Availability costs more than production

Ask an earner how long a job takes and you get the production time: forty minutes to write, twenty-five to record, ten for the admin. That number is accurate and it describes maybe half of the hours the trade actually takes, because the other half is spent being available for work that has not arrived yet.

Availability is unbilled, unmeasured, and the reason people who calculate a comfortable hourly rate still end up feeling like the work owns them.

Where the hours actually go

For a week with eight completed jobs, the split looks roughly like this. The production column is the only one anybody counts.

The last row is the one that has no natural size. It expands to fill whatever you allow, because there is always a plausible argument that one more hour with notifications on might produce one more booking.

Why the busy hours are not yours to choose

Demand in this trade clusters in evenings and at weekends, in whatever timezone your buyers live in. That is not a scheduling preference you can negotiate with. It is when people who have jobs are at home with time and a private moment, and it is precisely the window most earners were hoping to keep.

The practical consequence is that responsiveness during unsocial hours is worth more than competence during office hours, at least until you have a booked-out profile that buyers wait for. An earner who replies in four minutes at nine on a Saturday night beats a better earner who replies on Monday, and the gap is not close.

This is also why the first month feels disproportionately heavy. You are paying the full availability cost while receiving the smallest possible volume of work in return, and the cold start is essentially a description of that ratio at its worst.

The unconverted enquiry is the real tax

A job that never happens still costs you: reading the brief, deciding whether it fits, writing a reply, and often a second exchange before the buyer goes quiet. Call it eight to twelve minutes each.

If two-thirds of your enquiries do not convert - which is unremarkable early on - then for every hour of paid production you have spent a substantial fraction of an hour on people who paid nothing. The pattern is measurable on other platforms: an ILO survey of 3,500 crowdworkers (2018) found they spent on average 20 minutes on unpaid activities for every hour of paid work, including searching for tasks and researching clients. That fraction is the difference between the hourly rate you calculated and the one you are actually getting, and it is the single largest reason your effective hourly rate keeps moving without your price changing at all.

It also improves faster than any other number in the trade, because screening gets better with practice and because a visible price removes most of the enquiries that were never going to become work.

Structuring it so it does not eat the week

Four things do nearly all the work here, and none of them require being less responsive in the moments that count.

Publish windows rather than promising availability. "Replies 7-10pm most evenings, same-day turnaround on jobs booked before 8" is a stronger claim than being silently available all the time, because a buyer can plan around it. Stated hours also convert being offline from a failure into a policy.

Batch the checks instead of streaming them. Four deliberate checks in an evening beats a notification badge you glance at forty times, and produces roughly the same reply latency from the buyer's point of view. The badge is what makes an evening feel like work even when you did none.

Time-box the enquiry reply. Three sentences, one clarifying question, no more. Long, careful replies to thin briefs are how skilled people lose hours to buyers who were price-shopping five profiles. Answering quickly without being permanently online is a solvable problem, and the solution is mostly templates plus a decision rule.

Take the whole of one day off, weekly, and say so. Not scattered hours, a named day. Scattered rest does not read as rest to the part of your brain that is listening for a notification, and a named closed day is the only version of this that survives a busy month. Working hours that do not eat the week goes further into the scheduling mechanics.

What availability is worth, if you want to price it

You can convert the invisible half into a number, and it changes decisions.

Take your total hours in a week including everything above, divide by your take-home for that week, and compare it to what an hour of your time is worth elsewhere. Most people find the gap sits somewhere between mildly disappointing and fine, and the useful part is not the verdict but the components: if unconverted enquiries are eating a quarter of your hours, screening is worth more to you than a price rise.

There is a second use for the number. Being consistently available in the busy window is a service, and services can be charged for. Guaranteed same-evening turnaround, held slots and rush handling are the priced expressions of availability rather than free ones, and what urgency is actually worth is the conversion.

What the demand side sees

None of this is visible to buyers, who experience only the reply time. Their account of the interaction - how a commission normally opens and what a slow reply signals - is worth reading precisely because it contains no acknowledgement of the waiting, which tells you how much credit you get for it.

Automation is why the waiting exists at all in its current shape. Anyone who wants an instant generic answer already has one for free, so the human market has narrowed to buyers who want a specific person and want them now; what automated scoring does in seconds is the reason your response time competes against a machine's, and the tools packaging it are what a buyer tried before opening your profile.

Platform design also moves this number, sometimes sharply. Response-time metrics that feed ranking, auto-decline timers and enquiry routing all decide how much of your evening a marketplace consumes, and they differ between them - what Rate Cock asks of listed judges is the current statement of one such set, and it is the sort of thing worth comparing before you add a second marketplace rather than after.