A sous chef hands in notice in October. The venue in Causeway Bay budgets for it the way most venues do: a recruitment fee, a fortnight of overtime while the section is covered, a line in the November payroll for whoever arrives. The figure the operator writes down runs to a few tens of thousands of Hong Kong dollars, and it feels manageable.
It is not the figure. It is the invoice, which is a different thing. The invoice is the part of the cost that arrives with a reference number attached, and it is reliably the smallest component of what the departure will consume before the section runs properly again.
What follows is the rest of the arithmetic. None of it is exotic and all of it is knowable in advance. Operators who run the number once tend to make different decisions about assessment, about probation, and about the retention spending they had been treating as optional.
The fee is the smallest line
Start with the vacancy. From the day notice is served to the day a replacement works unsupervised, the work does not stop being necessary. It is absorbed by overtime, by casual cover at a premium, or by colleagues carrying an extra section. That last one never appears on an invoice. It shows up later as the second departure nobody connects to the first.
Then there is the training already spent on the person leaving. Every induction hour, every shift a senior member of the team spent shadowing them, every mistake absorbed while they learned the venue. That was an investment in an asset that has now walked out. It is not recoverable, and it will be spent again on the replacement.
Then the ramp. A competent hire is not productive on day one, and in a kitchen or on a floor with established service rhythms, full competence takes considerably longer than induction suggests. Throughout that period the venue pays a full salary for partial output, and the people around the new hire are watching them instead of their own sections.
Finally, management time. Recruiting, interviewing, inducting, and supervising pulls a head chef or a general manager out of the operation for hours already committed elsewhere. In a small venue this is often the largest component of all, because the person being pulled out is the one holding the operation together.
Where the published range comes from
Research on replacement cost puts the total at between 30 and 200 per cent of the role's annual salary. SHRM's 2025 benchmarking places it at 50 to 200 per cent; Gallup's breakdown by seniority starts at roughly 40 per cent for frontline roles and reaches 200 per cent for managers. A range that wide is not weak data. It reflects the fact that the same event costs radically different amounts depending on seniority, local scarcity, and how long competence takes in that particular venue.
Cornell's Center for Hospitality Research is the most granular source for the sector. The team studied replacement costs across 33 US hotels, breaking each departure into components, and found the same pattern in every property: lost productivity during the new hire's ramp to competence was the single largest cost driver, accounting for 55 to 69 per cent of the total. Training and orientation made up the next largest share. The finding explains why the invoice is so consistently misleading. The majority of the cost never arrives as a bill.
Hong Kong data brings the range closer to home. In 2026, HR Magazine Hong Kong itemised replacement costs at three salary tiers. A frontline worker earning HKD 18,000 a month costs HKD 25,000 to HKD 40,000 to replace. A mid-level hire at HKD 40,000 a month runs to HKD 275,000 to HKD 463,000 once recruitment fees, HR time, training, and productivity ramp are counted separately. A senior manager at HKD 80,000 a month can reach HKD 1.2 million.
Applied to a real salary, the range stops being abstract. A restaurant general manager paid HKD 540,000 a year sits somewhere between HKD 162,000 and HKD 1,080,000 in total damage. Two venues can lose the same job title in the same month and face costs that differ by a factor of five. Time to competence and local scarcity do more work than the salary figure.
The pre-opening multiplier
A departure during pre-opening belongs in its own category, because the currency is not only money. It is schedule.
Someone hired four months before opening is not simply doing a job. They are building menus, establishing supplier relationships, configuring systems, and hiring the team beneath them. When that person leaves during fit-out or in the first weeks of trading, the replacement does not inherit a working position. They inherit a half-built one, with no memory of the decisions that produced it and no relationship with suppliers who were persuaded to accommodate a business that did not yet exist.
This is why experienced operators commonly staff an opening at 10 to 15 per cent above target headcount. It is a rule of thumb rather than a published benchmark, but the logic is sound: the team trains in one cohort, the attrition that follows an intense opening is budgeted rather than reactive, and the venue avoids recruiting mid-service during the weeks when its reputation is being set. The overage costs money. It costs less than hiring into a live operation while the first reviews are being written.
It is also why the person who opens a venue and the person who runs it are often two different people. Planning that handover in advance turns a predictable departure into a managed transition, which is the cheapest form a departure can take.
Running the number for your own venue
The model takes an afternoon to build. Start with the annual package for the role rather than the base salary, including employer MPF contributions and allowances. Then estimate the vacancy period honestly, measured from notice served to unsupervised competence rather than to start date. The gap between those two dates is where most of the cost lives.
Cost the cover at what it actually costs: the overtime or casual rate, not the departing person's hourly cost. Add recruitment spending, which in Hong Kong means the statutory agency commission of 10 per cent of first-month salary for a frontline placement, or 15 to 25 per cent of annual base for a role that requires retained search. Add the training written off, valued at the hours senior staff spent rather than at a notional training budget. Then add the ramp: the proportion of full productivity the replacement delivers across their first months, multiplied by their salary over that period.
The last line is management time pulled out of the operation, valued at what those hours produce when they are spent on the operation instead. Most operators stop before it. It is usually the one that turns an uncomfortable number into a serious one.
The output is not there to alarm anyone. It exists to reprice two decisions that come before the hire and are usually made on instinct. The first is how much structured assessment a role justifies, because an interview question that reveals how somebody handles a disputed order costs a fraction of one failed hire. The second is how much retention spending a role justifies, whether that means rosters published weeks in advance, visible progression, or managers who correct privately rather than in front of the section.
Neither competes with the cost of a departure. Both are cheaper, usually by a wide margin, and the arithmetic above is the only thing needed to prove it to whoever controls the budget.
