At the end of the week, the dashboard showed that Elena and Raúl had each resolved forty-two cases. The number looked like a clean comparison. Raúl had received complete forms, attached documents, and one pending administrative decision. Elena had supported people without stable addresses, requests with conflicting data, and records requiring calls to housing, health, and social services. In several cases she could not mark “closed” until three teams confirmed an answer. The dashboard did not show those differences. It repeated the same number beside their names.
An equal score can hide a radically different caseload. When it is used to distribute incentives, assess performance, or decide who deserves an opportunity, it turns simplification into an employment consequence. The metric cannot see coordination, listening time, attempts to repair missing data, or responsibility for preventing a harmful decision. Those tasks do not cease to exist. They fall outside the screen and begin to look like inefficiency on the part of the person doing them.
The problem is not solved by deleting every indicator. A team needs to know whether there are delays, unequal loads, or failing procedures. But an isolated number must not decide a person's value or replace discussion of the work they received. Fair measurement requires showing the composition of each caseload, recognizing interdependence, opening a route to provide context, and prohibiting employment decisions based on a score alone.
A number makes one part visible and erases the rest#
Closures are easy to count because they leave a clear mark in the system. Preparing a complex case is not. Elena may spend an hour rebuilding a history from incomplete documents, explaining what is missing without blaming someone, and coordinating three services that use different deadlines. If she reaches a solution, the result counts as one closure. If one service is delayed and the case remains open, the dashboard may read her performance as worse. The metric turns someone else's dependency into an individual signal.
Closures are easy to count because they leave a clear mark in the system. Preparing a complex case is not.
The effect grows when the system rewards volume alone. People learn that taking simple cases protects their number and that accepting a confused situation carries risk without recognition. Over time, difficult records concentrate with those who cannot avoid them, have more experience, or show greater willingness to help. Inequality does not begin with stated bad intent. It begins with a rule that calls incomparable work comparable.
The first correction is to describe what kind of work enters the count. A dashboard can distinguish complete records, cases with missing information, cross-service coordination, urgent reviews, and later repairs. It need not turn these categories into minute-by-minute surveillance. It is enough for the team to see that forty-two closures do not necessarily mean forty-two tasks of the same difficulty or burden.
Showing caseload composition changes the conversation#
A caseload is not simply a list of cases. It is a distribution of conditions. It can include urgency, language barriers, incomplete data, decisions affecting housing or health, and dependencies on other teams. Showing this composition makes it possible to ask whether allocation is reasonable before asking who reaches the finish first. It can also reveal whether one person systematically receives work needing more time because of an automatic rule or an established team habit.
Information should be used to balance, not label. If Elena has handled many complex coordinations, the team can adjust her next load, offer support, or recognize that work in planning. It need not retain personal details about people receiving service to do so. Case categories and aggregate process times can be used with privacy controls and review by people who know the service.
Composition should be updated during the week. A seemingly simple form can become difficult when a signature is missing or an emergency appears. The person handling it needs to be able to change its category and explain why, without fearing that this will look like an excuse. If the system cannot let actual work correct the initial label, the label will end up governing a reality it no longer describes.
Coordination also produces an outcome#
Much of the work supporting a solution happens between teams. Calling a health professional, verifying a place, clarifying a rule with housing, or contacting a person who could not reply are not unproductive detours. They are the infrastructure of an answer that does not abandon someone with a less orderly case. When coordination does not count, the fastest exit is rewarded rather than the safest one.
Recognizing it does not mean celebrating purposeless meetings. It means separating necessary coordination from repeated bureaucracy and giving the former time, authority, and visibility. A team can record the type of dependency and result achieved, not every word of a conversation. It can then identify which services delay cases, where agreements are missing, and which repairs should become a shared procedure.
Later repairs deserve the same treatment. If a decision is corrected because new data appear or a rule was misapplied, that correction is evidence about the process. Penalizing the person who makes it teaches people to hide it. Recording it as learning permits review of the rule, improvement of the form, or redistribution of a burden. An organization that counts only closures loses the chance to understand why some cases return.
Context must be discussable without punishment#
A useful context route lets a person say, “This result does not represent the caseload I received.” It should be brief, accessible, and reach someone able to review an assessment. It cannot demand endless narratives or become a request for permission to do work properly. It can ask for a complexity category, an external dependency, and a concrete explanation of repair carried out.
Review needs a deadline and protection. If providing context delays promotion or brings more difficult cases the following week, no one will use it. Management must explain how this information is considered, prohibit retaliation, and review in aggregate whether people using the route suffer consequences. The remedy becomes credible when it changes decisions, not when it merely collects comments in a system.
Worker representation can participate in this review and look at patterns rather than private lives. If one role always absorbs invisible coordination, if certain shifts receive incomplete cases, or if a score disadvantages people who repair more, the problem is design. Correcting it requires changing allocation, time, and criteria—not asking each person to defend themselves alone against a number.
No employment decision should rest on one score#
A score can be a signal to ask a question, never a sufficient answer for cutting hours, denying promotion, or beginning a sanction. Before an employment consequence, there must be substantive human review, information about caseload composition, an opportunity to provide context, and a written reason that can be challenged. This rule protects people and improves the decision: it forces attention to work rather than obedience to a summary.
The next step is to take two caseloads that look equal on the dashboard and compare them with the team. Record which dependencies, repairs, and decisions each contained. Then add a composition view and agree that no pay or disciplinary assessment will rest only on the closure count. A score can organize information. It cannot decide which work deserves recognition when it has left out the very work that made resolution possible.
