Progress or Projection? The 63% No-Kill Number Doesn’t Add Up

Best Friends’ 63% claim rests on selective data, machine-learning estimates, and a sample that may represent fewer than 9% of U.S. brick-and-mortar shelters.

Having spent nearly three decades directing animal shelters-from small-town kennels to the sprawling complexes of Los Angeles and New York City-I know that rigorous, transparent data is the lifeblood of responsible animal welfare. So when Best Friends Animal Society (BFAS) announced that 63% of U.S. shelters have achieved “no-kill” status, I approached the claim not with skepticism or applause, but with professional concern.

That headline figure, it turns out, is not drawn from the full 4,915 brick-and-mortar shelters tracked by Shelter Animals Count (SAC), the nation’s largest independent animal welfare database. Instead, BFAS’ celebrated statistic is built on a proprietary, largely undisclosed sample of just 688 shelters-about 14% of the national total. And, since only 63% of those 688 shelters are classified as “no-kill,” the headline percentage is actually based on less than 9% of U.S. brick-and-mortar shelters.

This gap between headline and methodology deserves examination, not to undermine progress, but to ensure every community has reliable benchmarks to guide lifesaving work.

The Data Framework

BFAS’ 2024 Shelter Pet Lifesaving Report rests on two pillars: a public-facing Pet Lifesaving Dashboard and a closed-door Shelter Pets Data Analysis (SPDA).

• Public Dashboard: BFAS advertises “key insights and analytics from more than 10,000 shelters and rescue groups,” but offers no way for users to distinguish actual outcomes from modeled projections.

• Proprietary Sample: BFAS’ claim that 63% of U.S. shelters are no-kill is based on outcome data from just 688 brick-and-mortar facilities. By comparison, Shelter Animals Count—the most comprehensive national dataset—tracks over 4,900 brick-and-mortar shelters. That means the 63% headline reflects roughly 14% of U.S. shelters, and since only 63% of those qualify as no-kill, the actual share of no-kill shelters nationwide is under 9%.

  The proportion drops further when excluding shelters classified through machine-learning projections rather than direct reporting—but by how much, we don’t know. BFAS does not disclose how many of its 688 reporting shelters met the full 12-month threshold. While headlines suggest a national transformation, the underlying data points to a much narrower—and far less representative—story.

• Predictive Modeling: For any shelter lacking 12 consecutive months of recent data, BFAS uses a proprietary machine-learning model-trained on intake/outcome records and county-level demographics-to generate “real-time projections” of no-kill status. Unlike SAC’s peer-reviewed estimation model, SPDA’s algorithms have not undergone third-party audit, leaving communities unable to verify the 63% headline.

Bottom line: When 63% of 14% equals just 9%, who’s really being counted-and who’s being left out?

Core Concerns

1. Selection Bias

SPDA’s 688-shelter dataset underrepresents regions with persistently high euthanasia rates-especially parts of the South and large states like California and Texas, which together account for nearly half of all shelter killings nationwide. Voluntary reporting and selection criteria skew the sample toward higher-performing, better-resourced shelters.

2. Opaque Algorithms

Despite claims of comprehensiveness, SPDA’s inclusion rules, model weighting, and training methods have never been subject to independent audit. Communities see dashboard outputs but cannot verify how the 63% figure is calculated.

3. Advocacy-Funding Overlap

BFAS acts as both funder and advocate, partnering closely with many SPDA shelters. This dual role introduces potential bias, as organizations with strong BFAS ties are more likely to report data and be counted among the “no-kill” ranks.

BFAS’ Defense

In response to concerns about data sampling and transparency, Best Friends has defended its approach as both strategic and grounded in outcomes. In media interviews and public materials, CEO Julie Castle stated:

“Our focus is actionable data. By working directly with shelters committed to no-kill benchmarks, we allocate resources more effectively. Predictive models are validated annually against ground-truth outcomes and adjusted for regional disparities.”

BFAS highlights successes such as Rhode Island achieving no-kill state status in 2024 and a 10.5% reduction in feline euthanasia through community cat programs. But a 2024 court ruling in San Diego found that releasing domesticated cats without verifying a caretaker constitutes unlawful abandonment—a legal challenge that may undermine aspects of BFAS’ deflection-based strategies for reducing shelter feline intake.

