Part 1: The Results Are In

On May 29, Pro-Animal Future announced two ballot initiative campaigns for the 2026 election cycle. The measures in Denver, CO and Washington, DC seek to end force-feeding of farmed animals by banning force-fed products like foie gras. We will pursue similar laws through the city councils in the Portland, OR metro area this year. New chapters are emerging in yet more states, and we expect to announce additional campaigns early next year. In this post, PAF strategy lead Aidan Kankyoku discusses the background of these campaigns and how they fit into PAF’s 10-year strategy to build towards state-level campaigns directly challenging factory farming across the country. 

Key Points

• New campaigns launching: Pro-Animal Future has announced 2026 ballot initiatives in Denver, CO and Washington, DC to ban force-fed products, with a similar campaign in Portland, OR through the city council.
• 2024 results: Denver’s fur ban (42% support) and slaughterhouse ban (36% support) both failed, but provided invaluable lessons and a foundation to build on.
• Polling works: Professional polling accurately predicted final results within 1% – 18 months before the election and before any serious campaigning began from either side.
• Campaign impact follows a “logistic curve”: Most campaign spending has minimal effect until one side reaches a critical threshold.
• Opposition spent $2.5M but didn’t move the needle: Despite every advantage, the factory farm industry failed to persuade any voters to defect compared to pre-campaign polling, suggesting vigorous pro-animal campaigns can prevent well-funded opponents from reaching their critical threshold.
• Institutional power matters: Elected officials, media, and party endorsements significantly influence outcomes, requiring more strategic relationship-building. We neglected this last year in favor of focusing on direct voter outreach, and that was a mistake.
• Bold campaigns can “fail forward”: Even unsuccessful abolitionist measures advance long-term goals by shifting narratives and building momentum.
• Refocusing campaign efforts: Understanding these dynamics helps optimize resource allocation between voter outreach, institutional engagement, and message saturation in future campaigns.

A Brief History

In case you’re new to PAF, I’ll start by sharing a bit of context before I launch into discussing key lessons. Pro-Animal Future was launched in April 2023 as a direct outgrowth of research conducted by Pax Fauna. Our purpose at Pax Fauna was always to design a new strategy for the grassroots animal freedom movement using research, and PAF was the manifestation of that. PAF applied this research to two citizen initiatives that appeared on the ballot in Denver, CO in November 2024. The measures were a ban on fur sales and a ban on slaughterhouses. The latter would have shuttered the largest industrial lamb slaughterhouse in the US. However, after a costly opposition campaign led by factory farm industry groups, both measures fell short of the majority needed to become law. 

Despite the final vote tally, the campaigns were successful in a number of important ways. One goal was to start transforming animal rights from a niche consumer choice issue to a mainstream political issue. Pax Fauna’s research found that the public is most hostile to farmed animal advocacy when it is framed in terms of consumer choice (vegetarianism/veganism), and is much more receptive when presented with opportunities to act collectively through the political process. In this regard, the slaughterhouse ban acted as a proof of concept: 36% of Denverites voted to outright ban slaughterhouses. That’s more than 100,000 voters, the vast majority of whom regularly eat meat. Our exit polling suggests perhaps another 100,000 found themselves wrestling with the issue and came down on the “no” side more because of deceptive opposition talking points about the particular facility in Denver rather than due to a rejection of the overall cause of animals in the food system. Indeed, industrial animal farming was far and away the top local political issue in Denver in 2024, with over 80 local news stories in the months leading up to the election. 

An even more important goal of these campaigns was practice. This was PAF’s first go at running a ballot measure, and while the animal advocacy movement has used ballot initiatives before, these measures differed from past campaigns in important ways. Past ballot initiatives for farmed animals have pursued incremental welfare reforms, mainly banning the use of battery cages for hens and gestation crates for sows. These measures, most notably California’s Prop 12, passed by wide margins in both red and blue states. We wanted to test the limits of public support with measures that went further. We also wanted to experiment with campaign strategies that relied less on large sums of money and made better use of our movement’s relative advantage in people power. The Prop 12 campaign cost more than $12 million, an enormous share of the global farmed animal advocacy movement’s relatively tiny collective budget (somewhere between $200 and $300 million per year– .01% of worldwide charitable spending.) If our movement is going to scale up this strategy, we need to find ways to win with less money. We chose the slaughterhouse ban to deliberately challenge ourselves, forcing us to give our all knowing that’s how we’d all learn the most.

