Cost of Employee Turnover: An Auditable Worksheet

The cost of employee turnover is the financial and operating effect of an employee leaving during a defined period. Calculate it from local evidence, not a universal salary multiplier. Keep incremental cash, diverted staff capacity, and verified lost output separate, then reconcile offsets and uncertainty before presenting a total.

Define the departure before calculating cost

Start with a population and a decision. Total separations, voluntary quits, dismissals, retirements, internal transfers, and the end of temporary work are different events. Their replacement need and cost can differ.

The U.S. Bureau of Labor Statistics JOLTS definitions separate quits, layoffs and discharges, and other separations. JOLTS is useful external context, but its establishment-level definitions do not automatically match an employer's HR system. Create a local data dictionary before comparing rates.

For a simple period rate, use:

turnover rate = relevant leavers during the period / average active employees during the period x 100

State how average employment is calculated, who counts as active, and whether internal moves are included. Also report the count. A rate of 20% could mean two leavers from ten employees or 2,000 from 10,000, and those situations need different analysis.

Use separate views where the decision requires them:

• voluntary turnover for retention work;
• total separations for replacement workload;
• regretted departures under a documented business definition;
• early-tenure departures for onboarding review;
• role, site, tenure, or contract segments with enough observations for responsible interpretation.

Do not label every departure preventable. The cost model describes exposure. It does not establish why someone left or whether a particular intervention would have changed the decision.

Use three ledgers instead of one multiplier

A single percentage of salary hides assumptions and invites double counting. Build three ledgers.

Add a fourth section for offsets and timing differences. Vacancy payroll not paid, delayed spending, recovered work, insurance, and later revenue may change the accounting view. Keep them visible rather than quietly netting them into one headline.

The CIPD turnover and retention factsheet recommends understanding why people leave and measuring turnover consistently. Its retention guidance identifies recruitment, training, and knowledge loss among the practical effects worth examining. Those categories are a starting checklist. Your finance policy decides which entries become recognized expense, management information, or unquantified risk.

Build an auditable cost register

Use one row for each cost item:

For staff time, use a loaded rate that matches the decision. The BLS Employer Costs for Employee Compensation separates wages and salaries from benefits and publishes aggregate U.S. estimates. It is useful for understanding the components of employer cost, but an employer should use its own role, location, and benefit data. An aggregate national rate is not the cost of a specific manager hour.

Record knowledge and continuity effects even when they cannot be monetized. Examples include an undocumented customer history, a certification gap, or a delayed handover. Give each a mitigation owner and review date. An unpriced risk is more honest than a large invented amount.

Work through a fictional calculation

This example is fictional. It demonstrates arithmetic and accounting choices, not a benchmark or customer result.

A U.S. service team replaces one experienced employee. The review window runs from notice through the end of the successor's first four weeks. Finance and operations agree on the following evidence.

Ledger 1: incremental cash

The example excludes standard successor salary and equipment that would have been paid if the incumbent had stayed. It includes only additional cash caused by the replacement event under the fictional company's policy.

Ledger 2: diverted paid capacity

These are loaded fictional rates. The $3,188 is the economic value of paid capacity redirected from other work. It is not an extra cash payment and is not a cash saving available from a retention action unless staffing or payroll actually changes.

Ledger 3: verified lost output

During the six-week vacancy, the role would normally complete 240 service cases. Existing staff and the purchased temporary cover complete 204, leaving 36 cases. Finance verifies that 20 were cancelled at $180 contribution each. The other 16 were deferred and receive no value in this model.

vacancy contribution loss = 20 x $180 = $3,600

During the four-week ramp, the documented target is 160 cases and actual output is 112. Of the 48-case difference, finance verifies 30 cancellations at the same contribution. The remaining 18 are deferred and remain unpriced.

ramp contribution loss = 30 x $180 = $5,400

verified lost contribution = $3,600 + $5,400 = $9,000

The $180 is defined as revenue per case less case-specific variable costs avoided when a case is cancelled. It excludes incumbent and successor payroll, the temporary-cover invoice, and every other cash-ledger item above; those are handled separately in the reconciliation. The 204 vacancy cases already delivered through existing staff and temporary cover are excluded from lost output. The manager, HR, and peer hours in the capacity ledger are support activities and are not also valued as missed service cases.

Reconcile without hiding the lenses

The three reported figures are:

• incremental cash outlay: $5,000;
• diverted paid capacity: $3,188;
• verified lost contribution: $9,000.

The fictional vacancy also creates $7,800 of payroll timing variance because salary and benefits were not paid for part of the six weeks. Finance therefore records a provisional period P&L view of:

$5,000 cash outlay + $9,000 lost contribution - $7,800 payroll timing variance = $6,200

The $3,188 capacity value stays outside that P&L subtotal because those salaries were already paid. The $6,200 is not a universal turnover cost. It depends on the chosen period, contribution definition, treatment of payroll, and what happens to deferred cases. A longer review may add further coverage or overtime costs and establish whether deferred work was recovered or lost. Update those entries rather than assuming the current subtotal is final.

Stress-test the answer

Before a CFO or CHRO uses the result, ask:

1. Would this cost disappear, move, or remain if the employee stayed?
2. Does the quantity come from a record or an assumption?
3. Is the unit value appropriate for this role and location?
4. Does any invoice, hour, or output unit appear in two ledgers?
5. Is saved payroll shown as an offset rather than called a retention benefit?
6. Are recovered and deferred units updated later?
7. Are knowledge, customer, quality, and safety risks described without invented money?
8. Can another reviewer reproduce every subtotal?

Run lower, working, and higher cases only by changing named assumptions. Do not apply an arbitrary low and high salary multiplier to the final total. Sensitivity should reveal which evidence matters, not make the business case look larger.

Turn the model into a decision

Rank patterns by evidence and controllability, not just gross cost. A high-cost specialist departure may be isolated. Repeated early-tenure departures with the same onboarding failure may offer a clearer action. Keep cost, frequency, cause confidence, and the organization's ability to act as separate columns.

Use the employee retention strategy worksheet to choose owners and review points. Use the turnover reduction validation guide to test whether an action was delivered and whether outcomes moved without claiming causation from a simple before-and-after comparison.

Where Lontra may fit

Lontra can be evaluated when the cost review reveals an unanswered employee or manager context question, such as a recurring onboarding barrier or local work friction. Review the current platform capabilities, then use one trial campaign with up to 30 invitations over 60 days and no credit card to examine that question.

Conversation evidence supports human review. It does not calculate turnover cost, predict a named departure, or prove that an action prevented one. Managers receive a brief rather than raw employee responses, and aggregate HR views require at least five respondents.

Sources

• U.S. Bureau of Labor Statistics: JOLTS definitions
• U.S. Bureau of Labor Statistics: Employer Costs for Employee Compensation
• CIPD: Employee turnover and retention factsheet
• CIPD: Retention guidance for people professionals