Written by Jamie Padilla, Director of Strategic Engagement, FishChoice
In conversations about labor and human rights in global supply chains, there’s a familiar rhythm. We define expectations, interrogate accountability, and look for proof. We ask whether companies are doing the right thing.
The conversation moves quickly to what should happen, how to measure it, and how to hold companies accountable when it doesn’t.
What we spend far less time asking is a more uncomfortable question: do we actually understand what it takes for them to do it?
Because if we’re honest, a lot of the conversation skips over that part entirely. And then we build entire strategies on top of that gap. There’s a baked-in assumption at play—that once expectations are clear, the rest is just execution. If something isn’t happening, then someone chose not to do it, failed to enforce it, or isn’t serious enough.
It’s a clean way of reading the world. It’s also often wrong.
In challenging sectors (like seafood and agriculture), labor systems don’t operate in stable conditions. Workforces can change constantly—sometimes seasonally, sometimes week to week—and sometimes they don’t “change” so much as they were never uniform to begin with. You can have a workforce all at once that speaks multiple languages, comes from different regions, and operates with completely different norms around authority, communication, and risk, and you’re expected to design and operate systems that somehow work across all of that.
You can get something working—communication is flowing, worker reps are active, people are using the channels available to them—and then something shifts. A new group comes in. Or dynamics shift within the same workforce. Trust isn’t uniform. Comfort with speaking up isn’t uniform. What feels accessible to one group doesn’t land the same way for another, and suddenly what looked like a “functioning system” starts to fray at the edges. You’re not just running it anymore. You’re adjusting it, sometimes in small ways, sometimes rebuilding parts of it without really calling it that.
And then everything else starts pressing on those systems at the same time. Production changes. Buyer requirements change. Something as seemingly simple as a packaging decision made downstream changes the pace of work, which changes earnings, which changes retention, which changes how your labor system behaves whether you intended it to or not.
In seafood, you add another layer entirely. You don’t control fishing conditions. Weather shifts. Catch volumes fluctuate. Trips run longer than expected. Processing facilities get hit with sudden surges—large volumes landing at once that have to be handled immediately. Everything speeds up. Shifts stretch. Decisions get made under pressure. You can have systems in place and still see very clearly how they hold—or don’t—under those conditions.
Even when companies are trying—making deliberate decisions to stabilize their workforce, improve retention, create more consistency—the outcomes don’t line up as neatly as people assume they should.
Because the environment itself doesn’t hold still.
There are too many moving parts, some inside the operation, many outside of it, and they don’t respond in predictable ways. You can make what looks like the right decision and still not get the result you expected.
And yet we talk about this as if it’s something managers or HR should just inherently know how to do—as if there’s a clear playbook they’re failing to follow.
There isn’t.
In many cases, the people responsible for making these systems function didn’t come into their roles because of deep experience in managing diverse workforces or designing labor systems. They’re operators. They understand production. They’ve built expertise in the technical side of their business. Managing complex, multilingual, culturally diverse workforces—often under pressure and with limited support—is something they’ve had to figure out along the way.
That doesn’t make the responsibility any less real. But it does change how we understand what it takes to get it right.
Because this is the part we tend to move past the fastest—the messy part, the part that actually determines whether anything works.
And there are a lot of reasons why it gets skipped.
Sometimes it’s a lack of understanding. Not in a superficial sense, but in the very real sense that many of the people shaping these conversations have never had to make systems work inside operations where conditions are constantly shifting and work is inherently difficult. And from the outside, it’s easy to underestimate just how much is being navigated at once.
Even under the best of circumstances, the work itself is hard. We’re not talking about controlled environments or predictable workflows. We’re talking about physically demanding, time-sensitive, and often volatile conditions—where fatigue, pace, and pressure are part of the job, not the exception.
That matters.
Because when the baseline is already demanding, the margin for getting everything “right” all the time is thinner than people assume. Systems don’t operate in isolation from that reality—they’re constantly interacting with it.
There’s another reason this part of the conversation gets avoided.
Anything that sounds like it might explain why outcomes fall short can feel uncomfortably close to excusing them.
