A trade you have outgrown is not a failure

Almost nobody decides to stop this work on a good day, which is exactly the problem. The decision gets made in the worst week of a quiet quarter, on feelings that a spreadsheet would have contradicted.

There is a version of this decision made from figures you already have, and it produces a different answer often enough to be worth the hour it takes.

The three numbers that actually decide it

You are not asking whether the work is enjoyable this month. You are asking three separate questions that people collapse into one.

What is your effective hourly take-home, over the last three months rather than the last three jobs? Total money that reached your bank, divided by total hours including admin, enquiries that went nowhere, and the time spent being reachable. That last category is the one that gets omitted, and it is frequently the largest.

Is the trend flat, rising or falling? A falling rate over three months is a different fact from a low rate that is climbing. Most people who quit at a genuinely low number are quitting a rising one at the wrong moment, because the rise happens through speed rather than price and speed is invisible from the inside.

What is the next best use of those hours, honestly priced? Not the job you would like to have. The one available to you next month. What your time is worth outside this trade is the comparison that settles it, and a rate that beats the market band and loses to your alternative is still the wrong rate for you.

If the effective rate is below the alternative and the trend is flat or falling across a full quarter, that is a decision. Anything else is a bad fortnight.

The signals that are not signals

A quiet month is the most common trigger and the weakest evidence. Demand in this trade is seasonal, and a slow stretch that repeats at the same time each year is a calendar entry rather than a verdict - treating a quiet month as data separates the two failure modes, because no enquiries and unconverted enquiries have opposite fixes.

One difficult buyer is not a signal either. Neither is a bad review, a fee change, or a fortnight of silence after a price rise, all of which look catastrophic on the day and unremarkable on a twelve-month chart.

The real signals are duller. Dread before opening the queue, sustained for weeks rather than after one job. The World Health Organization's ICD-11 description of burn-out covers the same ground for work in general: "feelings of energy depletion or exhaustion", growing mental distance or cynicism about the job, and reduced professional efficacy. WHO classifies it as an occupational phenomenon rather than a medical condition, and anything persistent is worth raising with a doctor. A ceiling you have calculated and cannot move, in a trade where growth comes from price and format because every job still needs you personally. Or a simple substitution: something else now pays more per hour and you have checked rather than assumed.

Partial stopping is usually the right answer

Full exit is the least common correct decision, and the one people jump to.

You can drop the format that pays worst per hour and keep the two that do not, which is what tracking take-home per format exists to tell you. You can stop taking new buyers and serve repeat ones only, which cuts almost all of the unbilled enquiry time while keeping most of the income. You can go dormant for a defined period instead of closing, on the understanding that visibility decays and is recovered more slowly than it was lost.

Each of these is reversible. A closed account, a deleted listing and a cancelled payout method are not, or at least not cheaply.

What you owe in flight

Whatever you decide, the queue you already took is not part of the decision. Work that has been paid for gets delivered to the scope that was agreed, on the date that was agreed, at the standard the reviews were built on.

Stop accepting new work first, deliver the rest second, close third. That order costs you a fortnight and protects the one thing that is genuinely portable, which is a working name that has never left anyone holding nothing.

Buyers notice this more than earners expect. The commissioning side's own account of how a professional exchange is expected to run treats an unfinished job as the single unrecoverable breach, and it is the one thing that circulates.

Records outlive the account

The window in which somebody can ask you a question does not close when your profile does. Disputes surface months later; tax deadlines arrive yearly; a chargeback can land after the account that hosted it is gone.

Export everything before you close anything: transaction history, payout records, message threads for in-flight jobs, and the platform's own annual summaries. Then keep it for the retention period your jurisdiction sets rather than the one that feels reasonable - how long to keep what works through the general shape, and the specifics differ by country, so check yours.

Withdraw the balance before you close, and check the minimum threshold first, because a balance below it is stranded once the listing that was generating income has gone. What that threshold is, and which methods clear fastest, is platform-specific rather than general: Rate Cock's payout and rewards page is the current source for its own.

None of this is tax or legal advice, and no post is.

If the answer is stop

Do it as a sequence rather than an event, which is what winding down without burning anything covers in mechanical detail.

Two things are worth knowing on the way out. The skills that transferred were never about the subject matter, and the market you are leaving is not the market you joined - the floor of generic assessment has been taken by automation, and the technical account of what automated scoring now does for free explains why the low tiers stopped paying rather than why you did. The consumer-facing tools built on that scoring are what most of the volume at the bottom of the market went to, and it is not coming back.

What is left at the top is judgement by a named person, which is the part that still prices. If that part still pays you more than your alternative, the numbers are telling you something, and the bad week is telling you nothing.