The comparison that decides whether to continue

Your time here is worth doing only if its effective hourly beats the best alternative use of that hour, after fees, tax, unpaid admin and lost benefits. The market band tells you what buyers will pay, not whether the work is worth it.

Your real pricing floor is not a cost figure and it is not the bottom of the band. It is the best alternative use of the hour, adjusted for everything that alternative demands and this one does not.

Build the comparison properly

The naive version divides a monthly salary by 160 and compares it to a per-job fee. It is wrong in both directions at once, and the errors are large enough to flip the answer.

On the employment side, the take-home hourly figure understates the job, because it excludes employer pension contributions, paid leave, sick pay and whatever the employer covers in insurance and equipment. Divide annual take-home by hours actually worked, then add the value of the benefits you would have to replace.

On the self-employed side, the fee overstates the earning, because it excludes platform commission, processing, unbilled admin, tax you must set aside yourself and the hours spent on enquiries that never converted. The route from advertised fee to what lands is set out in how platform fees change your effective rate, and the hours nobody counts are in why your effective hourly rate keeps changing. The same correction applies when the alternative is not a job at all but another flexible gig, which is the ground comparing this to other online work covers.

Run both columns for a real month, not a good week. A month is the shortest period in which the quiet stretches show up, and quiet stretches are most of the difference between the two numbers.

The flexibility premium is real, and it is not infinite

This work has genuine advantages a job does not: no commute, no fixed hours, the ability to stop, the ability to work while something else is happening in your life. Those are worth money, and pretending otherwise leads people to abandon something that was actually serving them well.

The honest way to value them is to ask what you would pay to buy them back. If a job offered the same money and required four hours of commuting a week, what discount would you accept to avoid the commute? That number, whatever it is for you, is the premium you can legitimately subtract from the employment column.

The dishonest way is to treat flexibility as unbounded and use it to excuse any rate at all. Flexibility does not pay rent, and a schedule you control is not much comfort if the income is not enough to control anything else.

The variance discount runs the other way

Employment pays the same amount on the same date. This does not, and the difference is a cost, not a neutral feature.

Irregular income means holding a buffer, which is capital you cannot spend. It means the quiet month is entirely yours to absorb, which the piece on budgeting for the quiet season treats in detail. It means no sick pay - in the UK, Statutory Sick Pay requires you to be classed as an employee - so an illness is a double hit.

Price that as a discount on the self-employed column of somewhere between a tenth and a quarter, depending on how thin your buffer is. That is a judgement, not a measurement, and it is the point where the comparison stops being arithmetic - which is fine, as long as you make the judgement explicitly rather than by accident.

What the comparison is actually for

Not to decide whether to quit. Most people doing this are not choosing between it and full-time employment; they are choosing between it and the marginal hour, which is usually an evening.

That changes the benchmark completely. The relevant alternative for a Tuesday evening is not your salary, it is the other thing you could do with a Tuesday evening: overtime if it is available, a second job if you would take one, or nothing. Where the alternative is nothing, almost any positive rate clears the floor - which is exactly why the marginal-hour framing is dangerous if you let it set your published price.

A statutory wage floor is a useful sanity check on the hourly-work alternative. The UK National Living Wage for workers aged 21 and over has been £12.71 an hour since 1 April 2026, while the US federal minimum wage has stayed at $7.25 since July 2009. An effective hourly below your local floor means the work is paying less than the least an employer may legally offer.

Use the alternative-use figure as a floor for accepting work. Use the market band, described in reading the market band without guessing, to set the price you publish. When the floor rises above the band for a sustained period, that is the signal to reduce hours or stop, and it is a cleaner signal than any feeling about whether it is going well.

The comparison changes shape over time

Time per job falls sharply in the first few months, so the same fee produces a rising effective hourly with no price change at all. Running the comparison in month one and treating the answer as permanent is the most common way people talk themselves out of something that was about to work.

Re-run it at three months, then annually, alongside the yearly rate card review.

Two things from outside the trade belong in the calculation. The first is that buyers commission a person for judgement rather than for output, which is why the floor here has held up better than the floor for generic assessment - the buyer-side account of who gets commissioned and why is the clearest short statement of that. The second is that anything you sell which a widely available tool also produces is competing with a price of zero, and the current state of those tools is the boundary worth checking before you assume your hour is safe. Where the thing you sell is precision rather than opinion, the documented measurement conventions are what makes precision defensible and therefore chargeable.

If part of the answer is that the income is too irregular rather than too small, the fix is usually mix rather than rate, and the platform's own account of how judges are paid is worth reading against your own cadence before you conclude the work does not pay.