Lessons: The First Story

And boy, did we learn. A few months after the election, I released a writeup of some of our most hard-won lessons from 2024. I won’t rehash it all here, but the summary is that we sorely underestimated the importance of institutions and traditional centers of power. We originally thought of ballot initiatives as a form of direct democracy enabling us to bypass the corrupt traditional political process and go straight to the voters. While that is accurate de jure, I wrote that political power centers (especially elected officials, media companies, and the major political parties) de facto have numerous ways of shaping the course of a ballot measure campaign. I presented a story of our loss that read something like this: while we spent all our volunteer time going directly to voters, our opponents spent their money schmoozing with elected officials, media executives, and Democratic Party leaders, winning endorsements against us from every notable civic institution in the city. These endorsements turned voters against us and decided the campaign.

This is not a hopeless story– far from it. Now that we understand better how the game is played, we can all use our time much more effectively: a bit less knocking on doors of ordinary voters and a lot more rubbing elbows with politically influential people and groups.

However, while I still believe this story, and still believe we should implement the lessons from it, in this post today I will present a different story of what happened. I’ll be the first to admit I don’t yet know which of these two stories is more true. They could both be equally true. In the long run, that would present us with some interesting puzzles as PAF-style campaigns sprout up in more and more cities and states. Anyway, here goes the alternate story.

The Alternate Story

In the Summer of 2023, almost 18 months before the election, PAF commissioned a series of polls surveying Denver voters on their attitudes towards fur and slaughterhouse bans. There are a number of different ways to collect survey responses, and a complex industry of companies that offer polling as a service. These methods vary widely in cost. Counterintuitively, the cheapest way to get survey responses is to pay the respondents. The most affordable survey firms are platforms where people can sign up to complete surveys for a very small monetary incentive (often less than $1 for a five-minute survey.) On the other end of the spectrum are boutique political polling firms that use proprietary methods to gather well-balanced samples from unpaid respondents. It is generally known in the world of research that unpaid survey responses are usually more reliable than paid ones (if you interested in understanding why, Claude can do a great job explaining.) But we wanted to find out how much more reliable they are. To put it simply, we had sticker shock when we learned that testing two ballot measures in one city with the expensive political polling firms would cost $10,000-$15,000 a go.

We wanted to know whether the expensive pollsters were really worth it. So in 2023 we ran several concurrent surveys each presenting identical questionnaires about fur and slaughterhouse bans to ~500 Denver voters. The results that came back seemed like they couldn’t possibly be anything more than statistical noise. All of the cheap online panel services measured support for both measures somewhere in the 60s. The professional firm was flipped around the middle, measuring 43% support for the fur ban and only 35% for the slaughterhouse ban. Headlines from 2016 about the legendary failure of pollsters to predict Donald Trump’s first ascendance to the presidency echoed in our memory, and without anything solid to grab onto, we largely set these poll results aside.

So it was that on election day, the results hit us in much the way the heavily-foreshadowed climax of a Shakespearian tragedy befalls the witless protagonist: 42% yes for banning fur, 36% for slaughterhouses. Both results were within one percentage point of the pollster’s measurements taken 18 months before the election, before either side began vigorous campaigns that ended up costing over $2.5 million (almost all of that was the bad guys.)

Polling 101

Now hang on for a moment, before you emphatically close the browser tab containing this blog. I know what you’re thinking, especially if you were part of the campaign: “If we knew what the results were going to be before we even started, why did I bust my ass/donate my money for a lost cause?” You wouldn’t be wrong to be frustrated. I value your time and your money as highly as you do, not least because I know you could choose to spend your time/money on a different form of animal advocacy. I am fiercely committed to helping you use your time and money to make the biggest difference for animals. 

So let me break it down, because those numbers actually weren’t as simple as I first suggested. Even setting aside the fact that we were getting garbled signals from the lower-quality polls we ran at the same time, there was ambiguity in how we should interpret the data from our professional pollster. The structure of a ballot initiative poll is:

1. Gauge initial support, showing voters the language that will appear on the ballot and asking them how they would vote. (There’s a “not sure” option.)
2. Show them a battery of messages both in favor and against the measure, and ask them how convincing they find each argument.
3. After messages, ask them again how they plan to vote, again allowing a “not sure” option.