Ethical & Practical Impacts

Critics warn that when modeling and messaging outpace ground-level realities, even well-intentioned strategies can distort resource allocation, policy, and public trust. The following impacts illustrate how even well-intentioned data strategies can distort resource allocation, policy formation, and public trust.

• Funding Misalignment: Chasing an erroneous national “63% no-kill” headline risks channeling grants and support toward already high-performing shelters, leaving under-resourced communities-where euthanasia remains high-further behind.

• Preventive Care Sidelines: Some observers and critics note that BFAS began shifting funding away from large-scale spay/neuter programs as early as 2017, with this trend accelerating after 2020. This deprioritization has coincided with an 8% annual rise in overpopulation in Southern shelters, as tracked by SAC and regional intake data. Spay/neuter remains the most effective tool for reducing shelter intake and euthanasia; sidelining prevention risks worsening the very crisis shelters are trying to solve.

• Legal Exposure: The San Diego ruling underscores a growing legal risk for shelters that prioritize intake deflection strategies—like unmanaged community release—over documented care. As courts begin scrutinizing these practices under abandonment laws, shelters across the country may face mounting legal and ethical pressure to reassess what qualifies as responsible lifesaving.

• Community Trust: When projections and public messaging - such as the example below - replace real data, shelters and communities cannot trust the benchmarks used to secure local support and policy changes, undermining accountability and public confidence.

Best Friends CEO Julie Castle declares a 'tipping point' for no-kill — based on a sample representing just 14% of U.S. shelters. When marketing overtakes measurement, who is really being counted and for what end?

The 90% Benchmark: A Double-Edged Sword

The widespread adoption of a rigid 90% live release rate as the definition of “no-kill” can create perverse incentives. As communities succeed in reducing intake of healthy, adoptable animals, the proportion of animals with severe medical or behavioral challenges naturally rises - making the 90% threshold increasingly unrealistic for responsible, open-admission shelters.

This can pressure organizations to manipulate statistics or adopt risky policies, such as mass cross-country transports or “community sheltering,” where animals are moved or released without adequate tracking or care. These practices may improve reported outcomes but can obscure real welfare risks and undermine accountability. Even the most nuanced performance metrics lose their value without transparent reporting and clear public interpretation.

Transparency in Reporting

Dashboards must clearly differentiate between actual outcomes-based on 12 consecutive months of real data-and estimated projections used to fill data gaps. A live release rate below 90% should not be seen as failure if it reflects a community’s commitment to taking in more high-needs animals, such as those with serious medical or behavioral issues. Documented increases in complex intakes can signal real progress, while unexplained declines in live release rates warrant concern.

Without rigorous transparency and external oversight, there is a real risk of shelters misrepresenting their performance, whether intentionally or through selective reporting practices. Full disclosure and clear reporting are essential for genuine accountability.

Accountability & Next Steps

No-kill campaigns inspire, but they must rest on verifiable methods. Without transparency, the movement risks trading real progress for the illusion of it.

Stakeholders must call for:

1. Third-party audits of SPDA’s algorithms and sampling rules.

2. Inclusive reporting that brings high-need regions into the dataset, even if outcomes must be flagged as provisional.

3. Clear delineation on the dashboard between actual vs. estimated shelter statuses.

Given BFAS’s influential role in the no-kill movement and its data collection capabilities, it is uniquely positioned to lead by example. By providing transparent, verifiable data and clearly distinguishing between reported figures, projected estimates, and intake ratios, BFAS can set a new standard for integrity in shelter reporting. Such leadership would not only enhance the credibility of the no-kill initiative but also ensure that progress is both genuine and sustainable.

Conclusion

BFAS’ marketing has reshaped the national dialogue on shelter lifesaving, yet its most lauded metric depends on invisible mechanics-machine-learning projections and undisclosed selection criteria. As the 2025 no-kill deadline nears, communities need benchmarks they can trust and reproduce in their own backyards. Only then can “no-kill” transcend a slogan and become a reality for every shelter, in every state.

In the end, the future of the no-kill movement will be shaped not by bold promises, but by honest, transparent reporting.

Ed Boks is a former Executive Director of the New York City, City of Los Angeles, and Maricopa County Animal Care & Control Departments, and a former Board Director of the National Animal Control Association. His work has been published in the LA Times, New York Times, Newsweek, Real Clear Policy, Sentient Media, and now on Animal Politics with Ed Boks.

Stay Informed

For more analysis and updates on the evolving landscape of animal welfare policy, visit Animal Politics with Ed Boks.