There’s a strong pull toward clarity—this is the expectation, and there is no excuse for not meeting it. And in many situations, especially where harm is clear, that kind of clarity is necessary. It creates accountability. It signals urgency.
But the reality underneath doesn’t always organize itself that cleanly.
There are degrees of harm. Degrees of intent. Degrees of control. And it can feel uncomfortable—even risky—to say that out loud.
Because once you start engaging with the factors that shape outcomes, the picture becomes more complicated. You’re no longer just asking whether something happened. You’re asking how it happened, why it happened, and what would need to change for it to stop.
Take recruitment fees.
At one end, you have highly organized, often sophisticated systems—sometimes tied to trafficking—where workers are systematically exploited through debt, control, and coordination across networks (including organized crime, in many cases). At the other end, you have informal, opportunistic arrangements—a local middleman taking a cut to connect workers to jobs, operating loosely, inconsistently, and with varying degrees of harm. And then there’s a wide range of practices in between.
Acknowledging that range can feel like minimizing harm. That’s part of why we tend to avoid it.
The stakes are high. It can feel like a slippery slope—like once you start differentiating, you’re weakening the case for urgency or accountability.
But that’s not what’s happening.
It’s about understanding the full reality of what’s happening so that responses are actually grounded in how these systems function. Because if everything is treated as equally clear-cut, it becomes harder to see what’s actually needed.
Another barrier to engaging with this part of the work—the messy, real-world process of figuring out how to make systems actually function in practice—is the belief that the right model will fix things.
If outcomes aren’t where they should be, the conclusion becomes: we just haven’t applied the right approach yet. From there, the conversation quickly turns into an argument for one model or another—each positioned as the solution.
There’s a strong appeal to that way of thinking. It offers clarity. It offers direction. It offers the sense that if we can just get this right thing in place, the rest will follow.
It also offers something else: a way to make sense of failure.
If harm exists, it can be read as proof that the current approach isn’t working—and that something else should replace it. But that logic can be misleading. Because it assumes that outcomes are primarily a function of the model itself, rather than the conditions it is operating within.
It treats evidence of harm as evidence that a given approach is invalid, rather than asking what that approach was actually able—or unable—to influence in that context.
And it creates a kind of binary thinking: either this model works, or it doesn’t. But that’s not how systems behave in practice. Different approaches interact with different parts of the system. They operate with different levels of reach, different resource requirements, different mechanisms for influence. Seeing harm in a system doesn’t automatically tell you which model would have prevented it—or whether any single model could have addressed all of the factors at play.
And yet the pull toward “this is the answer” remains strong because it simplifies a problem that is otherwise difficult to hold. It creates a focal point, a strategy, a sense of momentum.
But it also pulls attention away from a harder set of questions. Not just what model should exist—but how any model actually functions once it meets the realities described earlier. How it holds up under pressure. What parts of the system it can reach. Where it depends on conditions that may or may not be present. And what happens in the much larger set of contexts where that model isn’t operating at all.
That’s where the conversation needs to expand.
Not away from models—but beyond the idea that any single model, on its own, resolves the complexity of how labor practices take shape in the real world.
There’s another assumption that can pull attention away from this part of the work: that stronger laws or better enforcement will solve it. Those things matter. Legal frameworks and enforcement mechanisms are essential. But they’re not an automatic path to change, and treating them as if they are can flatten how we understand where risk actually shows up—and why.
It can lead to an overfocus on operations in places assumed to have weaker governance, while overlooking how risk functions across the system more broadly. The reality is more uneven than that. Some of the highest-risk dynamics—like high worker-paid recruitment fees tied to migration—are driven by where workers are trying to go. Workers will often take on more debt to reach higher-paying destinations—places assumed to be lower risk because of stronger legal frameworks.
The risk is still there. It’s just less visible.
At the same time, protections and enforcement are often assumed to be stronger than they are in places considered lower risk. They aren’t always. There are entire categories of workers—farmworkers, domestic workers—who sit outside key labor protections even in countries with highly developed legal systems. And enforcement gaps are real.
None of this is an argument against stronger laws or better enforcement. It’s a reminder that they don’t eliminate the need to understand how systems actually function in practice—or the need to engage in the work of making them function better. Because even with strong frameworks in place, the question remains how those protections actually show up—or fail to—under real conditions.
