Syntha Lab RU

Cost of staff turnover: what replacing a seller costs

Adds the direct and hidden costs of replacing one employee (recruiting, training, vacancy, reaching full productivity), multiplies by the number leaving in a year and shows what pay rise would pay back if it lowers turnover.

Store staff and productivity For whom: Store managers and HR.

Calculate on your data →

At a glance

What question does it answer?
What a seller leaving costs and what raise would pay back.
What data is needed?
Number of sellers, turnover, revenue per seller and margin.
How to read the result?
One replacement costs 187,675, 81% of it hidden margin losses. At turnover of 40% and eight sellers that is 3.2 leavers and 600,560 a year. A 10 pp fall in turnover saves 150,140; a raise of up to 1,203 a month per person pays back. Example on the default values: Cost of one replacement: RUB 187,675; Hidden part of a replacement: 80.9%; Leavers a year: 3.2 people; Turnover a year, cost: RUB 600,560. The figures are illustrative: enter your own to see your result.

Why and how to reason

A salesperson’s departure seems free: no salary is paid. But replacement costs money and most of it is invisible in reports. Direct costs: recruitment 25,000 and training 24 hours × 450 = 10,800. Hidden: while the vacancy is open for 21 days the store loses part of its sales: 750,000 ÷ 30 × 21 × 30% × 45% = 70,875 of margin. A newcomer works at 60% productivity for two months: 750,000 × 2 × (1 − 0.6) × 30% × 45% = 81,000. Replacing one employee totals 187,675, of which 81% is hidden. With turnover of 40% of eight salespeople, 3.2 people leave a year and that is 600,560, or 0.83% of revenue. A measure that lowers turnover by 10 pp saves 150,140 a year: that is 1,203 a month per salesperson which can be added to pay, with contributions, and the measure still pays.

When to use it

Reach for it when discussing a pay rise, an onboarding program and retention bonuses. It suits a retail chain and a brand with stores. It does not suit a small team without turnover statistics: the estimate is then noisy. Assumptions: the share of lost sales during a vacancy and a newcomer’s productivity are your estimates; severance and compensation are not included; contributions on the increase are counted separately.

How to read the result

First “cost of one replacement” and its hidden part: it is the hidden one that is usually several times the direct. Compare “cost of turnover per year” with your pay budget: if it is 1% of revenue or more you have money for retention. “Pay increase that will pay back” shows the threshold: if the increase is smaller and you really lower turnover, the measure pays. But check the hidden part against experience: when someone left, did the point’s sales fall more than in neighboring ones? The decision: which measure (pay, schedule, training, mentor) and what effect you expect from it.

Common mistakes

Taking only direct recruitment costs and treating replacement as cheap. Assuming the measure will lower turnover by the whole amount claimed: part leave for reasons pay does not affect. Not including severance and holiday compensation. Comparing turnover in and out of season without adjustment.

What data is needed?

  • Number of sellers, turnover, revenue per seller and margin.
  • Recruiting, training, vacancy days, the newcomer’s ramp-up and the share of sales lost.

Inputs

FieldUnitValue in the exampleMeaning
Sellerspeople8Average number of sellers in the store.
Turnover a year%40Share of leavers in a year of the average headcount. Use your own from HR records. An example: enter your own data, there is no single standard in Russia.
Revenue per seller a month₽750,000Excl. VAT: store revenue ÷ number of sellers.
Gross margin%45Margin share in revenue excl. VAT.
Recruiting₽25,000Ads, an agency, the manager’s time on interviews. An example: enter your own data, there is no single standard in Russia.
Training hoursh24Mentor and newcomer together. An example: enter your own data, there is no single standard in Russia.
Cost of a staff hour₽450With contributions; see “Payroll and the full cost of an employee”. An example: enter your own data, there is no single standard in Russia.
Vacancy daysdays21From the leaving to the replacement starting. An example: enter your own data, there is no single standard in Russia.
Share of sales really lost%30The part of the employee’s sales not picked up by colleagues in a vacancy and while the newcomer ramps up. An estimate, replace with yours.
Months to full productivitymonths2How long a newcomer works below an experienced seller. An example: enter your own data, there is no single standard in Russia.
Newcomer’s productivity in that time%60On average over the ramp-up, against an experienced seller. An example: enter your own data, there is no single standard in Russia.
Fall in turnover from the measurepp10How many percentage points of turnover you expect a raise or program to remove.
Contributions on the raise%30The general contribution rate of 30% (2026); enter your own.

How it is calculated

StepFormula
Loss in the vacancyrevenue per seller ÷ 30 × vacancy days × lost share × margin
Loss in the ramp-uprevenue per seller × months × (1 − productivity) × lost share × margin
Raisefall in turnover × sellers × replacement cost ÷ (sellers × 12 × (1 + contributions))

Worked numbers

Default values from the app; in the interactive version you enter your own.

Cost of one replacement
187,675 ₽
Hidden part of a replacement
80.9 %Margin lost in the vacancy and the ramp-up: not visible in the income statement.
Leavers a year
3.2 people
Turnover a year, cost
600,560 ₽
Turnover in revenue
0.8 %
Saving from the measure a year
150,140 ₽
A pay rise that would pay back
1,203 ₽ a month per personSaving ÷ (sellers × 12 × (1 + contributions)). Above this raise the measure does not pay back through turnover.

What one replacement is made of

ItemAmountShare
Recruiting25,000 ₽13.3 %
Training10,800 ₽5.8 %
Margin lost in the vacancy70,875 ₽37.8 %
Margin lost while the newcomer ramps up81,000 ₽43.2 %

The hidden part depends on the share of sales really lost: this is an estimate, not an accounting figure. Test it on periods when someone left: did the store’s sales fall more than the neighbours’?

Severance, holiday compensation and other payments on leaving are not included: add them to recruiting if they are regular for you.

Norms (a guide as of 13.11.2026, check against the act text): Labour Code. Check with an accountant and a lawyer.

Nuances and limits

  • The share of lost sales is an estimate: test it on past departures.
  • Add payments on leaving to recruiting.

Terms used

Calculate on your data →