The rigorous anchor. The most defensible synthesis of turnover cost is the Center for American Progress review by Heather Boushey and Sarah Jane Glynn (2012), which pooled 30 case studies drawn from 11 empirical research papers. Its headline finding: it typically costs about one-fifth (≈20%) of a worker’s annual salary to replace them. More precisely, the median cost was 21% of annual salary across the 27 non-executive/non-physician case studies; 16% for jobs paying under $30,000; and 20% for jobs paying under $75,000 (which covers ~9 in 10 U.S. workers). The full range across all 30 studies ran from 5.8% up to 213% of salary, with the 213% figure attaching to highly paid executive/specialist roles that “skew the data upwards.” This is the number to lead with because it is a transparent synthesis of peer-reviewed empirical work, not a survey. (Note: the CAP brief’s body text says the studies span 1992–2007 while its methodological appendix says 1992–2012 — a minor internal inconsistency worth flagging.) Critically, CAP notes that only 2 of the 11 papers included indirect (soft) costs at all — so even this rigorous estimate is conservative on the soft component.

Hard vs. soft components. CAP splits cost into direct costs (separation/severance, exit interviews, higher unemployment taxes, temporary coverage/overtime, advertising, agency/search fees, screening, interviewing, background checks, onboarding and training) and indirect costs (lost productivity of the departing employee, lost productivity while covering the vacancy, ramp-time errors/waste as the new hire learns, reduced morale, lost clients and lost institutional knowledge). The hard costs are invoiceable and relatively easy to measure; the soft costs are where the multiples balloon and where rigor is weakest. This is the structural point Frank’s framework gets right: the visible recruiting bill is the tip of the iceberg.

Where Frank’s 3x actually lands. Frank Newman’s Rule of Three states that losing an employee costs roughly three times (300%) their annual salary. Placed honestly against the literature, 3x sits at or above the very top of published estimates. The rigorous CAP synthesis peaks at ~213% for executives and clusters at 16–21% for typical roles. The vendor/practitioner end runs higher: the Society for Human Resource Management’s commonly repeated rules of thumb include “six to nine months of salary” and “50–200% of annual salary” depending on seniority, and SHRM Foundation chair-elect Edie Goldberg is quoted in SHRM’s own 2022 article estimating total hire cost at “three to four times the position’s salary” — explicitly a practitioner estimate (Goldberg attributes ~30–40% to hard costs and ~60% to soft). So 3x is defensible specifically for senior, specialized, high-revenue-attached roles — e.g., a $100k senior manager whose work is attached to ~$2M of revenue, where prolonged vacancy, lost deals, and lost institutional knowledge plausibly compound past 200% — but it is well above the median for typical roles, where 16–50% is the evidence-based expectation. The multiple must be labelled directional: it is a planning heuristic that rises with seniority, not a measured average.

Ontario/SMB lens. For a 20–200-person K-W employer, the practical takeaway is to scale the multiple to the role. Use ~20–30% of salary for front-line and administrative turnover; reserve high multiples (100–300%) for senior leaders, specialized technical staff, and client-facing revenue generators. The dollar base should be the fully-loaded cost, not bare salary (see Note 2). Cost-rises-with-seniority is industry-consensus; the specific multiple is directional.

Anchor: Newman’s Rule of Three. Links to Note 2 (fully-loaded cost), Note 3 (bad hire), Note 4 (vacancy), Note 5 (fractional vs FTE).