When to quit — the signs that mean it, and the ones that only feel like they do

You've drafted the resignation email in your head three times this month, always right after a meeting that ran fifteen minutes long, never twenty-four hours later. That timing is the first clue. Most quitting impulses are weather — they pass through and clear. A small number are climate — they describe the actual conditions you're living in, and they don't change on their own. The problem is that weather and climate feel identical from inside Tuesday afternoon. You need a way to tell them apart that doesn't rely on how you feel at 4 p.m.

Three signals that hold

The first is a comp gap that's been priced and confirmed, not guessed at. Pull actual numbers from Levels.fyi, Glassdoor, or a recruiter you trust, and compare them to your current salary. If the market rate for your role sits 15% or more above what you're paid, and that gap has survived two raise cycles without closing, that's structural. A $70,000 salary against an $82,000 market rate isn't a bad mood — it's $12,000 a year, compounding, because raises are almost always calculated as a percentage of your current pay, not as a correction toward the market floor. Companies do this on purpose: annual raise budgets are set as a percentage of payroll before the fiscal year starts, which means your individual case is competing against a number that was fixed months before your manager ever heard your argument.

The second is scope that's been quietly reduced. Not "I feel less important" — actual, countable scope: a project you owned got reassigned, a report line disappeared, a budget you controlled went to someone else, and it's stayed that way for two consecutive quarters. Titles are sticky; responsibility is not. When the title stays and the responsibility leaves, that's the company telling you something it isn't going to say out loud.

The third is a broken promise with a date attached, repeated. Not a vague sense that leadership doesn't value you — a specific commitment ("you'll be considered for the senior role in Q2," "we'll revisit your comp after the merger closes") that came with a timeline, missed the timeline, and then got a second promise that also missed. One broken promise is a bad quarter for the business. Two is a pattern, and patterns don't fix themselves because you waited quietly through the first one.

Three signals that pass

The first is a single bad meeting or a blowup with one person. Test it against thirty days: if it doesn't repeat, it was an incident, not a condition. People have bad weeks. Managers have bad weeks. One data point is not a trend line.

The second is a temporary workload spike — a launch, a close, a deadline that's eating your weekends right now. Check your calendar for an actual end date. If there's a date on the calendar where the spike ends, you're not burned out by the job; you're burned out by a project, and projects end.

The third is comparison envy — someone else's new title, new offer, or humble-brag post lands in your feed and suddenly your own job looks unbearable by contrast. Run one test: would you still want to leave if you'd never seen the post? If the answer is no, what you're reacting to isn't your job, it's someone else's highlight reel, and that's not information about your situation.

The audit itself

Do this on a weekend, away from your desk, with two documents in front of you: your last two pay stubs or offer letters, and your calendar going back six months. Write down, in plain sentences, the specific promises made to you and the dates attached to them. Write down the specific scope changes and when they happened. Leave feelings out of it — you're building a timeline, not a diary entry. If what you've written describes a pattern across two cycles or two quarters, you're looking at climate. If it describes one bad week or one bad meeting, you're looking at weather, and weather clears.

Before you act on a durable signal, test whether it's actually fixed-in-place or just unaddressed. That means one direct conversation, stated as fact rather than complaint:

"I want to keep contributing here, but the last two cycles have shown a real gap between my scope and my compensation. I'm looking for a concrete plan to close that within the next two quarters — what would that require from my side, and who needs to sign off on it?"

What you're listening for isn't reassurance. It's specificity. A real answer names a number, a timeline, and a person accountable for it. A non-answer repeats the word "visibility" or "exposure" without attaching either to a date. The second response is itself a data point — it tells you the gap isn't a communication problem, it's a budget or a priority problem, and neither of those resolves because you asked nicely.

The replacement math matters here too: the commonly cited range for replacing a departing employee runs from six months' to two years' worth of their salary once you count recruiting, onboarding, and lost productivity. Companies know this. It's why a direct, specific ask sometimes moves faster than people expect — you're not asking for a favor, you're presenting a cheaper alternative to your own exit.

Open your calendar tonight and write down the last specific promise your manager made you, with the date it was made and the date it was supposed to happen. That single sentence is the test. If you can't find one, you're dealing with weather. If you can, and it's already missed, you have your answer in writing before the weekend's over.