Employee Retention Rate vs. Turnover Rate: What Each Metric Tells You

Retention follows the workforce you started with

A cohort-based retention rate starts with the employees who were present at the beginning of the period and asks how many of that group remained at the end. The denominator is the starting workforce, and employees hired during the period do not become retained members of that original cohort.

This makes retention particularly useful for understanding stability among a defined population. HR teams can use the same approach for a department, location, job family, tenure group, or other group as long as they define the cohort consistently.

Turnover measures departures against workforce size

This site's selected-separation-rate scenario compares documented voluntary and involuntary departures with a two-point average headcount for the period. Other turnover definitions may use different populations, events, or denominators, so the formula must be named rather than presented as universal.

Because turnover may use average headcount and retention follows a starting cohort, the two rates do not always add to 100 percent. Hiring, transfers, reorganizations, and different population definitions can all affect the relationship.

Use both metrics before choosing a response

Start with retention when the question is whether a defined starting workforce stayed. Use turnover when the question is how many departures occurred relative to average workforce size, then separate voluntary and involuntary movement where that distinction is meaningful.

If either measure changes, examine the time period, employee population, hiring activity, and data definitions before treating the change as an employee-experience conclusion. Combine the measures with turnover-cost and workforce-planning analysis to understand operational exposure as well as movement.

Why this matters

• Prevents teams from treating retention and turnover as interchangeable labels.
• Makes the role of new hires and average headcount explicit in workforce reporting.
• Creates a clearer starting point for retention, recruiting, and capacity decisions.

Method, scope, and sources

Cohort retention = starting-cohort members continuously employed through the stated end date ÷ members in that starting cohort × 100. Both compact and advanced views exclude later hires from retained members. The advanced view also displays the arithmetic complement as starting-cohort non-retention and shows ending headcount only as context.

Define continuous employment and the treatment of transfers, leave, status changes, contractors, acquisitions, and separated-and-rehired workers. Starting-cohort non-retention is not necessarily turnover or a formal separation rate. Ending headcount must not replace the retained starting cohort, and the result does not establish cause, culture, engagement, or a required response.

Method status: Supported.

Worked example

Hypothetical cohort: 84 people are in scope on June 1 and 75 of those same people remain continuously employed on August 31. Fourteen later hires are excluded. Retention = 75 ÷ 84 × 100 = 89.29%; starting-cohort non-retention = 10.71%. Neither figure is automatically the organization’s turnover rate.

Sources

• Society for Human Resource Management — Data Will Show You Why Your Employees Leave or Stay: Directly states the formula as employees remaining employed for the entire measurement period divided by employees at the start, times 100. It says not to count hires made during the period and permits targeted periods and worker groups.
• International Organization for Standardization — ISO/TS 30421:2021 Human resource management — Turnover and retention metrics: Supports treating turnover and retention as distinct workforce measures and notes interpretation issues, but its public abstract does not supply the calculator formula.

Method, scope, and sources

Both views use a publisher-defined selected-separation rate: included voluntary plus involuntary departures divided by two-point average headcount for the stated period. The advanced workspace adds a comparable prior-period calculation, category shares, a simple monthly/quarterly annualized run-rate, and a separate replacement-cost sensitivity based on a user-entered amount per included departure.

Define the population, dates, separation categories, and headcount method before use. The annualized value is an arithmetic projection, not an observed annual rate. The cost sensitivity is not part of the turnover rate and is not audited cost or savings. Retirements, transfers, deaths, disability exits, rehires, contractors, and internal moves require an explicit classification policy.

Method status: Publisher-defined.

Worked example

Hypothetical quarterly scenario: a defined unit has 100 starting and 104 ending employees, with three resignations and one employer-initiated termination under its documented policy. Selected-separation rate = 4 ÷ 102 × 100 = 3.92%. A simple annualized run-rate would be 15.69%, but it is only a constant-rate projection. Any per-departure cost remains a separate organization-specific sensitivity.

Sources

• U.S. Bureau of Labor Statistics — Job Openings and Labor Turnover Technical Note: Directly supports separations ÷ employment × 100 and the distinct quits, layoffs/discharges, and other-separations taxonomy. It defines quits as voluntary except retirements/transfers, and identifies layoffs/discharges as employer-initiated.
• Global Reporting Initiative / Global Sustainability Standards Board — Standards interpretations: Confirms GRI 401 reporters are not required to use end-of-period employee numbers to calculate turnover rates from 1 July 2018; it does not prescribe the calculator's two-point denominator.

Method, scope, and sources

The compact calculation multiplies included replacement-required departures by entered recruiting/selection, onboarding/training, and other non-duplicative per-departure costs. The advanced scenario adds organization-selected vacancy and ramp-up proxies, using either a salary-based daily proxy or a custom daily position value plus editable duration and percentage assumptions.

Use one defined period, population, currency, replacement policy, and allocation basis. Exclude non-backfills and internal moves unless separately documented. Vacancy and ramp-up values are scenario proxies, not measured revenue or universal productivity loss. Review entries for overlap. Avoided-departure output is gross modeled exposure before intervention cost and unavoidable components, not realized savings, return on investment, or a guarantee.

Method status: Publisher-defined.

Worked example

Hypothetical advanced scenario: two approved replacement-required departures have $2,000 documented recruiting/onboarding cost each and a $150 daily position-value proxy. A 10-day vacancy at 50% exposure adds $750 per departure, while a 20-day ramp-up at 50% productivity adds $1,500. Estimated component exposure is $4,250 per departure and $8,500 total, subject to overlap and source review.

Sources

• Chartered Institute of Personnel and Development — Benchmarking employee turnover: What are the latest trends and insights?: Supports organization-specific component review. It lists resignation administration, recruitment/selection, vacancy cover, and induction/training, plus indirect management/administrative time and direct advertising, agency, and assessment-centre expense.
• U.S. Bureau of Labor Statistics — Job Openings and Labor Turnover Technical Note: Supports defining the included departure population before modeling a per-departure estimate through its quits, layoffs/discharges, and other-separations taxonomy.

References

• U.S. Bureau of Labor Statistics — Job Openings and Labor Turnover Technical Note
• ISO — Turnover and retention metrics