groups involved
Notes: The number of actors is relatively small, involving basically just the supply chain from consumer to animal product producer.

Summary

General Summary of Area II (Certainty of Success). Summary may include relevant comments about the following topics: what has been learned from previous campaigns; how narrow or flexible are the goals of the campaign; how clearly is the necessary scientific, sociological, economic, and ecological science upon which the campaign’s success relies understood; what is the (metaphorical) distance between the campaign and the animals it is attempting to help; and what other uncertainties should be noted.

It seems clear that spending on online ads results in relatively large numbers of people being exposed to advocacy for veg eating, and that any dietary change will result, in expectation, in a large benefit for animals. We are very uncertain about how much dietary change, and other impact, occurs due to the online ads. We know that a substantial number of people who click on ads then pledge to go vegetarian or reduce their consumption of animal products, and some even contact the advertising organizations to ask questions about veg eating and say personally that they are changing their consumption. It is also unclear what impacts, if any, online ads have on the broader state of affairs for animal agriculture and how animals are viewed in society. We know of little, if any, historical precedent that shows major social change, such as a 20%+ reduction in the number of animals raised for food, is caused by individual advocacy and changes in personal consumption. This means that, although online ads might cause an immediate reduction in the demand for animal products, they might not have as much impact on animal welfare in the long-run as other interventions.

Error Tracking

• How rigorously documented were the reports from past similar campaigns? How reliable is the information that defines the precedents for the campaign?
  • The main organizations we know about running online ads all keep careful track of their metrics, such as cost per click. Robust evidence of the effect of these ads on behavior is lacking.
• What ambiguity exists in the scientific consensus regarding the mental abilities of relevant animals, the sociological and economic effects necessary for the campaign to make a difference, and the campaign’s ecological effects?
  • It is clear that many farmed animals, like chicken and fish, have mental abilities that give them at least some moral importance to most people. While there is substantial ambiguity in the scientific consensus about animal consciousness and suffering overall, online ads are not more subject to this ambiguity than other interventions.
• What assumptions were made regarding “other considerations?” What factors may not have been considered?
  • The technology used to place online ads, e.g. the targeting on Facebook, might change in a way that makes them more or less effective; our understanding of economics, that decreased demand decreases production, could be off.

Area III: Barriers to Entry

There are a myriad of factors to consider when evaluating how difficult an intervention will be to stage. This area of the evaluation examines the complexity of human skill required, the degree of work intensity needed, and the difficulty involved in obtaining the required material resources. This is not meant to duplicate the valuation mechanism laid out in Area V (Efficiency), but rather to provide a separate metric to measure the challenges inherent in putting on a campaign, with no reference to a conversion to fiscal units. This provides part of a cross-check for Area V, since the higher the barriers to entry for a campaign, the better its probable results must be to maintain a constant level of efficiency. In addition to providing a check on the efficiency as calculated below, knowing the barriers to entry for an intervention can help identify which types of group can likely conduct it successfully.

III.1 Skill Required

It is important to know what sort of expertise is required from a campaign’s staff. The more specific knowledge and/or special skills required, the more difficult a campaign will be. For example, a campaign to challenge an anti-whistleblower law will need significant legal expertise and political skill.

III.2 Work Required

The calculation of work required can be defined as a calculation of the intensity of effort required from an intervention’s staff members.

III.2.a Intensity of the work

Intensity is an attempt to quantify how exhausting, physically taxing, and mentally draining campaign-related work is. A more demanding workload indicates a more difficult campaign plan. For example, a campaign to engage strangers on a city street and talk to them about veganism may require exhausting days with lots of mentally taxing social interaction.

III.2.b Hours per week

The number of hours of work per week demanded from staff members compounds the effect of the intensity of their work per unit time. It will be useful to know how many hours per week are required from each type of staff member both at a minimum and at a maximum. A campaign may require a greater commitment around major events.

III.2.c Number of people needed

The number of staff members required for a campaign can be a limiting factor. Like the number of work hours required per week, this value may fluctuate over time.

III.3 Resources Required

It is further useful to examine the complexities involved in obtaining the necessary resources for an intervention. The procurement of materials for an intervention may require difficult-to-obtain permissions, such as copyright waivers, or an extensive search for uncommon objects. For instance, a campaign to develop a new meat substitute that requires a rare chemical has inherent difficulties associated with resource obtainment.

III.4 Evaluation and Error Tracking for “Barriers to Entry”

Evaluation for III.1

How many specialized skill sets are required?

1	2	3	4	5	6	7
1= None 7= Many unrelated and specialized skill sets
Notes: The campaign requires an understanding of the advertising system being used, such as Facebook. A non-specialized skill set of being good with numbers, adaptive, and creative in coming up with new ads, is also useful.

What special skills are required in conducting this campaign?
The only special skill required is an understanding of the advertising system being used, such as Facebook. This can be picked up quickly. Additional specialized skills can help create new materials, such as videos for the landing page.

How much expertise is required in the most demanding area?

1	2	3	4	5	6	7
1=Very little experience or skill 7= Many years’ experience or extensive professional training and advanced degrees
Notes: The expertise to run online ads on Facebook can be picked up in just a few days, although more time can lead to more effective campaigning. It takes more expertise to create a landing page, images, videos, or other materials that are part of the ads, but an organization doesn’t need to create those itself to run ads.

Comment on the expected ease with which qualified staff members will likely be recruited. This may involve such factors as the job market for a certain profession, the average interest in animal activism from people with certain qualifications, etc.
Because of the lack of specialized skills needed to run online ads, we expect recruiting to be relatively easy. Additionally, one staff member can run a large number of ads. However, there is a high skill ceiling in managing online ads (meaning the best employees can do a lot more than average employees).

Evaluation for III.2.a

How demanding is the typical workload associated with the campaign?

1	2	3	4	5	6	7
1= Light; working on the campaign involves little stress and is likely pleasant 7= Heavy; working on the campaign involves significant physical or mental labor at most times
Notes: While it can be challenging to run ads in a way that’s flexible and relatively cost-effective, running them isn’t particularly likely to cause stress or trauma. Of course, the ads frequently include graphic imagery, so working on them can be stressful insofar as the person running them is repeatedly exposed to that content.

