Chapter 34 — Key Takeaways

The core claims

1. Controls are not an accusation. They are the systems that let you know what happened. A control does not catch a thief; it produces a record, and a record is what lets you tell a failed compressor from a seven-ounce portion from an uncosted special from an invoice that crept eleven cents a pound. Without records, all of those look identical on a bank statement: less money than there should be.

2. The absence of a record is what turns a small dishonesty into a habit. A free round on Thursday that nobody ever mentioned becomes a free round every Thursday. That is on the operator. A restaurant where nothing is counted has not trusted its people; it has left them alone with a temptation and no structure.

3. Separation of duties is impossible at this scale, and pretending otherwise is the failure. Four managers cannot hold four functions cleanly. What you can build is the seven practical substitutes — an outside reconciler, cross-review at the top, rotation, blind records, owner review of exception reports, mandatory absence, and systems that log. Knowing the difference is the actual skill.

4. The single most valuable control an independent operator can buy is one they already pay for. The bookkeeper never touches cash, never signs an invoice, never approves a comp, never writes a schedule — which makes them the only genuinely independent party in your financial life. Ask for a list of exceptions, not a statement. It costs a conversation.

5. Cash is the smaller risk and the bigger ritual. \$1,395 a year at Bellwether — 2.6% of the \$53,122 — routinely absorbs more than half the control attention in the building. Count the drawer every night anyway: the ritual protects staff, protects safety, and a business that cannot count \$612 cannot be trusted to count \$25,488 of walk-in. Then go read the reports where the money is.

6. A void is invisible on the profit and loss statement; a comp is not. A void removes an item that was never sold. A comp removes revenue from a sale that did occur while the product cost stays in cost of goods sold — depressing the denominator and holding the numerator, moving both halves of the food cost ratio the wrong way at once.

7. The rate at which reason codes go un-entered is the best single measure of whether your comp policy is real. Report it as a percentage, not as dollars. One un-coded comp on a Friday is \$16; it is also 8.6% of that night's comp dollars, permanently un-reviewable, and that share only grows.

8. The daily sales report is the one control to install if you can install only one. One page, three ties, closed while the people who handled the money are still in the building. The flash report is not a separate document — it is seven daily sales reports added up. If yours comes from the point-of-sale weekly summary instead, you have a report and not a control, because nothing in it was ever compared to a physical count of anything.

9. The money crossing your bank account is never equal to your sales, and never will be. Larger by the charged tips and the gift certificates — neither of which is yours — and smaller by the processing fees. An operator who cannot separate revenue from receipts eventually spends the sales tax.

10. A threshold that generates more work than you will actually do is worse than no threshold, because it teaches you to ignore your own reports. Use the greater-of rule at the category level so small categories stay quiet, and the smaller-of rule at the whole-book level so the tripwire stays sensitive. Sensitivity where a look is cheap; specificity where a look is expensive.

11. One period is a data point. Two of three in the same direction is a pattern. Three periods bouncing across zero is your counting error plus the natural volatility of the product. A single period over threshold gets a look. A pattern gets a project. A negative variance gets a recount — a gift is a mistake you have not found yet.

12. A variance report tells you WHERE to look and never WHAT you will find. It is a flashlight, not a verdict. Period 8 said meat and poultry; the ladder found stale cost cards, an uncosted special, an unchased credit, a scale nobody used, and birds that came in heavy.

13. Theft is seventh, and the order is both fairer and more accurate. Base rates favor the boring explanations — \$997 of Bellwether's \$1,062 resolved onto rungs one through six. And an investigation that begins at rung seven gathers confirmation rather than evidence, typically leaving the real cause running while a good employee is quietly pushed out. The stale cost card is still there after the firing, and so is the variance.

14. If you punish the waste log, you convert waste into theft — not in reality, but in your numbers. Spoiled product and stolen product land on the same line, indistinguishably. The only thing separating them is a piece of paper somebody was willing to fill out, and whether they are willing depends entirely on what happened the last time somebody did.

15. The leaks are almost never surveillance problems. They are a scale, a jigger, a spec written as a range, a delivery window with a name on it, and a cost card someone updated in April. \$233 of jiggers closes \$7,631 of over-pour — a 32.7-times return available in every bar in America for the price of a conversation about how drinks get made.