Let’s look at what our Denver poll actually found:

	Pre-message ballot	Post-message reballot
Fur ban	46% Yes | 10% Unsure | 44% no	43% Yes | 9% Unsure | 47% no
Slaughterhouse ban	39% Yes | 17% Unsure | 45% no	35% Yes | 15% Unsure | 50% no

Imagine looking at this data before starting the campaign. Starting with the pre-message numbers, it seems that while it’s going to be an uphill battle, it’s not hopeless. If anything, fur is leaning in our favor, with more “yes” than “no” votes. And for slaughterhouses, we need to win over a large majority of the “unsure” voters, but hey, that’s exactly the kind of campaign we were looking for: the most ambitious, transformative policy we could pass if we ran an effective campaign.

Things admittedly look worse after messaging. Seeing messages from both sides moves voters against both measures. The slaughterhouse ban now needs to win over every single unsure voter, and the fur ban has lost its plurality. But maybe that just means the messages we tested in the poll weren’t the best we could come up with. After all, surely it shouldn’t be that hard to convince somebody that brutally murdering a fox for a stupid ass $5,000 jacket is a bad idea, right?

More importantly, looking at all these possible numbers and seeing that one way of interpreting them accurately predicted the result would be careless at best. Why assume that the post-messaging “yes” votes will be the correct number, that all the “unsure” votes will turn against us? This would be the classic scientific faux pas of choosing your hypothesis after seeing your data. Surely it would be more natural to assume the unsure votes break down along the existing yes:no ratio. And maybe most voters will be so uninformed that the pre-messaging numbers are a better gauge. At the very least, there’s no way we could have known which numbers to rely on in advance.

The Truth Comes Out

Well, I’m sorry to admit, but the preceding three paragraphs are all a lie. Sort of. Because in reality, we had experienced advisors looking at the polling data with us and telling us exactly how to interpret them. In particular, PAF’s dauntless mentor Josh Balk was at the center of all the aforementioned state ballot measures against intensive confinement, culminating in California’s Prop 12 in 2018. Josh told us that in his extensive experience with polling, the number that matters is the “yes” votes after messaging. He assured us that everyone still undecided by the end of the poll would indeed turn against us due to voters’ bias in favor of the status quo.

Why did we ignore Josh? Well, besides all the reasons I’ve already given about mixed polling data, we weren’t ready to give up just because a few people told us to, no matter how experienced they were. We had to find out for ourselves whether we could campaign hard and smart enough to win over the undecided voters. (We still haven’t totally given up on this– more on that to come.)

On top of that, the slaughterhouse campaign was one we were willing to risk losing. With a bold abolitionist campaign like this, we knew that even if the measure didn’t pass the first time, we could be confident all our voter outreach efforts would move the needle in the right direction, perhaps more effectively than any other public opinion strategy we could spend our time and money on instead. Looking back on the campaign, I feel total confidence that this was the right assessment, and I’m not alone in that. At the Pro-Animal Colorado chapter meeting in December when we first started deliberating on what to do next, a large contingent wanted to come right back around and run the slaughterhouse ban again, to keep chiseling away at the industry’s fortress of lies one doomed campaign at a time. There are, after all, successful historical precedents for this strategy of social movements using bold ballot initiatives to fail forward (see women’s suffrage, cannabis legalization.)

Not So Fast

But wait, if the pollster was able to predict the outcome of these campaigns within a single percentage point 18 months before the vote, before either side started campaigning, does that mean everything each side did to campaign was a complete waste of time and money? That all the canvassing and endorsements were meaningless? That campaigns don’t matter at all? It seems too unlikely that the two sides just happened to perfectly cancel each other out.

Whether that conclusion sounds depressing or seems like a relief, we have to rule it out due to one more data point: Measure J, the 2024 ballot measure campaign run by Direct Action Everywhere in Sonoma County to ban factory farms. The first time DxE worked with a professional pollster to test their ballot language in Sonoma, they found 45% support and a large number undecided– better numbers than either of the Denver campaigns. But on election day, Measure J suffered a devastating defeat: 15% in favor, 85% opposed. The Denver slaughterhouse ban looked like a dead heat in comparison.