What challenges are involved with the daily work of the campaign?
The main challenge is intellectual, as the ad manager uses creativity and analysis to improve the effectiveness of the ads and create as much dietary change as possible. It can also be challenging to work with the survey software and representatives of the advertising companies, such as Facebook.

Evaluation for III.2.b

How many hours per week does the typical staff member work on the campaign when it is making its minimum demands?

1	2	3	4	5	6	7
1= 1 or fewer 7= 80 or more
Notes: Usually it doesn’t take a full 40 hours per week to run the ads of one organization, although this does vary. Ad managers can also fill other roles in the organization, such as running different forms of outreach.

How many hours per week does the typical staff member work on the campaign when it is making its maximum demands?

1	2	3	4	5	6	7
1= 1 or fewer 7= 80 or more
Notes: Workload can increase when creating a new campaign or designing new ads. It can be especially demanding if the staff is creating new content to use in the ads, such as a new video.

Note that if the campaign involves several different types of staff roles with different time demands, it may be helpful to answer the above questions for each role rather than for the typical staff member.

Evaluation for III.2.c

How many staff members are involved with the campaign when it is ongoing but at the time when fewest people are working on it?

1	2	3	4	5	6	7
1=It runs automatically with at most one person supervising it 7= Over 50
Notes: There is usually one person involved in managing an organization’s ads for most of the time, since they just need to make relatively simple edits to the materials, change targeting and other ad factors, and do other day-to-day activities to ensure cost-effectiveness.

How many staff members are involved with the campaign at the time when the most people are working on it?

1	2	3	4	5	6	7
1=It runs automatically with at most one person supervising it 7= Over 50
Notes: Usually other staff members help to develop new materials or make other substantial changes to the ads. For example, if the organization wants to create a new video, this could take one or two additional staff working full time, depending how quickly it will be created.

Evaluation for III.3

How many rare or difficult-to-obtain materials does the campaign require?

1	2	3	4	5	6	7
1= None 7= Several which are crucial to the success of the campaign
Notes: The materials involved such as images and videos are publicly available online.

How severe is the difficulty of obtaining the hardest-to-obtain material needed for the campaign?

1	2	3	4	5	6	7
1= There are no materials which will be difficult to obtain 7= There may be an insuperable obstacle to obtaining this material
Notes: There are no hard-to-obtain materials.

What resources does the campaign require that may be difficult to obtain?
None.

Summary

General Summary of Area III (Barriers to Entry). Summary may include relevant comments about the following topics: the level of complexity of the skill sets needed for the campaign; the intensity of the work; and the feasibility of acquiring needed material resources

Running an online ads program involves getting familiar with an advertising platform, such as Facebook; creating or using ads from another organization, including a landing page; then purchasing the ad space and running them. Usually, organizations monitor the ads and continuously try new ones. This usually takes one part-time staff member to implement, but more work can be required, especially when creating new content such as videos.

Error Tracking

• What differentiates the campaign from past similar campaigns? Could any of these factors affect the difficulty of achieving the campaign’s goals? Keep in mind that it may be difficult to ascertain which factors may have contributed to a past campaign’s success/failure.
  • New ads can use existing resources, such as images and videos, as well as the knowledge accumulated over time of what is most cost-effective. Additionally, as the perception of veg eating and animal farming changes in society, this can affect the success of new ads.
• How rigorously documented were previous campaigns? Could estimates of staff time spent etc. be in error?
  • It seems easy for organizations to keep track of budgets and staff time spent on online ads. Breaking this down further, such as how much time is spent on specific tasks like creating new ads or learning about new advertising strategies, can be less reliable.
• Have environmental factors changed which would affect either the skills required to conduct the campaign or the difficulty of obtaining necessary materials?
  • As more ads are run, more resources such as images and videos used in the ads are accumulated. These materials can go “out of date,” for example, if they refer to a celebrity who is no longer in the public eye, but that effect is not very worrisome in the short term.

Area IV: Expected Indirect Effects

Oftentimes, a campaign will have effects that reach beyond its intended goals. These may be positive (i.e. in-sync with the ethics driving the campaign) or negative (i.e. working against the change effected by the campaign). There are multiple factors to consider when evaluating external influence. These effects are especially difficult to estimate precisely because they are often diffuse and because even when campaigns have carefully tracked their progress towards their explicit goals, they may not have attempted to track other effects. However, diffuse and indirect effects of some campaigns may be significant parts of the campaign outcome, so we consider any evidence about these effects as a factor in evaluating interventions.

IV.1 Positive Influence

Good externalities may come from a campaign.

IV.1.a Chain effect

A chain effect is when those affected by a campaign spread the message of the campaign and thus extend the influence of the campaign’s message. For example, an online advertising campaign might have a particularly high positive chain effect, due to the ease with which a viewer can share a website or video link with others.

IV.1.b Benevolent slippery slope

It is possible that the success of the campaign will indirectly inspire other positive victories due to social pressure, or desire to imitate. For example, if multiple states enacted bans on gestation crates, some pork producers in other states might bow to public pressure and voluntarily phase them out themselves.

IV.1.c General education

General education is when a campaign instills a certain mindset in some subsection of the public that will go on to produce beneficial outcomes. For example, a campaign to oppose “ag-gag” laws laws might cause some people to develop a negative attitude towards animal agriculture in general, regardless of whether or not the campaign itself is successful.

IV.2 Negative Influence

Bad externalities may also arise from an intervention

IV.2.a Alienation

An imperfect campaign may run the risk of turning some potential supporters away from the cause and/or inciting a backlash against the goals of the campaign. For example, a public awareness campaign that uses offensive imagery to make its point may risk alienating potential supporters and/or fueling the influence and credibility of its detractors.

IV.2.b More harm than good

There is also be a chance that the intended results of a campaign might actually be more harmful than beneficial to animals. This is a difficult-to-spot type of negative externality because it requires Researchers to recognize when something is being overlooked in how a certain abusive system would respond to a successful attack against it. For example, a campaign against the consumption of beef might merely push people to consume more chicken (which would be particularly bad as it takes the meat of roughly 200 chickens to equal the same amount of meat from a single cow).

IV.3 Evaluation and Error Tracking for “Expected Indirect Effects”

Evaluation for IV.1.a

How large a chain effect does the intervention likely have?

