Cohort Ease: Measuring Whether Systems Become Easier Over Time
On Cohort Ease, System Friction, and the Work of Making Things Easier
There is a form of administrative violence that is not recognized as such.
It is visible in systems that appear to work, but do not improve. In training programs where placement rates hold steady, but each new cohort must take just as long to secure employment. In service pathways where participants are required to submit the same documentation repeatedly, because institutions do not recognize or trust each other’s data. In enterprise programs where access to finance expands, yet each cohort encounters the same collateral demands, pricing uncertainty, and approval delays as the last. In each case, success is recorded. But nothing has become easier.
This is not only a problem of design or delivery. It is a failure to learn.
When systems do not systematically learn what makes participation easier, and redesign accordingly, they do something quite specific. They ask each new cohort to carry the same burden as the last. They must navigate the same barriers, repeat the same steps, and absorb the same delays, costs, and uncertainties. When they succeed, we call it impact. But this is not improvement. It is repetition. And the cost of that repetition is not borne by the system. It is externalized to those who must pass through it.
This argument builds on two earlier pieces. In On Ease, Cohorts, and the Limits of Tracking Outcomes, I argued that tracking individual success without asking whether systems become easier over time produces a false sense of progress. In Administrative Violence II, I explored how bureaucratic and procedural systems impose hidden costs on those they are meant to serve, often invisibly, often normalized. This piece extends both arguments into a practical proposition: that we should not only observe these dynamics, but discipline ourselves to change them, and to test whether that change holds over time.
Ease, in this framing, is not a matter of perception or satisfaction. It is structural. It is the amount of time, effort, cost, and uncertainty required to move through a pathway, whether that pathway leads into education, training, enterprise, finance, or employment. When these conditions do not improve across successive cohorts, it tells us something important. Not that individuals are failing, but that the system is not learning.
Most monitoring and evaluation systems are not designed to detect this, and in doing so, they create a particular kind of misrecognition. By focusing on whether individuals succeed, they allow systems that do not improve to appear effective. Success becomes evidence of function, even where the underlying conditions remain unchanged. A cohort completes training and finds work, and this is taken as proof that the pathway works. An enterprise accesses finance and grows, and this is treated as impact. What remains invisible is whether the next cohort faced the same documentation burden, whether onboarding took as long, whether the same drop-off points persisted, and whether the same bottlenecks remained untouched.
What is required, then, is not simply another metric, but a different way of seeing. Cohort ease tracking names this problem directly. It begins from a simple premise: that impact cannot be claimed if the conditions of participation do not become easier over time. This requires not only observing friction, but ensuring that it is defined from the standpoint of those who experience it, rather than solely by the institutions that manage it.
This friction is not abstract. It appears at entry, in the form of documentation requirements, eligibility ambiguity, and the time it takes to onboard. It appears in progression, where participants encounter repetition of steps, re-verification, delays, and points of attrition. It appears in conversion, where the transition from participation to outcome is constrained by placement timelines, access to capital, or weak market linkages. And it appears in stability, where sustaining an outcome requires navigating income volatility, re-entry barriers, or the risk of falling back into informality. To take friction seriously is to make visible the structural drag that is otherwise absorbed silently by those moving through the system.
Yet observing friction is only the beginning. Much of what makes participation difficult does not sit neatly within program boundaries. A training system that cannot signal quality to employers will reproduce long placement cycles no matter how many cohorts pass through it. A financial system that prices risk conservatively will extend access without materially reducing the cost or uncertainty of borrowing. A public system that requires repeated verification will continue to shift administrative burden onto individuals, regardless of how many are processed.
Many of these constraints are not incidental. They reflect underlying distributions of power, inclusion, and recognition, and do not yield easily to technical adjustment alone. What presents as friction is often the visible surface of deeper political and institutional arrangements that determine who is seen, who is served, and on what terms.
This raises a more difficult question about responsibility. Development actors are typically held accountable for what they control, which includes program design, delivery quality, and partner performance. Everything else is treated as context. But context, left unaddressed, hardens into structure, and structure, left unchanged, begins to look inevitable. The question is not whether any single actor can control these broader conditions. They cannot. The question is whether they are willing to engage them, deliberately and systematically, rather than routing around them indefinitely.
Cohort ease tracking therefore requires a shift in posture. For each recurring friction point, it asks a simple but demanding question: what did we do about it, and did it make a difference for the next cohort? Sometimes the answer will lie within the program itself, where processes can be simplified, requirements reduced, and sequencing improved. Sometimes it will sit at the interface with partners, where expectations can be reset and practices adjusted. And sometimes it will require engagement beyond the immediate sphere of implementation, whether through coordination with public institutions, alignment with regulatory processes, or contribution to broader system reform. These responses operate on different time horizons, and not all will benefit the current cohort. But without them, the system remains static, and the burden persists.
What this introduces is a second layer of impact that sits alongside, but is distinct from, individual outcomes. It asks not only whether people succeed, but whether the system becomes easier to navigate over time. It pays attention to whether solutions developed within programs begin to travel beyond them, whether partners adopt new defaults, whether institutional processes shift, and whether barriers begin to reduce not only for participants, but for those who never enter the program at all.
These burdens are not evenly distributed. They fall differently across gender, legal status, class, and geography, and any serious attempt to reduce them must take that unevenness into account. Without this, the easing of systems for some can coexist with the persistence, or even deepening, of difficulty for others.
This is the difference between delivering outcomes within a system and contributing to the gradual easing of that system itself.
This way of thinking is not without precedent. It draws on a long tradition that understands development as a process of problem-solving rather than blueprint execution, as argued by Albert Hirschman, who showed that progress emerges through iterative engagement with constraints rather than their prior resolution. It reflects Amartya Sen’s insistence that well-being must be understood as the expansion of real freedoms, not simply the achievement of outcomes, and therefore must take seriously the conditions that enable or constrain action. It recognizes, following Michael Lipsky, that systems are not experienced as policy but as practice, shaped in their daily operation by those who implement them. And it aligns with Elinor Ostrom’s work in showing that outcomes are produced through the interaction of multiple institutions rather than any single point of control. What this framing adds is a requirement that is often implied but rarely enforced: that systems should not only function, but should become easier to move through over time, and that this easing should be observable, deliberate, and cumulative.
This introduces a different form of accountability. It is no longer sufficient to report results. It becomes necessary to account for the persistence of difficulty. If successive cohorts encounter the same barriers, in the same places, with the same intensity, then something is not being addressed, even if outcomes remain strong. Indeed, strong outcomes can obscure this condition, because they demonstrate that people are capable of navigating constraint. But capability should not be confused with fairness, nor resilience with system performance. People find ways through, often at considerable cost. The question is whether they should have to bear that cost repeatedly.
None of this displaces the importance of tracking individual outcomes. Programs must still deliver, and people must still benefit in tangible ways. What cohort ease tracking proposes is an additional lens, one that runs in parallel and asks whether the passage of one cohort leaves the system any easier for the next. It introduces a discipline of identifying friction, acting on it, and then testing whether that action has reduced the burden that future participants must carry.
If taken seriously, this changes the nature of the work. Monitoring systems begin to track not only flows, but points of resistance. Program design begins to concern itself not only with delivery, but with the reduction of unnecessary burden. Partnerships begin to extend beyond implementation into joint problem-solving across institutional boundaries. And impact begins to be understood not only as movement within systems, but as the gradual easing of those systems themselves.
Ultimately, this is what improvement looks like. Not that some people make it through, but that it becomes easier for others.