Chapter 17 — Key Takeaways

The core claims

  1. Turnover is a real expense with no line on the P&L. At Bellwether it runs about \$38,070 a year — 2.5% of revenue, 7.6% of the labor line, more than half the annual debt service — hidden across advertising, management salary, hourly labor, and food cost.

  2. One line cook costs about \$2,180 to replace, and advertising — the component operators name first — is \$150 of it. The largest single component is unproductive training wages.

  3. Compute it for your own positions. Until turnover is a number an operator can state, nobody funds the thing that prevents it.

  4. The cheapest sourcing channels produce the best retention. Referrals and returning alumni beat job boards on both cost and tenure. Boards are where you go when the pipeline has already failed, and you pay for that failure twice.

  5. Post the wage. A posting without a number selects disproportionately from candidates who have no alternatives — the opposite of what you want. It is also increasingly unlawful.

  6. Unstructured interviews predict poorly because they measure rapport and similarity rather than performance. The same scored questions, in the same order, for every candidate, work better.

  7. Score before you discuss. Two managers who confer first produce one opinion held by two people.

  8. Pay for the stage. Four hours at roughly \$88 is about 4% of a replacement cost and the cheapest insurance available against both a bad hire and a wage-and-hour claim.

  9. The correction test is the most informative ninety seconds of a stage — not because you want compliance, but because you are learning how information will travel between you and this person.

  10. Most turnover money is lost in the first ninety days, at three specific moments: day one, week two, and day thirty. All three sit before break-even, so those departures return nothing.

  11. Hiring is necessary and nowhere near sufficient. Early departures point at hiring and onboarding; late departures point at schedule, wage, path, and culture — which is Chapter 21.

The arithmetic

$$\text{Turnover rate} = \frac{\text{separations in the period}}{\text{average positions}}$$

Cost of turnover, nine components: advertising · management screening time · the stage · orientation · unproductive training wages · the trainer's diverted attention · reduced output · learning-curve waste · vacancy overtime.

Bellwether's ledger: 29 non-owner positions → ~26.8 separations\$38,070.

The process, in order

Step The discipline
Description six parts, wage posted, process and timeline stated
Sourcing referrals first, bench second, boards last
Screening structured, scored, same questions, scored before discussion
Stage paid, identical tasks, watch mise / pressure / correction / the close
References one question — "would you hire them again?"
Background consistent written process; FCRA and fair-chance obligations
Offer in writing; at-will stated carefully
I-9 Section 1 day one, Section 2 within three business days, employee picks documents
Onboarding day one prepared · named owner in week one · real check at week two · decision at day thirty

Key terms

turnover rate · cost of turnover · job description · sourcing channels · structured interview · stage · working interview · reference check · background check · Form I-9 · at-will employment · onboarding

What you should be able to do Monday morning

Compute the fully loaded replacement cost of one position in your restaurant, multiply it by that position's actual annual separations, and put the number in front of whoever controls the wage budget. Then ask the diagnostic question: of the people who left last year, how many left inside ninety days? If most did, the problem is in this chapter. If most didn't, the problem is in Chapter 21.