1	2	3	4	5	6	7
1= No reason to expect a chain effect 7= Campaigns often go viral, with a high percentage of people who see hear the message sharing it in close to the original form
Notes: Although some viewers see online ads and share them with others, usually on social media, it generally seems to be more of a one-by-one intervention.

Evaluation for IV.1.b

How likely is the campaign to inspire other victories?

1	2	3	4	5	6	7
1= No reason to expect the campaign to inspire indirect victories 7= Most similar campaigns have inspired successful imitations or victories beyond their expected scope
Notes: The effects of online ads on other campaign victories are likely just the effects that come from having a larger veg eating population. For example, this can mean increased demand for vegan options that increases the institutional adoption of vegan foods.

Comment on any evidence (especially taken from observations of past campaigns) suggesting ways in which this intervention may directly or indirectly inspire other activism movements.
Mercy For Animals’ use of online ads, and the large numbers of people they’ve been able to reach at low costs, has been a driving factor in the adoption of online ads programs by other organizations. The availability of videos, landing pages, and other resources accumulated by running ads also makes the creation of more online ad programs more likely.

Evaluation for IV.1.c

How significant a change in public mindset is this campaign likely to cause?

1	2	3	4	5	6	7
1= No change; the public will be entirely unaware of or uninterested in the campaign 7= A significant change; the campaign will reach most of the population with a message that is highly contagious and represents a significant departure from current norms
Notes: Although potentially many people change their diets due to online ads, the individualized focus of the ads likely leads to a limited impact on the public mindset as a whole. If one thinks human decision-making is more individual-determined, rather than socially driven, and thinks online ads create large amounts of individual change, then this score could be much higher. Online ads also tend to discuss veg eating as a personal choice, which is already a view most people hold, as opposed to stronger views like veg eating as a moral or social imperative.

Evaluation for IV.2.a

How much alienation is this campaign likely to cause?

1	2	3	4	5	6	7
1= None; this campaign has no apparent controversial features 7= Significant alienation; a large contingent of potential allies will likely take offense at some aspect of the campaign, and the campaign may be perceived as representing the overall position of a large number of animal advocates
Notes: Online ads usually frame veg eating as a personal choice, and as an opportunity to help animals. They don’t focus on the harm done by people who eat animals or make accusations against those people. Recently, ads have become increasingly focused on the “how” of veg eating rather than the “why,” which also makes them seem less alienating. However, any promotion of veg eating risks being taken as an attack on people who eat meat. And since people are used to online ads trying to get them to change their behavior and online ads don’t disrupt the viewer’s other activities, they are probably perceived as less aggressive than, e.g. protests.

Comment on the ways in which this campaign might alienate certain members of the public. Specifically, who might be alienated, and how damaging might the alienation be, in terms of negative impressions carrying over to other similar organizations?

Online ads risk alienating people who feel that whether one eats animal products is a personal choice and one shouldn’t try to change other people’s behavior. This concern is reduced the more the ads present themselves as sharing information to encourage better decision-making rather than outright trying to change behavior. Some viewers could object to the content in the ads. For example, they could feel that investigation footage misrepresents the animal agriculture industry. This objection doesn’t seem to be very common, since most people outside of the animal agriculture industry seem to agree there’s a lot of mistreatment in the industry, even if they believe some farm animals live good lives, usually on smaller farms. Viewers could also feel the ad was inappropriate because it showed gruesome footage of suffering animals. This objection also seems very limited, especially because viewers usually have to click on the ad and watch the first part of the video before seeing any upsetting footage.

Evaluation for IV.2.b

How high is the risk that the campaign may be more harmful than beneficial?

1	2	3	4	5	6	7
1= Extremely low risk; the campaign will have few, highly controlled effects 7= High risk; the campaign ignores significant parts of the system it is working in that pose a threat to animals
Notes: In the short-term, online ads are unlikely to increase farm animal suffering because they simply reduce the number of animals being raised for food. In the long-term, it’s possible that online ads could increase the perception of veg eating as a personal choice, trend, or fad, which could prevent a further reduction in animal product consumption due to seeing it as a moral or social imperative.

What specific practices does the intervention seek to change? Would the replacement practices necessarily be an improvement for the animals involved?

Online ads primarily seek to reduce consumption of animal products, which leads to fewer animals being bred and raised for food production. Most animal advocates see this as beneficial, because farm animals live in circumstances which appear to be extremely unpleasant and demonstrate a variety of stress behaviors in response to these conditions. However, it should be noted that population ethics is difficult even when applied to humans who can actually express a preference between their current circumstances and never having been born, and there is necessarily more uncertainty with regard to the preferences of animals who cannot communicate about abstract topics.

Summary

General Summary of Area IV (Expected Indirect Effects). Summary may include relevant comments about the following topics: indirect reactions to the campaign (positive and negative); and long-term externalities of the campaign (positive and negative)

Online ads seem to have a limited effect outside of the behavior of viewers. For example, ads don’t receive press attention, and it seems like they have less of a community-building impact than other interventions. However, given the large numbers of people exposed to ads, they could indirectly affect the general attitudes and behaviors of consumers. They are unlikely to alienate or anger their audience, although they could have negative effects by shifting the public perception of veg eating and farmed animal suffering towards viewing it as a matter of personal choice.

Error Tracking

• What assumptions have been made?
  • Our reasoning in this section is based largely on intuition and speculation due to limited empirical evidence, but we have tried to account for all possible effects rather than making definitive assumptions.
  • We also note that the effects of animal agriculture on wild animal suffering are extremely unclear, and that uncertainty affects several parts of this review. If we were to view animal agriculture as beneficial to wild animals, perhaps by reducing the number of organisms who live miserable lives that don’t seem worth living, that could lead to sweeping changes of our current views.
• How reliable is the data regarding past campaigns?
  • We do not have data on the long-term effects of any specific campaign, and we have very limited data on the long-term impact of online ads as a whole. We can observe changes in public attitudes over time in regions where ads have been run, but it seems very difficult to discern the effects of ads from other factors.
• How reliable is data predicting expected responses to various campaign tactics?
  • We have little data predicting long-term or indirect impact.