Clearly, it’s possible for campaigns to change the outcome of a ballot initiative. Yet there’s also strong evidence that the default is for skilled pollsters to be able to predict the outcome before any campaigning takes place. That’s not just based on the prophetic Denver polls– polling and election results from two decades of animal welfare measures (Prop 12 and its predecessors) suggest a similar trend.

How do we reconcile these two observations? If this question fascinates you as much as it did me, you might enjoy this obnoxiously long internal dialogue I wrote out when first attempting to sort out my thoughts. Otherwise, here’s a summary of the conclusions:

Returns on investments in ballot initiative campaigns follow what is called a logistic curve (pictured below). In this case, the x axis is time and money put into campaigning, and the y axis is voters who change their minds. When you first start campaigning, the effect on vote share is minimal, and it stays minimal until one campaign reaches a critical threshold of how densely they have infiltrated their message into the electorate. This triggers an information cascade, a social phenomenon where people start making a certain choice en masse not because they’ve been persuaded through rational argument, but because they receive sufficiently powerful social signals that everyone else is making that choice. At a relatively narrow window around this critical threshold, marginal campaign expenditures yield enormous results, before tapping out the available population of voters whose decision could possibly be influenced. Additional expenditures once again cease having any significant effect.

So does that mean all our campaign efforts in Denver were wasted? Not at all. Remember, relative to polling, we didn’t make a dent in the number of people voting our way. But neither did our opponents. Despite spending over $2 million spreading incredible disinformation, the meat industry failed to persuade even one percent of animal-friendly voters to defect. I have no doubt that this is because the strength and vitality of our campaign stopped them from doing so. 

For our opponents, the force of their campaign is a combination of dollars spent, disparaging news articles published, and politically powerful allies activated. How high is the threshold all of that must reach? It’s a function of a number of factors including the population size and the economic power of the targeted industry in that community. This is why the opposition campaign to Measure J in Sonoma was able to reach that threshold, where the slaughterhouse industry in Denver was not. Both opposition campaigns spent about the same amount of money ($2.5 million), but this similarity is deceiving. Sonoma’s population is half the size of Denver, so those dollars went twice as far. More importantly, while most city council members, local newspapers, and the Democratic Party in Denver endorsed against both pro-animal measures, they didn’t do much beyond that. By contrast, the Sonoma Democratic Party, County Commission, and Press Democrat (local newspaper of record) all made crusading against Measure J their top priority for an entire year. The opposition saturated Sonoma County 10 or 100 times more intensely than in Denver, leaving the critical threshold in their dust.

Ultimately, though, one of the most important factors pushing that threshold up is the strength of the campaign they are fighting against. The more vigorously we campaign, the harder it becomes for our opponents to reach the critical campaign threshold. In close campaigns like the ones we’ll examine in part two of this blog, this will mean the difference between winning and losing. And in bold, boundary-pushing campaigns like the slaughterhouse ban, vigorous campaigning is essential to having the biggest impact we can on the narrative. It may not change the vote share in the short term, but it plants the seeds that will enable us to fail forward and start winning these truly transformative campaigns in the decades to come.

So That’s the Bad News…

In 2024, we worked hard. We ran a campaign that reflected our status as rookies. In the end, we didn’t make a dent on our earliest poll numbers. But we didn’t let the industry win anyone over either, and I haven’t even mentioned the fact that we forced them onto defense. They never tried to defend industrial animal slaughter on ethical grounds– instead, they sowed confusion about the impact the ban would have on the slaughterhouse workers, instrumentalizing the working-class Latino community who ended up being the only neighborhood in Denver where a majority of voters supported the ban. They know we’ve found a crack in their armor. It’s not enough to kill them outright, at least not yet. But it was enough to send them into a panic. Perhaps because they know even better than we do what we can accomplish if we keep at it.

Dear friend, if you want to emphatically close the tab on this blog at this point, I won’t try to stop you. But if you’re willing to hear some good news about how we ended up making sense of all this after last year’s election and deciding how to move forward for 2026, read on to part 2 (coming soon).