16. Controls protect the innocent at least as much as they catch the guilty. A manager who counts with a second person cannot be accused of a shortage. A bartender counted on the same rotation as everyone else has evidence their variance is normal. In a restaurant with no controls, an unexplained loss lands on whoever the owner happens to suspect — which is a far worse place to work than one with a clipboard.

17. Respond to a suspected loss soberly. Document before you speak. Establish the pattern, not the incident. Write down every innocent explanation and rule each one out on evidence. Involve counsel before the first interview, not after the termination. Keep the circle small. Never confront on the floor, never interview alone, never "handle it quietly" by cutting hours, and never decide about law enforcement in the first hour.

The rules and the arithmetic

THE FOUR CONTROL FUNCTIONS
   AUTHORIZE  ·  RECORD  ·  CUSTODY  ·  RECONCILE
   Separation of duties: no one person holds two of these for the same asset.
   At four managers it is impossible. Substitute, and know that you are.

INVENTORY VARIANCE
   actual usage       = beginning inventory + purchases - ending inventory
   theoretical usage  = cost cards x POS item mix
   variance           = actual - theoretical, by category, per period

VARIANCE THRESHOLDS
   CATEGORY, one period     > GREATER of (5% of that category's ideal usage, $250)
     liquor and wine        > GREATER of (3% of ideal usage, $150)
   CATEGORY, persistence      over threshold in 2 of 3 periods, same direction
   WHOLE BOOK, investigate  > SMALLER of (1.0 point of food cost, $400)
   WHOLE BOOK, escalate       +/- 2.0 points SUSTAINED ACROSS TWO PERIODS
   NEGATIVE variance        > 1% under theoretical  -->  RECOUNT, do not investigate
   CASH, per night          > GREATER of ($5.00, 0.5% of cash sales)
   CASH, persistence          3 nights out of tolerance in a period,
                              or any single night over $25

   Greater-of downstairs, smaller-of upstairs.
   The total decides WHETHER to look. The category decides WHERE.

WHAT A PATTERN IS
   one period over threshold ............ a DATA POINT   -> a look
   2 of 3 periods, same direction ....... a PATTERN      -> a project
   three periods crossing zero .......... VOLATILITY     -> fix the count first
   hovering at the line, period after
     period, never dramatic ............. CHRONIC WASTE  -> a waste log, not a meeting

THE DAILY SALES REPORT TIES THREE TIMES
   1   sold             =  tendered                       tonight
   2   cash expected    =  cash counted                   tonight
   3   card tenders     =  processor BATCH  =  the bank   in 1-2 days
   Tie 3 is a PREDICTION until the batch report matches it. Match it to the
   batch, never to sales.

THE INVESTIGATION LADDER
   1 stale cost cards -> 2 uncosted specials -> 3 menu-mix drift ->
   4 purchasing -> 5 over-portioning -> 6 unrecorded waste -> 7 PEOPLE
   Rungs 1-4 are paper you already own, settled in an afternoon, harm if wrong: none.
   Rung 7 costs you a relationship even when you are right.

PRICING A CONTROL PROGRAM
   break-even recovery rate = program cost / measured exposure
   $4,849 / $53,122 = 9.1%   <- the number to quote to anyone who says
                                "you'll never recover all of it." You won't.
                                You don't have to.

Bellwether at a glance

Key terms

Internal control · Separation of duties · Cash handling procedures · Bank · Cash drop · Over/short · Comp / void / discount authorization · Reopened check · Exception report · POS audit trail · Daily sales report (DSR) · Inventory variance · Variance threshold · Blind count · Surprise count · Sweethearting · Over-ringing · Shorting · Vendor fraud

What you should be able to do Monday morning

Print last week's exception summary — voids, comps by reason code, reopened checks, no-sale opens — and read all four in forty-five minutes; build one daily sales report tonight before the building empties and tie it in three directions; write your variance thresholds on a card and post them where the count sheets live; email your bookkeeper and ask for a list of what does not tie, weekly, forever; and put the investigation order on the wall in seven words so that the next time a number moves, everybody in the building already knows the first six places you will look — and that none of them is a person.