Area V: Expected Direct Effects and Overall Efficiency Analysis

This area of evaluation is designed to produce a single ratio of dollars spent per unit of suffering reduced or per number of animal deaths averted. We believe this is a useful factor to consider when planning an intervention, as different types of campaigns can have vastly different cost/benefit ratios. Given the limited amount of funds available to an organization of any size, this section explores ways of making sure that those resources are used optimally. We break down this calculation into component expenditures and results. If we could rely entirely on the results of this section, for many purposes we could omit the other sections of the evaluation entirely. However, in reality our estimates of both costs and expected results will be fallible, so reasoning about an intervention’s likely effectiveness from other perspectives (as above) allows us to be more confident of our conclusions than we could be about a direct efficiency analysis alone.

V.1 Expenditures

A successful intervention requires the input of resources, both financial and otherwise. It is important to understand and quantify the full extent of the costs to run a campaign.

V.1.a Monetary costs

Monetary costs include both upfront costs to begin an intervention and maintenance costs to sustain an intervention over time. Upfront costs are expenditures that go towards things such as the recruitment of volunteers, the purchase of office supplies, consulting fees, etc. Such expenses can be a limiting factor for an intervention if insufficient liquid capital is available. Maintenance costs are expenditures that are expressed in the form of dollars per unit time or per event. They include things such as material expenses (e.g. pamphlets, gas for vehicles, etc.) and money allocated to train and compensate workers (e.g. salaries for staff, costs involved in educating staff, costs involved in training volunteers for an event, etc.). Maintenance costs also involve non-specific overhead, such as office leases, travel expenses, and salaries/training costs that are not specific to any one event within a campaign. The amount of monetary resources needed for a campaign directly relates to how long a campaign can be sustained for and whether it can even be undertaken in the first place.

V.1.b Personnel

Personnel includes the number of people needed and the level of expertise needed to run an intervention. The number of people needed includes staff, long-term volunteers, single-event volunteers, outside help, consultants, and media coverage by animal activist journalists. While the cost of salaries is covered under “monetary” expenses, there is a clear opportunity cost of using up the time of people who are committed to working on animal activism projects. A poorly planned campaign could pull activists away from other, more effective interventions. The level of expertise required from the personnel includes both pre-existing expertise and the amount of training it is necessary to provide to new recruits. This could range from little to no expertise (e.g. one-off volunteers recruited for discrete events), to high expertise in animal law, high expertise in policy, high expertise in campaign management, high expertise in data analysis, etc. Recruiting the services of a person with a high level of expertise relevant to animal activism carries a higher opportunity cost, given that person’s theoretical ability to do more for a different activism campaign than the average activist.

V.1.c Time

Time refers to the time expenditure by an intervention’s staff, including both time committed during the intervention and time committed to pre-campaign training and planning. The former category involves such things as hourly shifts at events, weekly staff time, one-off consulting appointments, etc. This again is proportional to the magnitude of opportunity cost. TIme committed before an intervention begins includes time for event preparation, market research, experiment conducting, data collection, etc. This is time invested in laying the groundwork for a campaign. A campaign with a very short startup time may in some cases be preferable to a campaign that has slightly more potential effectiveness per unit time, but which also has a long startup time.

V.1.d Unknowns

Unknowns describes the category of unplanned expenses. While unknown expenses are by definition unexpected, it is possible to attempt to consider a range of possible emergency expenses and multiply the cost of each by the likelihood of the unplanned event occurring. This category includes such expenses as legal payments from being sued by a group or person, replacement costs for event materials lost on-site, etc.

V.2 Results (“Profits”)

After estimating expenditure of resources, we estimate the expected “return on investment,” measured by the degree of success a campaign achieves in reducing animal suffering or saving animal lives.

V.2.a Breadth

The breadth of an intervention’s results is defined as the number of people reached by the campaign. It can be expressed as the product of the number of people contacted per event or per time, and the number of events or units of time within the intervention. It is important to understand what is meant by the phrase “reached by the campaign”. A person does not have to be directly contacted by a staff member in order to be “reached”. A legal campaign to change the way a certain farmed animal is raised would (if successful) “reach” every person who consumes that animal product.

V.2.b Depth

Depth measures the extent of impact upon animal welfare the average person reached by an intervention will have. This is expressed as the product of multiple terms: the percent of people contacted that end up reducing animal product use (this is to be broadly construed as any lifestyle change that reduces a person’s negative impact on animal welfare, either by completely abstaining from the use of certain animal products or by switching to more humane animal use infrastructures or taking other actions that reduce animal suffering such as volunteering or donating to reduce animal suffering); the amount of time a person who reduces animal product use continues that reduction; the number of animals affected by the animal product use reduction; and the improvement in welfare for each of those animals affected. Each of these factors will here be defined.

The percent of people reducing animal product use is the fraction of people reached by an intervention that make certain measurable changes to their lifestyles (whether voluntarily, as in a dietary change, or involuntarily, as in consuming animals that are raised more humanely under new laws). There may be multiple categories of animal use reduction, with different accompanying conversion rates. For example, from the group of all people reached by an intervention, it might be found that 1% converted to veganism, 6% lowered their overall chicken consumption by a fifth, and 10% will only purchase cage-free eggs.

The amount of time a given lifestyle is adopted is defined as the length of time before a person adopting a new lifestyle is expected to experience recidivism back to their old lifestyle or back to a less humane lifestyle. This time may be different for each type of lifestyle change. The net good done by any given convert is proportional to the amount of time they stick with their newly adopted habits.

The number of animals affected by animal product use reduction measures the number of animals that are expected to have their welfare improved by a lifestyle change made by a person reached by an intervention. Again, this number might be different for each type of lifestyle change. This compensates for the fact that reducing, for example, an average omnivore’s beef intake has a much weaker effect than reducing the same person’s chicken intake, given the relative values of pounds of food produced per animal for cows versus chickens.

Finally, improvement in welfare per animal affected quantifies the decrease in suffering per animal affected by the campaign. Not all changes to an animal’s life are equal. For example, there needs to be a way to compare and rank the gain associated with a chicken being moved from a battery cage to an open pasture with the gain associated with a less painful slaughter method being used on a pig. This is a complex determination. It takes into consideration the improvement in living conditions for each animal, the mental complexity of each animal (i.e. their ability to experience those living changes as ethically meaningful increases in mental well-being), and the number of animals saved from slaughter.

