Here is a thought experiment worth running: take your store's last ten scheduling "fires" (sudden absences, desperate swaps, the shift refused with drama) and ask how many were genuinely unforeseeable. In most cases the answer is uncomfortable: half of those fires were constraints the person had known about for weeks. The Tuesday course, the children on alternate weeks, the medical appointment booked a month ago. Nobody had asked, so nobody knew, so the plan ran straight into them.
That is what shift preferences are, viewed without sentimentality: operational information you already own and are not using.
Constraints versus wishes: the distinction that makes it all manageable
The typical owner's resistance ("if I ask for preferences, I'll have to honour them all") comes from a confusion worth dissolving immediately, with the team too: constraints are not wishes. "I have a course on Tuesdays" is a constraint: ignoring it produces an absence or a resignation. "I'd rather never close" is a wish: honoured when possible, declined when not.
Once that distinction is made, the numbers become small and tractable: the genuine constraints of a six-person team fit on half a page, and a plan that respects them takes the same effort to build as one that ignores them. It just does not blow up.
Asking once is not enough
Preferences age: the course ends, the family situation changes, the part-timer starts university. The form filled in at hiring and never updated is a two-year-old photograph hung up as if it were a mirror. You need a flow, not a census: one place where each person updates their own availability when it changes, and a monthly deadline before planning ("updates by the 20th").
The detail that changes everything: the workers update, you do not chase. If collecting preferences means making six phone calls a month, you will not do it. If it means opening a page that is already filled in, you will.
When you cannot accommodate: the explained no
Collected preferences create an expectation, and it is right to govern it immediately: collecting is not promising. The honest pact sounds like this: "hard constraints we always respect; wishes we respect when possible, and when it is not possible we tell you why". The explained no ("three of you want Saturdays off, we rotate") preserves trust; the silent no erodes it, and the impossible yes promised for a quiet life destroys it.
There is also a balance to watch: that the accommodated wishes are not always the same people's, typically the most insistent. Here too, a visible tally is the only antidote to bookkeeping by feelings.
The return on investment is measurable
Whoever keeps track finds, after a few months of preferences done properly, the same three numbers falling: last-minute absences, emergency swaps, early departures. Not by magic: because the schedule built against people's lives was the factory of those three numbers. You called it flexibility you were owed; it was emergency maintenance on a plan drawn up blind.
In Sked Solve, availability and preferences are collected through the app and the schedule generator respects them on its own: the site shows how it works.