Availability is a decision, not a default

Keep this work from eating your week by picking two availability windows a day and one weekend block, publishing them, batching jobs inside them, and treating everything outside as closed. Enquiries arrive in evenings and weekends, and fast replies are a newcomer's main advantage, so without windows the work expands into every free hour.

Where the hours actually go

Almost nobody is working eight hours and calling it two. What happens is smaller and harder to see: fifteen minutes of checking, six times a day, across sixteen hours.

Count the categories honestly.

• Producing the work. The only part that is billed, and usually the minority of the total.
• Enquiry handling. Replies to people who will not book, which is most of them.
• Waiting. The dead time between sending a clarifying question and getting an answer, during which you are not free but not working.
• Checking. Opening the app for no reason because a notification might have arrived.

The last two are the ones that eat the week, and neither shows up in any hourly estimate you would give someone who asked what this pays. An honest effective rate divides take-home by all four, and why that number keeps moving is mostly a story about the unbilled three shrinking rather than the billed one growing.

Windows beat availability

Pick two windows a day and one weekend block. Inside them you answer within minutes, work through the queue and deliver. Outside them the app is closed.

Two windows works because it matches the actual arrival pattern: something in the late afternoon, something in the evening. Where your peak sits depends on where your buyers live rather than where you do, which is worth checking against your own order timestamps before you assume it matches your local evening.

The competitive advantage survives this almost intact. Being first to reply during the busy hours beats being available during all of them, because the buyer comparing three profiles at nine in the evening is not comparing your midnight response time to anyone's. What you lose is the enquiry that arrived at two in the morning and was answered at five in the afternoon, and that buyer was mostly not going to book anyway.

State the windows on your listing and in your auto-reply if the platform has one. A stated response time is a promise a buyer can check, and it converts better than vagueness - the same mechanism that makes a stated turnaround worth having, covered in turnaround as a product feature.

Batching, and why five jobs in a row take less than five jobs apart

Setup is the fixed cost of every job: opening the material, re-reading your rubric, getting into the mode of paying attention properly. That cost is paid once per session rather than once per job, so five assessments in one sitting take meaningfully less time than the same five spread across five evenings.

The corollary is that a same-day promise on everything destroys the ability to batch, because each job has to be started when it arrives. A stated turnaround of two days, honoured reliably, buys you the right to schedule your own week, and it is worth more than the small conversion advantage of same-day.

If you do want to sell speed, sell it explicitly as a priced option rather than giving it away as a default - the argument is in what urgency is actually worth.

Capacity is a number, so write it down

Decide how many jobs a week you will take before demand tests you, because deciding in the moment always produces yes.

Work it out from your own timings: minutes per job, times jobs, plus roughly a third again for the unbilled parts, against the hours you actually want to spend. The result is usually smaller than people expect and it is the honest ceiling of this work at your current price. Set it with the evidence on long hours in mind: a joint WHO and ILO analysis (2021) estimated that working 55 or more hours a week is associated with a 35% higher risk of stroke and a 17% higher risk of dying from ischemic heart disease, compared with 35-40 hours. Evening rating work stacked on a day job reaches that total faster than it seems.

Once the number exists, exceeding it becomes a visible decision rather than a slow drift. It also converts a queue into information: consistently turning work away is the only reliable evidence that your rate is below what the market will bear, which is a better trigger for a price rise than nerve.

The alternative to a capacity number is the trade that expands until the hours are gone and the per-job pay hides the fact - which is the whole mechanism described in how per-job pricing conceals overwork.

What to do about the hours you cannot move

Some demand genuinely sits outside any window you would choose, particularly if your buyers cluster in a timezone six hours from yours.

Three options, in order of how much they cost you. Publish a turnaround rather than a response time, so a delayed reply is not a broken promise. Shift one window rather than adding a third, and accept that you are trading a local evening for a remote one. Or price the awkward slot, which is the honest version and the one people skip - a booking that costs you a Saturday morning should not cost the same as one that does not.

Buyers are less demanding about timing than earners assume. The commissioning-side account of how a job normally proceeds reads as people who expect a considered reply within a day, not a reply within a minute, and the pressure to be instantly present is largely self-imposed. Where instant genuinely is the expectation, it is because the buyer has been using something automated first, and what an automated score delivers in two seconds is the standard they have been comparing against - competing with it on speed is not a contest you can win, and the packaged tools that make it one tap are why it is not worth entering.

Platforms increasingly let you set availability, pause a listing or schedule quiet hours without a ranking penalty, and the details matter because the penalty is not always zero. Rate Cock's judges page documents its own version, which is the difference between a paused week and a disappeared listing.