Setting Goals That Do Not Become Theater

Every leadership team I have worked in eventually discovers the same uncomfortable truth: the goals you write down quietly become the goals you optimize for, whether or not they are the goals that matter. In a regulated fintech environment, this is not an abstract management concern. It is the difference between a quarter where we genuinely reduced settlement failures and one where we merely got better at not recording them.

I have set goals that changed how a team operated and goals that became elaborate theater performed for a slide deck. The line between the two is thinner than most people admit, and crossing it rarely feels like a mistake at the time. This is what I have learned about staying on the right side of it.

What Goal Theater Actually Looks Like

Theater does not announce itself. It shows up as a beautifully formatted objectives document, a confident review meeting, and a green status on every line item. The teams performing it are usually busy, sincere, and tired. Nobody thinks they are pretending, because the work is real even when the goal is hollow.

The clearest tell I have found is a goal whose completion produces no argument. If we hit a target and nothing about how we work, what we ship, or what we say no to has changed, the goal was probably decorative. Real goals create friction. They force a trade-off that someone in the room did not want to make. A goal that everyone can comfortably agree to in five minutes is often a goal that asks nothing of anyone.

The second tell is retroactive redefinition. When the metric is going to miss, a theatrical organization quietly renegotiates what the metric meant. The payment success rate did not drop; we simply reclassified those timeouts as customer-side events. I have watched honest people do this without malice, because admitting a goal was wrong felt more expensive than bending its definition.

Why Regulated Fintech Makes This Worse

In most software companies, a vanity goal wastes a quarter. In ours, a vanity goal can produce a control that looks like compliance but does not actually protect a customer or satisfy an examiner. We are surrounded by measurable proxies: number of alerts reviewed, percentage of tickets closed within SLA, count of controls marked effective. Every one of those proxies can be hit without the underlying risk moving at all.

I once inherited a fraud-operations team with a glittering metric: ninety-eight percent of flagged transactions reviewed within the hour. It was true. It was also nearly meaningless, because the review consisted of a human clicking approve on a queue tuned to almost never surface anything genuinely suspicious. The goal was being met perfectly. The fraud was getting through somewhere else entirely, in a flow nobody had instrumented.

The regulatory environment amplifies this because so much of what we do is judged on evidence of activity rather than evidence of outcome. It is easy to build a year of green metrics that an auditor will accept and that a determined attacker, or an unlucky settlement window, will walk straight through.

Anchoring Goals to Outcomes, Not Outputs

The single most useful discipline I apply is forcing every goal to name the outcome it is a proxy for, and then asking whether we could hit the proxy while the outcome got worse. If the answer is yes, the goal is incomplete and we keep working on it. Shipping the new reconciliation engine is an output. Reducing the number of payments stuck in an unresolved state for more than twenty-four hours is an outcome. The first can be declared done; the second has to actually happen.

This sounds obvious and is surprisingly hard to hold. Outputs are concrete and within the team's direct control. Outcomes are messier, partly dependent on factors we do not own, and sometimes embarrassing to commit to publicly. The instinct to retreat into output goals is strong precisely because output goals are safe. But safe goals are the raw material of theater.

If hitting a goal would not change the experience of a single customer, a single auditor, or a single engineer on call, I have learned to treat it as a status update wearing the costume of an objective.

Building the Trade-Off Into the Goal

A goal without a stated cost is a wish. When I set a target, I now insist on naming what we are willing to give up to reach it, and what we are explicitly not willing to give up. Reducing onboarding fraud is not a goal until we say whether we will accept more legitimate-customer friction to get there, and how much. Otherwise the team will optimize fraud down by quietly rejecting good customers, hit the number, and call it a win while the growth team wonders why conversion collapsed.

This is where goals stop being theater, because the trade-off creates the friction I mentioned earlier. Someone has to accept a worse number on their dashboard so that a more important number improves overall. That negotiation is uncomfortable, and the discomfort is the point. A goal that has survived an honest trade-off conversation is far harder to fake, because everyone already knows what it cost.

I keep a short set of questions for any proposed goal:

• What outcome is this a proxy for, and how could we hit it while the outcome got worse?
• What are we willing to sacrifice to achieve it, stated as a specific number or behavior?
• Who in the organization will be unhappy if we succeed, and have we talked to them?
• How would we know within weeks, not quarters, whether we are actually moving?
• What would tempt us to redefine the metric, and how do we lock the definition now?

