Outcome goals vs process goals — pick the wrong one and you'll stall

Lose two stone. Get the promotion. Publish the book.

Every one of those is a real thing to want, and none of them tells you what to do this evening. That's not a flaw in how they're written — it's what an outcome is. Outcomes describe a destination, and destinations are silent about the walking.

A goal that only describes where you're going will stall, reliably, and it'll stall in a way that looks like a character problem when it's actually a specification problem.

The two kinds, and how each one breaks

An outcome goal names a result you don't fully control. Lose two stone. Get the promotion. Have a thousand subscribers. You influence these, sometimes heavily, but so does biology, a hiring committee, and other people's attention.

A process goal names a behaviour you do control. Swim twice a week. Send two applications a month. Write six hundred words before work. Whether you did it is a matter of fact, and it's entirely up to you.

The obvious point is that you need both. The less obvious one is that they fail differently, and knowing which failure you're looking at tells you what to fix.

Outcome-only goals stall at the beginning, and the mechanism is ordinary.

Every evening, an outcome goal requires you to translate it into an action before anything can happen. What does get the promotion imply for tonight? There is an answer, but finding it is work — invisible, unrewarded work that has to be redone every single time, and which competes with doing nothing.

So the goal is never quite abandoned and never quite started. It sits on the list looking legitimate, producing a small wince, and the gap grows to six weeks without any decision having been made.

There's a second failure that's worse. Because the outcome isn't fully yours, effort and result come apart. A month of genuine work can produce no visible movement — the weight doesn't shift, nobody replies, the subscriber count sits still — and an outcome goal has no way to record that the month happened. Judged against the destination, you've failed. Judged against what you actually did, you've done well. Nothing in the goal can tell those apart, and people stop doing things that never register.

Process-only goals fail the other way round, and it's less discussed because it looks like success.

A process goal you're hitting produces steady satisfaction whether or not it's working. Send two applications a month can run for a year while the applications go to the wrong jobs, or the CV is the problem, and every month you'll have hit your goal. The behaviour is intact and the outcome is nowhere, and the process gives you no reason to ask why.

This is the mechanism behind a lot of very disciplined stagnation. Someone runs three times a week for two years and gets no faster, because the process was never checked against the thing it was supposed to produce. The process is the loyalty; the outcome is the audit, and without the audit the loyalty can be pointed anywhere.

Both failures share a root, which is worth naming. A goal has two jobs — telling you what to do tonight, and telling you whether it's working — and each type does one of them. Choosing between them is choosing which question you'd rather be unable to answer.

The pairing that works

Outcome as the reason, process as the goal.

Concretely: the thing you write down and track is the behaviour, and the outcome lives in the one-line why underneath it. Swim Tuesday and Thursday before work. Because my back stops hurting when I swim regularly.

That structure does three useful things. The tracked thing is the thing you control, so the record is honest and effort registers even in months where nothing visible moves. The outcome stays present as a reason, so the process can't drift into an unexamined ritual. And when the two come apart — the swimming happens, the back still hurts — you've got both halves in front of you and can ask the right question, which is whether the process is wrong rather than whether you are.

Then review the pairing rather than the behaviour. Every couple of months: is this process still producing the outcome? That's the audit, and it's the step almost nobody does.

Where the distinction is oversold

An honest caveat, since this framework gets recited more often than it gets examined.

Some outcomes are genuinely the right thing to track, because the process is unknowable in advance. If you're trying to solve a problem nobody has solved, there is no correct daily behaviour to commit to — inventing one produces a comfortable ritual that has nothing to do with the result. Ship the beta is a legitimate goal, and reducing it to code for an hour a day would be worse.

And the split is less clean than it's usually presented. Send two applications a month looks like pure process until you notice it depends on there being jobs to apply for. Most real goals are a mixture, and forcing them into one category is a diagnostic aid rather than a truth about the world.

The rule that survives: whatever you write down should be something you can act on tonight without a translation step. Sometimes that's a behaviour. Sometimes it's the next concrete move toward an outcome. What it must never be is a destination with nothing underneath it.

Two job hunts

Two people leave the same company in September.

The first writes get a new job by Christmas. It's on his list for eleven weeks. Some weeks he does a lot and some he does nothing, and because the goal only measures the destination, both feel identical — which is to say, like failure, because he hasn't got a job. By November he's stopped opening the list, and the goal is at 34 days quiet. Nothing about those eleven weeks is recorded anywhere.

The second writes two applications a week, Tuesday evenings. Because I want out of contract work and into something permanent. Tuesday is a fixed appointment with a defined task, so there's no translating to do. Ten weeks in, she has twenty applications, a record of all of them, and no offers.

That's the interesting moment. The process is intact and the outcome hasn't arrived, which is exactly the situation the pairing is for. She reads back the twenty and finds seventeen were for roles a level above anything on her CV. That's a diagnosis, and it's only available because the process left a record and the outcome sat next to it asking a question.

Neither of them got a job by Christmas. Only one of them knows why, and only one of them has anything to change in January.

What the gap counts

goaldrift counts days since you last wrote something, which suits process goals naturally — you swam, you logged it, the counter clears.

For outcome goals it does something more useful than it first appears. It can't tell whether you're closer to the promotion. What it can tell you is that you haven't touched the goal in nineteen days, which is the fact you'd otherwise never notice, because an outcome goal makes no noise while nothing is happening to it.

The destination stays in the description. The days are what get counted.

---

Related: Write the goal so tomorrow-you knows what to do · The thirty-minute unstick · Logging beats tracking