There’s also the question of pressure—on buyers and brands in particular.
That pressure is real, and it’s increasing. Legal exposure. Due diligence requirements. lawsuits. Withhold Release Orders. Reputational risk. All of it pushes in a very specific direction—toward clarity, defensibility, and proof.
Toward verification. KPIs. Compliance.
That makes sense. But it also shapes how the problem gets approached.
Because when the priority is to demonstrate that risk is being managed, the focus naturally shifts toward what can be shown, measured, and defended. And the harder, messier work of building systems that actually function can get pushed to the side—not because it doesn’t matter, but because it’s less legible.
The irony is that this doesn’t reduce risk in the long run. If anything, it exposes a different kind of fragility. Because when systems aren’t actually working on the ground, the signals used to demonstrate performance become less reliable. The gap between what can be shown and what is happening widens.
And that’s where things tend to break.
So this shifting landscape doesn’t weaken the case for engaging with the messy realities of implementation. It strengthens it. Because if expectations are rising—and scrutiny is increasing—then the question isn’t just how to prove something is happening. It’s how to make it more likely that it actually is.
That brings the focus back to implementation—what it actually takes to make expectations around responsible labor practices function in practice.
And this isn’t unknown territory.
Across different models and approaches—and in operations that have gone through real, often significant change processes—we’ve already learned a lot about what stronger systems look like in practice. We’re not starting from zero.
But companies are starting from very different places. Different supply chains, different operating contexts, different levels of maturity. There isn’t a single entry point, and there isn’t a one-size-fits-all path.
There are, however, patterns.
The work is a mix of things that are highly concrete and operational, alongside things that are less tangible but just as critical.
At one level, it’s about building the relational and structural foundations that allow systems to function—building trust over time, not assuming it’s there, and creating structures for worker voice that actually work inside the workplace: committees, representatives, and processes that people understand and use.
It also means getting the fundamentals of workforce entry right—how workers are recruited, what they’re told before they leave, how they’re onboarded, and how they’re supported once they arrive. In some cases, this extends further upstream, into recruitment practices and whether what workers arrive to matches what they were promised.
Communication is another core part of this work—designing systems that function across different languages, norms, and levels of familiarity, so that information flows, concerns surface, and grievance mechanisms are actually used in practice, not just formally available.
And then there are the day-to-day operational decisions that ultimately shape outcomes—how schedules are set, how tasks are assigned, what a clear contract actually looks like, and why systems that exist on paper don’t always function consistently across different groups within the same workforce.
This is the kind of work that deeper, hands-on models take on. And where those exist, they matter. Partnerships with worker organizations, NGOs, certification programs that include real engagement, and technical advisors can all play an important role in helping companies build and strengthen these systems.
But that level of support or intervention isn’t available everywhere. It’s time-intensive, resource-intensive, and for much of the supply chain, simply not present.
That doesn’t mean there’s no way to engage.
Because over time, through those deeper efforts, a wide range of tools and resources has been developed—practical things that reflect what it actually takes to build and run functioning systems–systems that work for workers.
Those tools are valuable. But they don’t move on their own. Access to them isn’t consistent, and in many parts of the supply chain—especially upstream—companies are still largely left to figure things out on their own.
That’s where something like Seafood LAB comes in (modeled on ECIP LAB, built for this same purpose for the produce sector).
It creates a way to get those tools and resources into the hands of the people responsible for making systems function—and a structure for actually using them over time. At the same time, it provides downstream actors with visibility into who is engaging in that work, and to what extent, creating a clearer picture of how effort is distributed across the supply chain.
It also connects users to a network of vetted service providers, organizations, and programs for those who want or are ready for more.
It’s a way of engaging with the work, not skipping over it.
At the end of the day, change isn’t just about expectations or proof. It’s about whether we’re willing to engage in the work that actually makes either of those meaningful. That work is messy. It’s uneven. And it doesn’t fit neatly into the kinds of conversations this field tends to default to.
That’s the part of the problem Seafood LAB is built for—and the conversation FishChoice and Verité are inviting others into, in partnership with the Equitable Food Initiative.
If this is the conversation you’ve been wanting to have—or are already having—let’s have it together.