How to Forecast Daily Covers So You Know Exactly How Much to Prep

The short answer: forecast tomorrow's covers by averaging the same weekday from the last four to six weeks, adjust that number for trend, weather, events, and what's on the books, then convert covers into prep quantities using each dish's attachment rate and yield percentage.

That's the whole system. Everything below is how to actually run it without spending an hour a day on spreadsheets.

Step 1: Build a same-weekday baseline

Never average across a whole week. A Tuesday and a Saturday are different businesses. Pull your last four to six same weekdays from your POS and average them.

Example for a Thursday dinner service:

| Date | Covers |

|---|---|

| Thu (4 weeks ago) | 118 |

| Thu (3 weeks ago) | 131 |

| Thu (2 weeks ago) | 126 |

| Thu (last week) | 140 |

Average = 129 covers.

Throw out obvious outliers — the Thursday you were closed for a burst pipe, the Thursday a food influencer posted about you. Note why you threw them out. Those notes become your event calendar later.

Step 2: Apply a trend factor

If your four data points are climbing (118 → 131 → 126 → 140), a flat average under-forecasts. Compare the most recent two weeks against the two before:

• Recent two: (126 + 140) / 2 = 133
• Prior two: (118 + 131) / 2 = 124.5
• Trend factor = 133 / 124.5 = 1.07

Apply it conservatively — half the trend is a safe rule when your sample is small: 129 × 1.035 ≈ 134 covers.

Step 3: Adjust for the things a spreadsheet can't see

Now layer on judgment. Keep these adjustments as percentages so they're repeatable:

• Weather. Patio-heavy rooms get hit hard by rain; delivery-heavy operations get a lift. Log actual vs. forecast on bad-weather days and you'll build your own real multiplier within a season.
• Local events. Home games, concerts, conference dates, school holidays, university move-in and finals weeks. Build a shared calendar and put dates on it a month out.
• Your own marketing. A promo email that goes out Wednesday morning affects Thursday.
• Competitor closures or openings. A new restaurant opening two doors down usually dents you for a few weeks.
• Holiday adjacency. The Tuesday before a long weekend is not a normal Tuesday.

Write the adjusted number down: 134 covers × 0.90 (forecast rain) = 121 covers.

Step 4: Reality-check against reservations pace

If you take reservations, you have a leading indicator. Track, for each weekday, what percentage of final covers are booked at 48 hours out.

Say your Thursdays typically sit at 55% booked two days ahead. Today you have 78 covers on the books for Thursday:

78 ÷ 0.55 = 142 projected covers

Now you have two numbers: 121 from history, 142 from pace. Don't average them blindly — ask which signal is stronger. A rain forecast three days out is soft; 78 confirmed reservations is hard. Lean toward the book, temper it slightly: call it 135.

Build this pace curve once for each weekday (booked % at 7 days, 3 days, 48 hours, morning-of) and it becomes the single most useful forecasting tool a reservation restaurant has.

Step 5: Convert covers into menu items

Covers are useless to a prep cook. What they need is how many pounds of short rib.

For each menu item, calculate its attachment rate: units sold ÷ total covers, over the last 30 days on that weekday.

| Item | Units (last 4 Thu) | Covers (last 4 Thu) | Attachment rate |

|---|---|---|---|

| Short rib | 128 | 515 | 0.25 |

| Roast chicken | 93 | 515 | 0.18 |

| Cacio e pepe | 149 | 515 | 0.29 |

| Beet salad | 82 | 515 | 0.16 |

At 135 forecast covers:

• Short rib: 135 × 0.25 = 34 portions
• Roast chicken: 135 × 0.18 = 24 portions
• Cacio e pepe: 135 × 0.29 = 39 portions
• Beet salad: 135 × 0.16 = 22 portions

Attachment rates drift when you change the menu, reprice, or move an item's position on the page. Recalculate monthly, and immediately after any menu change.

Step 6: Apply yield to get raw purchase and prep quantities

Portions are still not prep quantities. You need yield.

Short rib: 34 portions × 6 oz cooked portion = 204 oz cooked. If braised short rib yields 55% of raw weight:

204 ÷ 0.55 = 371 oz raw ≈ 23 lb raw short rib

Beet salad: 22 portions × 3 oz roasted beet = 66 oz. Beets yield roughly 75% after peeling and trimming:

66 ÷ 0.75 = 88 oz ≈ 5.5 lb raw beets

Do this once per recipe and it never changes. The only variable that moves day to day is the cover count at the top.

Step 7: Prep to par, not to forecast

Here's the step most kitchens skip. Your prep list isn't the forecast — it's the forecast minus what's already in the walk-in.

Prep quantity = (forecast requirement + safety buffer) − on-hand usable

Safety buffer depends on the ingredient:

• Long-shelf-life, cheap (braises, stocks, pickles, dressings): 20–30% buffer. Running out costs more than the food.
• Short-shelf-life, expensive (fish portions, cut herbs, dressed greens, plated garnishes): 0–10% buffer. Prep to order or prep twice.

Split your prep list into these two categories explicitly. It's the fastest waste reduction most kitchens can make.

Step 8: Track forecast error weekly

Every day, log forecast covers vs. actual covers. Once a week, calculate average absolute error:

|forecast − actual| ÷ actual, averaged across the week.

A good, mature forecast for a stable restaurant lands in the single digits to low teens percent. If yours is worse, look at which day is breaking. Usually it's one weekday with an event pattern you haven't captured, or a menu item whose attachment rate has shifted.

Also log 86s and end-of-night waste by item. An item that 86s twice a month needs a bigger buffer; an item you dump three times a week needs a smaller one.

A realistic daily rhythm

• Morning (10 min): Pull yesterday's actuals. Log forecast error. Check reservations pace for the next 3 days. Set tomorrow's cover forecast.
• Pre-service: Print the prep list, generated as forecast covers → attachment rates → yields → minus on-hand.
• Post-service (5 min): Note 86s, waste, weather, and anything unusual. That note is next month's accuracy.

Common mistakes

• Averaging across all seven days instead of same-weekday.
• Using a six-month average — it buries seasonality and trend.
• Forecasting revenue instead of covers. Revenue moves with pricing and check average; prep moves with people.
• Ignoring lunch/dinner split. Forecast each daypart separately.
• Keeping the forecast in one person's head. Write it on the board so the whole line can see it.

---

Once your cover forecast feeds real portion and yield numbers, the next question is what each of those portions actually costs you. You can run a single recipe through the free food cost calculator at getprepsheet.com/calculator, or grab a ready-made Excel costing template from the shop if you'd rather build the whole menu out at once.