V.2.b.2 Notes about improvements in welfare

The determination of the variable “improvement in welfare” also requires a decision regarding how to weigh the unit of “lives saved” against “lives improved”. Some animal activists take “reduction of suffering” as their bottom line, and value saving a life only so far as this prevents the suffering associated with animal death. Other activists also include the increase in pleasure as a factor in calculating animal welfare; they would regard lives saved as valuable both for the avoidance of the pain of death and for the gain of life’s pleasures available to an animal spared from death. FInally, some animal activists find intrinsic value in preventing animal deaths, regardless of suffering averted or pleasure increased.

The determination of suffering reduced requires some sort of an understanding of animal neurology, as discussed above in the form of “mental complexity.” This is a difficult variable to quantify, but remains incredibly important to the overall assessment, since less sensitive (with respect to the ability to experience pain) animals should be given less ethical weight compared to more sensitive animals. It is our hope that our ability to rank the welfare of various animals will improve as research into nonhuman animal consciousness progresses. For now, this crucial variable remains a grey area within the analysis.

V.3 Evaluation

Please comment on what has been empirically observed regarding the effect of relevant human lifestyle changes on animal welfare and on what has yet to be empirically determined.

We have empirical evidence that some ad-viewers click on the ads, and that some ad-clickers also sign up with their email address, usually as a pledge to go vegetarian. We lack empirical evidence that a sizeable number of viewers actually reduce their consumption of animal products, but assuming that gap is bridged, we have a good understanding from economics that reduced consumption reduces the number of animals raised, even though that effect cannot be directly observed.

Evaluation for V.1.a

What upfront and maintenance costs does the intervention have, including costs of supplies, salaries, and office space? How consistent are these costs between previous campaigns?
There can be upfront costs in the form of creating materials, but it’s relatively easy to use the materials of other groups. The ads have costs of their own and require staff time to monitor the campaigns and continue optimizing demographics and ad content.

Evaluation for V.1.b

How many staff and volunteers are needed for the intervention? Do staff or volunteers require special skills and expertise? If the skill and expertise of staff and volunteers is not fully reflected by the cost of paying and training them, consider including the gap (the opportunity cost of shifting them to this project from any other) in the costs of the campaign accounted for in the calculation sheet.

An organization can choose to put more or less staff time into an online ads program. Usually it can be done with less than one full-time equivalent. No unusual training costs are required.

Evaluation for V.1.c

How long does the pre-campaign phase of the intervention last? How much time is committed during the intervention? How consistent are these times between different past campaigns? (Be sure to include all the time the campaign takes, including pre-campaign activities, when calculating the costs of the intervention.)
It can take some time, perhaps weeks or months, to set up an online ads program, depending on whether new materials are created. Running the online ads has no fixed time period.

Evaluation for V.1.d

What expenses have occurred for some past campaigns but not others? What other emergency or unplanned expenses might a campaign implementing this intervention incur? How likely is each one? Which of these expenses would be one-time expenses, and which might happen repeatedly over the course of the campaign (or become more likely the longer the campaign continues)?
Staff time varies, as well as the costs of the ads. It also costs to produce new materials.

Evaluation for V.2.a

How many people are reached directly by a typical campaign? How many people change their behavior due to the campaign without being aware of this? How much does the breadth of campaigns of this type vary?

Ads are usually purchased in quantities of “impressions,” which measures how many people are shown the ads. People might not be aware of the effects of the ads, perhaps because the ads prime them for another exposure to pro-veg information, which then inspires a conscious change, or because the ads lead them to make decisions that seem to be based on other criteria, such as finding animal products less appetizing in the future.

Evaluation for V.2.b

What changes do people reached by the campaign make that will affect animals? What percentage of people reached by the campaign make each of these types of change? How much does this vary between different campaigns implementing this intervention? How well-documented and understood are the changes that people make in response to the campaign? For each change, how long do people typically retain the change? How deep of an impact does this make upon animal welfare? How many animals are affected, and how do their conditions change? Again, how well understood are these effects?

The primary change people can make because of the campaign is a reduction in their consumption of animal products. It’s unclear what percentage make this change and how long they make it for. It likely varies quite a bit between ad programs, but it’s unclear how much. These changes themselves are not well-documented, but some indications are, such as email sign-ups. The impact of dietary change is substantial. The average American consumes approximately 30 animal equivalents per year (excluding wild-caught fish and shellfish). Although we discount this figure when estimating the number of animals spared entirely of suffering, the number is still large. Again, we have empirical evidence that some ad-viewers click on the ads, and that some ad-clickers also sign up with their email address, usually as a pledge to go vegetarian. We lack empirical evidence that a sizeable number of viewers actually reduce their consumption of animal products, but assuming that gap is bridged, we have a good understanding from economics that reduced consumption reduces the number of animals raised, even though that effect cannot be directly observed.

Evaluation for V.2.b.2

Does this intervention result in animal lives spared or in animal lives improved, or both? Is there exceptionally strong or weak evidence that the animal experiences the results of the intervention as a reduction of suffering or increase of pleasure?
The primary outcome is animals spared by dietary change. Assuming this occurs, it is fairly clear that it reduces net suffering.

V.3.1 Calculating efficiency

We offer a way to calculate as objectively as possible the efficiency for a given campaign measured in the units of animal suffering or animal deaths avoided per US dollar equivalent expended.