The Case for Fewer, Heavier Goals

Theater thrives on volume. When a team carries fifteen objectives, no single one can fail loudly, because the noise of the other fourteen drowns it out. A long goal list is also a quiet abdication of prioritization: instead of deciding what matters most, we list everything that might matter and let the team sort it out under pressure. They do that by doing the easy ones and narrating the hard ones.

I have moved toward a small number of goals that are genuinely hard, where missing one would be a real event the leadership team has to reckon with. Three goals that can each fail visibly are worth more than a dozen that can all succeed quietly. Fewer goals also make trade-offs legible. With three priorities, you can reason about which one yields when they conflict. With fifteen, every conflict gets resolved invisibly by whoever happens to be holding the keyboard that day.

The objection I hear is that important work falls off the list. It does not disappear; it moves into the category of expected, ongoing engineering rather than headline objectives. Keeping the platform available and the books reconciled is not a goal in my framing. It is the floor. Goals are the things we are deliberately trying to change, and there should be few enough that the team can recite all of them without checking a document.

Choosing Measures That Resist Gaming

Every metric will be gamed if the incentive is strong enough, but some metrics are far easier to game than others. Single-number targets are the most fragile, because there is usually a path to the number that bypasses the intent. I now prefer paired metrics, where improving one in a fake way visibly damages the other. Pairing fraud loss with false-decline rate is the canonical example: you cannot quietly cheat one without the other moving against you.

I also try to measure things closer to the customer and further from the team's own instrumentation. A team measuring its own ticket-closure rate controls both the work and the scorekeeping, which is an invitation to drift. A team measured on whether customers stopped contacting us about a problem has far less room to redefine success. The further the measurement sits from the hand that benefits, the more honest it tends to stay.

None of this makes gaming impossible. It makes gaming expensive and visible, which is the realistic goal. I am trying to build a metric where distorting it costs more effort than simply doing the real work, so that the path of least resistance happens to be the honest one.

Running Reviews That Permit Real Failure

Goals become theater largely because of how we review them. If every status meeting punishes red and rewards green, people will manufacture green. The most important thing I do as a leader is make it genuinely safe, and even respected, to walk in and say a goal is failing and here is what I have learned. The first time someone does that and is thanked rather than grilled, the culture shifts more than any framework could shift it.

I structure reviews around learning rather than verdicts. The question is not did you hit it, but what do we now know that we did not know when we set this, and does that change whether the goal is still worth pursuing. Sometimes a missed goal is a failure of execution. Often it is a signal that the goal was wrong or we misunderstood the problem. A review that cannot distinguish those cases will train people to hide all of them equally.

This is also where I watch for the redefinition reflex. When someone proposes that a metric did not really miss because of some reclassification, I treat that as a moment requiring care rather than a convenient escape. If the reclassification is genuinely correct, we fix the definition deliberately and write down why. If it is a way to avoid an uncomfortable conversation, naming it gently in the room is usually enough to stop it, because everyone already suspects what is happening.

Knowing When to Abandon a Goal

Persistence is a virtue right up until it becomes theater of its own. A goal that has clearly stopped making sense, but that we keep on the list because abandoning it feels like an admission of failure, is one of the purest forms of organizational pretending. We are no longer pursuing an outcome; we are protecting a decision everyone privately knows no longer matters.

I try to give every significant goal an explicit condition under which we would drop it. If the assumption underneath it turns out to be false, we stop, openly, and reallocate. Naming that exit condition at the start removes most of the shame from later invoking it, because we are not improvising an excuse; we are executing a plan we agreed to when we were thinking clearly. Killing a goal on schedule is a sign of a healthy culture, not a weak one.

The hardest version is the goal that is mostly working but no longer the most important thing. Sunk effort makes it tempting to ride it to completion. I have learned to ask whether, if this goal did not exist today, we would choose to start it now. If the answer is clearly no, finishing it for the sake of finishing is just a disciplined form of theater.

Conclusion

Goals that do not become theater share a few traits: they name a real outcome rather than a comfortable proxy, they carry a stated trade-off that creates honest friction, they are few enough to fail visibly, they are measured in ways that punish gaming, and they are reviewed in a culture that treats a missed goal as information rather than a crime. None of that is exotic. It is mostly the willingness to keep goals uncomfortable on purpose, which runs against every instinct that wants the quarter to look tidy. In a regulated business, that discomfort is not a cost to be minimized. It is the evidence that the goal is doing its job, pointing at something that genuinely needs to change and refusing to let us pretend it already has.