Almost everybody in this field gives work away. It is how anybody finds you, it is how a body of work becomes visible, and it is the normal entry cost of building an audience. It is also frequently done without any thought about what it is supposed to achieve, which is where it stops working.
The useful framing is that free work buys attention. Attention is genuinely valuable and it is not income, and the gap between the two is where most of the disappointment in this area lives.
Attention has to convert into something
A large audience with nothing to buy, nothing to join, and no way to support the work produces very little. That sounds obvious and it is the commonest structural mistake: years of free posting with no destination for the readers who wanted to do more.
Something has to exist at the end of the path, whether that is a subscription, a shop, commissions or a book. The routes are described in the piece on where the money comes from, and any of them work. Having none of them does not.
Free work as a sample
The most reliable model is that free work demonstrates what paid work is like. It should be genuinely good and genuinely representative, since a sample that oversells produces disappointed buyers and one that undersells produces none at all.
This also settles the perennial question of whether to give away your best work. Give away work that is characteristic and good; keep the things that require the most labour, or that only make sense as part of a larger whole, on the paid side.
What giving everything away costs
If the complete work is always freely available, supporting it becomes an act of generosity rather than an exchange. Some readers will do it and most will not, and that is not a moral failure on their part; it is what happens when there is no difference between supporting and not.
The workable versions preserve some difference: earlier access, higher resolution, collected editions, extras, or simply the ability to keep the work going. Any of these is enough, and readers respond well to an honest arrangement stated plainly.
Timing is a lever
Releasing paid first and free later is the least contentious arrangement available. Everybody eventually gets everything, supporters get it sooner, and nothing is permanently withheld. It converts well and it avoids the resentment that hard paywalls can generate.
It also suits serialised work particularly well, since the schedule provides a natural rhythm. The serialisation considerations in the piece on collected volumes apply here as a business structure rather than only a formal one.
The visibility argument has limits
Free work does create visibility, and visibility does not scale indefinitely into income. Past a certain point additional attention adds very little, because the people who were going to support the work already have. Continuing to increase free output past that point mostly increases the workload.
Recognising where that point sits, for your own audience, is worth more than any general rule. It is usually visible in the numbers: a month where attention rose substantially and support did not is the signal.
Free work for other people
Requests to work unpaid for somebody else's project are a different question entirely, and the answer is usually no. Exposure offered by a third party is rarely delivered, and the arrangement transfers your labour into their asset. Free work should be work you own and publish yourself.
The exceptions are genuine collaborations where everybody is unpaid and everybody shares the result, and even then the terms belong in writing, as described in the piece on collaboration.
Deciding how much to give away
Decide what the free work is a sample of, make sure there is somewhere for interested readers to go, preserve some real difference for supporters, prefer timing over hard walls, watch for the point where attention stops converting, and keep free work your own.
Our own publishing position is described in the editorial principles.