Most hospitality businesses measure turnover as an annual figure. It's the wrong lens. Split that number by length of service and a very different picture usually appears: a large share of leavers never made it past three months.
That matters, because early leavers are the most expensive kind. You paid to advertise the role, you paid a manager to interview, you paid for the induction, you paid wages while the person was still slow — and you got almost none of the productive period that was supposed to repay all of it. Then you start again.
It's also the most fixable kind of turnover. Someone who leaves after four years has usually run out of road. Someone who leaves after four weeks has almost always been failed by a process.
Why the first weeks decide it
New starters form a working theory about an employer far faster than most operators expect. Within about two weeks they've answered three questions for themselves:
- Do these people know what they're doing?
- Does anyone here care whether I do well?
- Can I picture myself here in a year?
Nobody asks those out loud. They get answered by whether the rota was ready, whether anyone knew they were starting, whether their trainer had actually been told they were training someone, and whether a manager spoke to them by name in week one.
Get those wrong and you don't get a formal complaint. You get someone who quietly keeps their options open, takes the first offer that comes along, and gives "personal reasons" on the way out.
The problem with how most inductions work
The typical hospitality induction is a fire-hose morning followed by silence. Paperwork, a tour, fire exits, a login, a uniform, then straight onto the floor to shadow whoever is on shift. After that, nothing scheduled. The next real conversation about how it's going happens either when something goes wrong or at a probation review nobody diarised.
Two things break here. The first is volume — nobody retains a shift's worth of new information in one morning, so "we covered that on day one" is rarely true in any useful sense. The second is the gap. Between day one and the probation meeting there is usually no checkpoint at all, which means a struggling new starter has no natural moment to say so. They work it out alone, decide they're bad at the job, and leave before anyone notices they needed help.
A 90-day structure that actually works
The fix isn't a bigger induction. It's spreading the same content over a longer arc and putting scheduled checkpoints in it. This is the single highest-return change most operators can make to staff retention.
Before day one. Confirm the start date in writing, say who to ask for and what to wear, and tell the team a new person is starting. Have the uniform, locker, login and rota ready. None of this is difficult; all of it signals competence.
Day one. Keep it short and human. Introductions by name, the tour, the genuinely urgent safety and legal essentials, and one clear statement of what good looks like in this role. Then let them observe rather than perform. Do not schedule a new starter as a working body on day one to cover a gap — it is a false economy that shows up in your turnover figures a month later.
Week one. Assign a named buddy — a specific person, not "the team". The buddy's job is to be the one who answers the stupid questions, and they should know they've been given that job. End the week with a ten-minute check-in from a manager. Two questions do most of the work: what's been confusing, and what do you need.
Day 30. A proper sit-down. By now they know enough to have real opinions, and they're still close enough to their first impressions to tell you what was badly explained. This is also the point to give genuine feedback on how they're doing — specific, not "you're doing great" — because a month of silence reads as indifference.
Day 60. Shift the conversation forward. What do they want to get better at, and what would the next step here look like? This is where you find out whether you've hired someone who wants a career or someone passing through, and it's early enough to influence which one they become.
Day 90. Confirm the outcome properly rather than letting probation lapse by default. Either they've passed, and you say so and mean it, or there's a specific gap and a plan. Letting a probation period quietly expire teaches people that nothing here is really tracked.
The checklist most operators are missing
Underneath the checkpoints, one document does a lot of unglamorous work: a written statement of what competent looks like in each role. Not a job description — a list of the things someone should be able to do unsupervised by week two, week six and week twelve.
Without it, "trained" means whatever the trainer remembered to cover that day, standards drift between shifts, and nobody can say whether a new starter is behind. With it, training becomes checkable, feedback becomes specific instead of personal, and two managers give the same answer about whether someone is ready.
What it costs you not to do this
Structured onboarding has no line in most P&Ls, so it loses every argument against something that does. That's a measurement problem, not a value problem.
Track one number for a quarter: the share of your leavers who were with you under 90 days. If it's high, you don't have a recruitment problem or a labour market problem — you have an onboarding problem, and it's costing you a full replacement cycle every time it repeats. Most of the fix is scheduling and attention rather than spend.
At WHY Hospitality we build this structure with operators — the standards, the checkpoints and the manager conversations that hold it together — usually as part of wider work on retention and engagement. If your leavers are going out of the door faster than your induction can bring them in, that's the place to start.