General Information
Description	Value	Notes
Unit of intervention length	Clicks	How the length of the intervention is measured. E.g. “days”, “weeks”, “number of events”
Estimated length of intervention, in intervention units (a1)	100	Use unit defined in a1

Expenditures (unit=US$)	Pessimistic (Highest) Estimate	Realistic Estimate	Optimistic (Lowest) Estimate
	Based on highest costs we’ve seen for an ads program (when the program has just started)	Based on average cost we’ve seen for ads programs	Based on lowest costs we’ve seen for an ads program
[simple_tooltip content=’Ad space purchased from the advertising platform, e.g. Facebook’]Cost of materials[/simple_tooltip]	$5.00	$3.00	$1.00
[simple_tooltip content=’For pessimistic, assuming it takes 40 hours of work with an annual staff cost of $40,000 a year to recruit and train someone to get 500,000 clicks. For others, assuming program has already been running so recruitment/training costs are negligible.’]Cost of recruitment and training[/simple_tooltip]	$0.16	$0.00	$0.00
[simple_tooltip content=’Calculated from hours worked per number of ads provided by advertising organizations and wage estimates for average nonprofits.’]Cost of Personnel[/simple_tooltip]	$1.03	$0.12	$0.05
Results
[simple_tooltip content=’For pessimistic/optimistic totals, see Guesstimate model’]Total Expenditures[/simple_tooltip]	$5.66	$3.12	$1.45
[simple_tooltip content=’The unit by which the results of an intervention are measured.’]Unit of suffering[/simple_tooltip]	Equivalent animals spared / Years of suffering spared
	Pessimistic (Lowest) Estimate	Realistic Estimate	Optimistic (Highest) Estimate
	Based on the most pessimistic staff opinions after considering the limited studies done, intuition, psychology research, etc.	Based on the average staff opinion after considering the limited studies done, intuition, psychology research, etc.	Based on the most optimistic staff opinions after considering the limited studies done, intuition, psychology research, etc.
[simple_tooltip content=’Lowest, average, and highest estimates based on data given to us from charities. Reached by comparing CPM (cost per thousand impressions) to cost per click for different organizations. Used to inform “number of vegetarian equivalents.’]Number of impressions (~non-clicking viewers)[/simple_tooltip]	2,631.58	7,272.73	10,526.32
[simple_tooltip content=’Lowest, average, and highest estimates based on data given to us from charities. Reached by comparing cost per click to cost per pledge for different organizations. Used to inform “number of vegetarian equivalents.”‘]Number of vegetarian pledges / email sign-ups[/simple_tooltip]	0.64	18.48	33.66
[simple_tooltip content=’This estimate has been the result of a substantial amount of research and discussion, but remains quite subjective. Please see the table on Row 44 for some examples of relevant evidence.’]Number of vegetarian equivalents[/simple_tooltip]	0.20	0.70	1.80
[simple_tooltip content=’The choice of this figure is abritrary as long as the “number of vegetarian equivalents” takes it into consideration.’]Average ‘lifespan’ of a vegetarian[/simple_tooltip]	7.03	7.03	7.03
[simple_tooltip content=’See table below “Animals spared / Years spared per year of vegetarianism”, using figures from the ACE Impact Calculator, excluding wild fish and shellfish.’]Animals spared per year of vegetarianism[/simple_tooltip]	4.99	8.68	12.28
[simple_tooltip content=’See table below “Animals spared / Years spared per year of vegetarianism”, using figures from the ACE Impact Calculator, excluding wild fish and shellfish.’]Years spared per year of vegetarianism[/simple_tooltip]	1.28	2.05	2.89
[simple_tooltip content=’For pessimistic/optimistic totals, see Guesstimate model.’]Results totaled (animals)[/simple_tooltip]	10.60	42.73	135.54
Results Total (years)	2.52	10.10	31.85

Animals spared / Years spared per year of vegetarianism
	[simple_tooltip content=’Pessimistic’]Pess.[/simple_tooltip]	[simple_tooltip content=’Realistic’]Real.[/simple_tooltip]	[simple_tooltip content=’Optimistic’]Opt.[/simple_tooltip]	Elasticity (pess.)	Elasticity (real.)	Elasticity (opt.)	Pess.	Lifespan (years, real.)	Opt.
Cow – beef	0.11	0.118	0.126	0.45	0.67	0.89		1.25
Pig	0.34	0.37	0.41	0.57	0.76	0.96		0.5
Chicken – broiler	24.3	24.7	25	0.34	0.63	0.91		0.123
Turkey	0.75	0.78	0.81	0.01	0.28	0.55		0.32
Farmed fish	2.07	2.92	3.72	0.16	0.43	0.71		1.5
Discount rate for country differences	0.5	0.5	0.5	[simple_tooltip content=’Online ads seem to be most cost-effective in Latin America, so these figures are discounted based on Latin Americas consumption rates relative to the United States: Global Poultry Trends 2014 List of Countries by Meat Consumption’]Notes[/simple_tooltip]
Total animals spared	5.1	8.68	12	[simple_tooltip content=’For pessimistic/optimistic totals, see Guesstimate model.’]Notes[/simple_tooltip]		Total years spared	1.3	2.05	2.9

Final Total: the proposed intervention has a calculated efficiency of Total Years or Lives Spared / Total Expenditures, for a campaign of 100 clicks, with results being measured in the unit of lives or years spared.
	Pessimistic	Realistic	Optimistic
Lives	2.80 lives/dollar	13.68 lives/dollar	54.78 lives/dollar
Years	0.63 years/dollar	3.23 years/dollar	12.95 years/dollar

Evidence that affects our estimate of number of vegetarian equivalents per 100 clicks
Description of evidence	Pessimistic	Realistic	Optimistic	Notes
Estimate from general psychology/sociology, “intuition”	0.001	1	5	Based on the general, hard-to-quantify experiences of ACE Researchers and other well-informed parties.
Cost per vote from political grassroots advocacy research	0.1	0.5	1	Based on data from Get Out The Vote, 2nd Edition by Donald P. Green and Alan S. Gerber. Speculative adjustment for factors such as (i) vegetarianism being a bigger ask than voting, (ii) most registered voters already planning to vote, (iii) online ads undergoing more testing/optimization for messaging based on intermediate outcomes, (iv) fewer people being familiar with the basic arguments for vegetarianism.
Intermediate outcomes such as % pledging to go vegetarian	2	5	10	Based on information provided by charities, combined with speculation of how many pledgers actually go vegetarian
Low-powered studies done on leafleting and online ads	-10	-3	3	Most heavily weighted by the MFA online ads study

V.4 Calculation Sheet Discussion

A few points on the sheet:

• To express the magnitude of animal suffering avoided within a formula, it is necessary to convert to a common unit. It is difficult to express this as an objectively defined value, so the suggested way to measure the results of a campaign is to list the types and number of animals affected, and the specific ways in which their suffering is expected to be reduced. Ideally, the various forms of suffering prevented by a campaign could all be converted to a common unit. The unit for “results” (i.e. measurement of campaign success) has been defined here as years of factory-farm level suffering (i.e. the amount of suffering experienced across one year on a factory farm) averted. This is a crude unit, as it does not allow for distinctions to be made between different types of animals (with different types of consciousness), and makes it hard to deal with suffering that happens outside of the factory farm industry. It should also be noted that the calculations in the “lifestyle multiplier” chart within the calculation sheet only takes into consideration farmed fish, not wild-caught fish. If an alternative unit is desired, one “YBHS” (“years of battery hen suffering”) could be defined as the negative utility produced by keeping an average hen in a battery cage for one year. A year of non-battery conditions for hens could be estimated as a factor of the YBHS (e.g. it has been suggested in the Wageningen Journal of Life Sciences (Del Mol, et al., 2006) that cage-free facilities reduce the suffering of an egg-laying hen by half; thus, a year of cage-free hen suffering prevented would be equivalent to only 0.5 YBHS avoided). Other non-hen animals could have their yearly suffering also calculated as a factor of the YBHS, after taking into consideration both the nature of their environment and the capacity for the animals to feel pain relative to that of a hen.This is perhaps a more controversial method of calculation, as it requires making estimations about the relative abilities of different animals to experience suffering. While it is clear that the expected suffering of, say, a gestation crate-confined pig is many orders of magnitude stronger than that of a fly caged in a jar, it is not as clear that the same pig should be given greater weight than a battery-caged hen. However, estimations must be made for the purposes of this calculation. To count all animals as equal is itself an estimation, so the decision cannot be easily avoided.
• The expenditures/results multiplier sections include somewhat arbitrary (though intuitively reasonable) coefficient values. This is an attempt to quantify non-directly-calculable variables, in order to adjust the final values accordingly. As more research is done and more information is known about specific campaigns, these coefficients could change dramatically.
• The “total expenditures” and “total results” boxes (f1 and m1) give values specific to a campaign of the length described in a2. It is theoretically scalable to any campaign length, though unless upfront costs scale with campaign length, accuracy will be improved by revising the calculation to address the new length directly.
• The final output is given in terms of unit of animal welfare increase per cost in US dollars.
• The “lifestyle multiplier” key refers to the number of years of factory-farm level animal suffering are expected to be averted due to a person beginning to adopt a certain lifestyle (See I.2.b for more details). This is calculated by: [mean years person is expected to maintain lifestyle] * [number of years of animal suffering avoided per year of lifestyle].

V.5 Error tracking for “Efficiency”

• How reliable is the scientific data on the causal relationship between campaigns and behavioral changes?
  • We can directly observe the effects on vegetarian pledges, ad-clicks, and other intermediate evidence. There is no robust data on the impact on dietary change, although a few studies have been attempted and provide weak evidence.
• How reliable is the scientific data on the causal relationship between behavioral changes and improvements in animal welfare?
  • We estimate the impact of dietary change on animal welfare in the Impact Calculator on our website. Our certainty regarding whether changes in diet lead to improvements for animals is much greater than our certainty for whether the dietary changes occur, and to what degree.
• How much direct knowledge exists about the costs and outcomes from past campaigns?
  • We have substantial knowledge of the costs of these campaigns, and a significant amount of knowledge of the immediate outcomes.
• How reliable are budget reports from past, similar campaigns? Could budgets mistakenly fail to list some costs and/or record unrelated expenses as campaign costs? Could some budgets fail to temporally synchronize expenditures with results?
  • The budget information seems very reliable relative to the outcome information, although there are some difficulties in, for example, assessing future campaign costs based on previous campaign costs.
• Has there been significant economic or social change since the time of past, similar campaigns that might invalidate their quantification of costs/outcomes? What other changes have happened (such as with supply chains for campaign resources) since past, similar campaigns?
  • There haven’t been noticeable changes, although trends in animal product consumption and social change could affect effectiveness.
• How reliable are estimates about things like the number of people expected to attend a campaign event, the number of campaign events to be held within a given time span, etc.?
  • N/A

Summary of the Evaluation

Activism Philosophy	Activism Approach
Online ads are usually an incremental, consumer change approach. They fit into a model of social change driven by consumption, but don’t specify an abolitionist or welfarist position except insofar as they assume reducing the number of farmed animals is a good outcome.	Internet users are shown ads, usually in their Facebook feed, that invite users to click and be directed to a landing page. On that page, pro-veg information is displayed, often in the form of a video showing footage from undercover investigations.

Target Demographics	Target Animal(s)	Size of Campaign	Time Delay between Intervention and Results
Targeted at consumers, usually Facebook users. Most frequent demographic is young and female.	Usually not targeted at a specific species, but the ad or video may include specific animals.	Can be done with very small amounts of time and money, although some organizations spend large parts of their budget on it.	Dietary change resulting from the ads could be immediate or might happen over time if viewers are more likely to change behavior based on future exposure.

General Summary of Area I (Type).

Online ads are typically done on Facebook by showing images with text in the news feed. Viewers who click on the ads are directed to a landing page (for example, www.meatvideo.com) that typically shows an auto-play video (for example, What Cody Saw, where an undercover investigator narrates his experience on animal farms) and invites visitors to pledge to go vegetarian and enter their email addresses. Those who sign up will receive several weekly emails giving information on how to eat a more plant-based diet. These ads could benefit humans by reducing the size of animal agriculture, an industry that many believe is harmful to public health, the environment, and the global food supply. It is relatively easy and low-risk to run online ads without much organizational capacity, especially if they use the materials of other organizations, such as the video or even the landing page itself. However, with more experience and resources, online ads can be optimized to greatly increase their impact, as measured by conversions from dollars spent to pledges to go vegetarian (or take some other action on the landing page). Online ads can bring in support from other organizations, most commonly in the form of advice or materials to help create the ads. The intervention has been growing in prevalence in the past few years as organizations have begun to see it as a cost-effective form of veg outreach and social media has become a more popular advocacy platform. In theory, online ads can also inspire new activists and donors that can benefit the movement, but we’re not sure if this happens much in practice. Finally, insofar as online ads increase demand for animal-free foods, this benefits organizations that sell and promote these products.

