When a volunteer stops, the explanation that circulates is nearly always the same: "they ran out of time". It is a convenient explanation because it is true on the surface and asks nobody to change anything.
Underneath there is usually something else. Someone who joins an organisation has already decided to give time: if they stop, it is generally not because the time ran out, it is because that time started costing more than it gave back. And a large part of that cost is organisational.
The four kinds of friction visible in a rota
Uncertainty. Knowing your shift three days ahead means being unable to organise the rest of your life around it. Anyone with a family or shift work of their own cannot sustain it for long, and they will not say so: they will fade out.
The pointless shift. Turning up and finding you were not needed, or that three people are doing one person's job, is more demotivating than a heavy shift. Donated time that gets wasted is the most effective insult an organisation can deliver without noticing.
Visible imbalance. It is not the load in itself: it is the load relative to others. Someone covering three nights a month while others cover none stops feeling generous and starts feeling used, and the second feeling is not corrected by thank-yous.
Changes communicated late. A shift moved the day before costs the person receiving it far more than it costs the person moving it. Repeated, it communicates something precise: that the time of those doing the work is worth less than the time of those organising it.
None of these four is about motivation. All four depend on how the rota is built.
What can genuinely be changed
Publish early and leave it alone. A four-week rota, published and stable, is worth more than any recognition scheme. Stability is the form of respect an organisation can offer without spending anything.
Declare the minimum cover and stop there. If a service needs two people, calling four "to be safe" does not add safety: it consumes two volunteer shifts and hands them back empty.
Make the counts visible. Shifts worked, nights, holidays, per person. Not to reward: to make impossible the feeling, often justified, that the load is distributed at random.
Ask before assigning the difficult slots. The difference between "I have put you on Sunday night" and "can you do Sunday night?" is the whole difference between an assignment and a request, and in an organisation without salaries it is the only correct form.
The window where you can still act
Volunteers about to stop signal it long before they say it, and they signal it measurably: declared availability narrows, swap requests increase, shifts accepted at the last minute turn into refusals.
They are the same signals that apply to employed staff, gathered in burnout signals you can read in a rota. The difference is that here the window to act is narrower: an unhappy employee stays because they have a contract, a volunteer does not.
The sum worth doing
Training a new volunteer costs more time than you save by asking an experienced one for one more shift. It is a trivial calculation and hardly any organisation does it, because the cost of training is visible and the cost of attrition is not.
Those who do it usually stop asking the same people every time. Which is also the most effective way of not having to replace them.
In Sked Solve rotas are published in advance, per-person counts are visible and everyone declares their own availability: three things that reduce friction without asking anything of anyone. See how it works at Sked Solve.