Updates to Our Impact Estimation Process

How FWI Currently Helps Fishes

FWI improves the lives of farmed fishes primarily through the Alliance for Responsible Aquaculture (ARA). Within the ARA, farmers commit to reducing stocking densities and improving water quality—two key welfare challenges. Our field teams conduct free water testing and give corrective actions where needed. They then return to assess whether welfare conditions have meaningfully improved.

For fishes affected by reduced stocking density, we already use what we consider to be a robust system: we count fishes as helped only when harvest data confirms a reduction of 20% or more. We believe this threshold is sufficiently rigorous and are not further changing it.

This post focuses instead on the changes we’ve made to how we count impact from water quality improvements.

Figure 1. FWI’s process of helping fishes through water quality improvements.

Our Previous Impact Counting Process

From 2022 through mid-2025, our approach relied heavily on staff discretion. Field staff would nominate improvements they believed significantly benefited fishes, and these nominations were reviewed by our Fish Welfare Experts. Confirmed instances were then added to our public Impact sheet. See the full process in Figure 2.

The following criteria guided decision-making:

• The farmer had implemented the corrective actions we recommended.

• The change in water quality was meaningfully better for fishes, as assessed by staff's informed judgment.

• At least 80% of fishes in the pond were likely to have benefited, based on staff’s informed judgment.

Over time, we found that this system had limitations in consistency and transparency.

Figure 2. The ARA’s current process of counting a water quality improvement as fishes helped.

Why We Made Changes

We identified two key issues with the old system:

1. Inconsistency: Staff varied in how and when they nominated fishes as helped, making our impact numbers harder to interpret or compare across time.

2. Limited transparency: Although individual cases were shared publicly, the reasoning behind each decision was not always sufficiently detailed.

Our Updated Approach

Using two additional years of data, our experts’ opinion, and insights from scientific literature, we’ve adopted a more standardized system:

• Nominations: We now use specific thresholds to determine whether a water quality improvement is eligible for review.

• Standardized Assessment: We classify each improvement’s intensity and duration using standardized categories aligned with the Welfare Footprint Project’s pain intensity scale.

• Transparency: We publicly share further details about each recorded instance of fishes helped on our ARA Dashboard (see "Fishes Helped Details" at the bottom).

  • Note that the table only shows instances from 2026 onwards, so it may not display any yet.

Early Results from Testing the New System

We began piloting this system in August 2025. To evaluate it, we compared how many fishes were counted as helped from August to December 2025 versus the same period in 2024 (see the full dataset we used for analysis here):

• In 2025, nine instances of fishes helped were counted (and three more are pending review at the date of writing).

• In 2024, 13 instances were counted.

Having more instances counted in 2025 as opposed to 2024 raises concerns about whether the new system might be underestimating impact. Upon further review, we were not concerned about that because:

1. Nomination rigor improved: Fewer severe water quality issues met the new threshold in 2025 (22) compared to 2024 (57), suggesting stricter standards with the new process. It seems to us that the old system’s issue was nominating too many instances compared to the ones ultimately counted.

2. Imbalance complicates analysis: Having much fewer nominations in 2025 may explain why ultimately fewer instances were counted. If approved, the pending instances may actually balance the 2024 and 2025 counted instances.

3. Caution principle applies: Ultimately, we’d rather underestimate our impact—as opposed to overestimating it. We built this new process based on the best available evidence for fish welfare and feel confident that it accurately reflects our impact.

Next Steps

In line with our values of evidence-base and transparency, we will continue evaluating this system and will publish updates here should our data and conclusions change. In the meantime, we invite you to review our process and instances of fishes helped on our ARA dashboard (shows only instances from 2026 onwards, so it may not display any yet).

For feedback and thoughts, please reach out to jennifer@fwi.fish.