Code	Question an (R) in the Score column indicates that a score of 1 would be generally considered preferable to a score of 7	Score (1-7)
II.1.a	How many times has a similar intervention been attempted before?	5
	Has a similar intervention ever been attempted in the same context as the one being evaluated?	6
II.1.b	How widely have previous campaigns varied in total results?	5
	How widely have previous campaigns varied in efficiency?	4
	How well have previous predictions of results matched actual results?	5
	How well have previous predictions of costs matched actual costs?	7
II.2	How many different avenues of success does the campaign have available?	2
	How many intermediate positive outcomes does the campaign have?	7
II.3.a	How well understood are the neurological abilities of the primary animal group targeted by the campaign?	6
	How well understood are the neurological abilities of any other animal groups targeted by the campaign?	6
II.3.b	How well understood are individual and social group responses to this type of campaign?	2
	How well understood are economic factors affecting the outcome of this campaign?	6
II.3.c	How well understood is the primary ecological impact of this intervention?	3
	How well understood are any other ecological impacts?	1
II.4.a	How many known uncertainties affect the outcome of the campaign?	2(R)
	How strongly do known uncertainties affect the outcome of the campaign?	5(R)
II.4.b	How many steps separate the actions of the campaign from a change in animal welfare?	3(R)
	How many different actors are involved in the chain of events between the success of the campaign and the change in welfare?	3(R)

General Summary of Area II (Certainty of Success)

It seems clear that spending on online ads results in relatively large numbers of people being exposed to advocacy for veg eating, and that any dietary change will result, in expectation, in a large benefit for animals. We are very uncertain about how much dietary change, and other impact, occurs due to the online ads. We know that a substantial number of people who click on ads then pledge to go vegetarian or reduce their consumption of animal products, and some even contact the advertising organizations to ask questions about veg eating and say personally that they are changing their consumption. It is also unclear what impacts, if any, online ads have on the broader state of affairs for animal agriculture and how animals are viewed in society. We know of little, if any, historical precedent that shows major social change, such as a 20%+ reduction in the number of animals raised for food, is caused by individual advocacy and changes in personal consumption. This means that, although online ads might cause an immediate reduction in the demand for animal products, they might not have as much impact on animal welfare in the long-run as other interventions.

Code	Question an (R) in the Score column indicates that a score of 1 would be generally considered preferable to a score of 7	Score (1-7)
III.1	How many specialized skill sets are required?	2(R)
	How much expertise is required in the most demanding area?	2(R)
III.2.a	How demanding is the typical workload associated with the campaign?	2(R)
III.2.b	How many hours per week does the typical staff member work on the campaign when it is making its minimum demands?	3(R)
	How many hours per week does the typical staff member work on the campaign when it is making its maximum demands?	4(R)
III.2.c	How many staff members are involved with the campaign when it is ongoing but at the time when fewest people are working on it?	1(R)
	How many staff members are involved with the campaign at the time when the most people are working on it?	3(R)
III.3	How many rare or difficult-to-obtain materials does the campaign require?	1(R)
	How severe is the difficulty of obtaining the hardest-to-obtain material needed for the campaign?	1(R)

General Summary of Area III (Barriers to Entry)

Running an online ads program involves getting familiar with an advertising platform, such as Facebook; creating or using ads from another organization, including a landing page; then purchasing the ad space and running them. Usually, organizations monitor the ads and continuously try new ones. This usually takes one part-time staff member to implement, but more work can be required, especially when creating new content such as videos.

Code	Question an (R) in the Score column indicates that a score of 1 would be generally considered preferable to a score of 7	Score (1-7)
IV.1.a	How large a chain effect does the intervention likely have?	2
IV.1.b	How likely is the campaign to inspire other victories?	2
IV.1.c	How significant a change in public mindset is this campaign likely to cause?	2/3
IV.2.a	How much alienation is this campaign likely to cause?	2(R)
IV.2.b	How high is the risk that the campaign may be more harmful than beneficial?	4(R)

General Summary of Area IV (Expected Indirect Effects)

Online ads seem to have a limited effect outside of the behavior of viewers. For example, ads don’t receive press attention, and it seems like they have less of a community-building impact than other interventions. However, given the large numbers of people exposed to ads, they could indirectly affect the general attitudes and behaviors of consumers. They are unlikely to alienate or anger their audience, although they could have negative effects by shifting the public perception of veg eating and farmed animal suffering towards viewing it as a matter of personal choice.

Final Total: the proposed intervention has a calculated efficiency of (_______), for a campaign of 100 ad views, with results being measured in equivalent animals spared or equivalent years of farmed captivity averted.

	Pessimistic	Realistic	Optimistic
Lives	2.80 lives/dollar	13.68 lives/dollar	54.78 lives/dollar
Years	0.63 years/dollar	3.23 years/dollar	12.95 years/dollar

Were the optional modifiers used in calculating this total?

No

Final Determination (Overall Summary and Recommendations)

Overall we think ads seem to have limited long-term and indirect benefits, which is likely where a majority of the potential impact lies in animal advocacy. We think the short, impersonal interactions of online ads programs have less of an effect on, e.g. creating new activists, than other forms of outreach. Also concerning are the potential negative impacts of online ads, at least in their current form, such as shifting attitudes towards veg eating as a personal choice rather than a social imperative. We currently don’t recommend that organizations create new online ads programs or expand existing programs, at least when that funding could be used for more promising interventions such as corporate outreach and undercover investigations. We feel that the main upside of online ads, reaching a large number of people at very low costs, is outweighed by the concerns of low and uncertain per-viewer impact and the limited long term and indirect benefits. Even if we focus entirely on short-term impact, corporate outreach still seems more promising. However, if corporate outreach becomes more difficult in the future, e.g. once cage-free reforms are fully implemented, then online ads or other forms of dietary change outreach might be justified as the most cost-effective intervention for short term impact. We think that some ads programs, such as those that run VSG ads or have spent significant time minimizing their cost-per-pledge, tend to be more promising than others.

Related Studies

2013 Analyzing The Hidden Face of Food

2013 Comparing the Effectiveness of Videos and Ads

2013 The Humane League/Farm Sanctuary Video vs. Ad Study

Archived Evaluations

2013 Online Ads Intervention Evaluation