The plain-language guide

How much should you order for a pop-up event day?

A market day sells faster and runs longer than a normal Saturday, so last Saturday is the wrong starting number. Use your past events, work in units per open hour, and decide where in the range to order.

Published October 27, 2026

The short answer

Take the units per open hour from your two or three most similar past events, multiply by the hours the event will run, and read the result as a range. In the example below, 18 to 24 buns an hour over 12 hours gives 216 to 288 buns, with 252 in the middle. An ordinary Saturday scaled to 12 hours gives 144, which is 43% below the middle.

What an ordinary day gets wrong

The example is a burger stall at weekend markets. On an ordinary Saturday it opens 12:00 to 20:00, which is 8 hours, and sells 96 buns. That is 12 buns per open hour. Scale that to a 12-hour event and the answer is 144 buns. Two things have changed, and the ordinary day captures neither. The hours are 1.5 times longer (12 against 8), and the crowd buys faster. The three past markets in the table all sold at 18 to 24 buns an hour, which is 1.5 to 2.0 times the ordinary rate.

DayOpen hoursBuns soldBuns per hour
Ordinary Saturday8.09612.0
Market A9.016218.0
Market B10.021021.0
Market C11.026424.0

All four days are invented for this example. Together the longer hours and the faster rate make an event day 2.25 to 3.0 times an ordinary Saturday in buns (216 to 288 against 96).

Use more than one past event

Forecasting by analogy, the method this page uses, is covered in section 6.4 of Hyndman and Athanasopoulos, Forecasting: Principles and Practice. The authors recommend basing a forecast on several analogies, because a single analogy can carry a bias. They report that experts with more than two analogies, and direct experience of them, produced the most accurate forecasts.

The table shows the cost of using one. Market A alone gives 18 buns an hour, which is 216 buns over 12 hours. Market C alone gives 24 an hour, which is 288 buns. The two answers are 72 buns apart, a 33% swing on the lower figure, from a choice that has nothing to do with the coming event.

Turn the rates into a range

Multiply each rate by the hours the event will run. The mean of the three rates is 21.0 buns an hour, which is the middle case.

CaseBuns per hourHoursBuns for the day
Low (Market A)18.012216
Middle (mean of A, B, C)21.012252
High (Market C)24.012288

With 3 events, the lowest and highest rate mark the range. That makes a rough range from a small sample. A statistical interval needs more events than a stall usually has, so treat 216 to 288 as the plausible spread and keep a note of each new event to firm it up.

Test the longest hour

The 12-hour event is longer than all three past markets, which ran 9, 10 and 11 hours. The extra hour is a guess, because no past market shows what the twelfth hour sells. Test the guess against 3 cases for that hour, using the high rate of 24 buns an hour.

Twelfth hour sellsBuns in that hourBuns for the dayChange from 288
The full rate242880
Half the rate12276-12 (-4.2%)
Nothing0264-24 (-8.3%)

Even a dead last hour moves the high case by 8.3%. The hour count is a smaller risk than the choice of rate, which moves the day by 72 buns. Spend the effort on the rate.

Where in the range to order

The place to order depends on what each mistake costs. The overordering guide gives the ratio: the cost of running short divided by the cost of running short plus the cost of running over. Apply it per item. A straight line between the low and high cases is a rough way to turn the ratio into a quantity: order = low + ratio × (high - low). Here a burger that cannot be sold loses ₱45 of margin, an unsold bun loses ₱6.50, and an unsold patty loses ₱38.

ItemCost per unitMargin lost if shortRatioOrderPack sizePacks
Burger buns₱6.50₱45.000.872791224 (288 buns)
Beef patties₱38.00₱45.000.542551026 (260 patties)

Buns are cheap to over-buy, so the order sits near the top of the range at 279. Patties cost nearly as much as the margin, so the order sits near the middle at 255. Both round up to whole packs. Rounding up adds, on average, half a pack of extra stock, a figure the case-or-unit guide works through.

Event-day order

24

packs to order

Order deadline and stock already held

Subtract what you hold. With 30 buns on the shelf and none on the way, the 279-bun target needs 249 more, which is 21 packs of 12 (252 buns). Then check the clock. A supplier that needs 2 days of notice and delivers on Friday needs the order by Wednesday for a Saturday event. The order sheet works from the supplier's lead time, and an event does not move that deadline.

Record the event while it is fresh

Write down 3 numbers after the event: units sold, units left and the hour you ran out, if you did. Suppose the stall orders 288 buns and sells 258. That leaves 30, which is 10.4% of the order, and a sell-through of 89.6%. Now suppose it sold out at the 11th of 12 hours with 288 sold. The true rate is at least 26.2 buns an hour (288 over 11), above the 24.0 of Market C, because the last hour of sales was lost to the empty shelf. A sell-out hides demand, so record the hour it happened. Add the event to the table of past rates with its own hours and units, and the next range will rest on 4 events, up from 3.

What ForeFlux does

ForeFlux forecasts per item and per branch, and it works from the opening hours you set for each location, in 15-minute steps, with 10:00 to 20:00 assumed when none are set. A one-off event day is a single day among the days it learns from. For the dates of the event, the Adjust demand for a date range panel on a stock item lets you set your own figure, and Supply orders against it. We suggest, you decide: the event figure from this page is yours to enter.

Sources

  • Hyndman, R. J. and Athanasopoulos, G., Forecasting: Principles and Practice (3rd ed.), chapter 6 on judgmental forecasts, section 6.4 on forecasting by analogy. Use multiple analogies, and experts with more than two analogies and direct experience forecast best.
  • The stall, its 4 days of sales, its pack sizes and its pesos are an invented example for arithmetic. They come from no customer and no measured result.

Questions owners ask

Past the basics.

What if I have no past events to compare?

Use your ordinary rate times the event hours as the low case. In the example that is 12 buns an hour over 12 hours, or 144 buns. Then choose a high case yourself, such as 2 times the ordinary rate, which is 288 buns, and mark it as a guess. Order to your ratio inside that range, and record the day so the next event has one real comparison to use.

Should event days sit in my normal forecast?

Keep them labelled. Suppose 57 ordinary days sell 96 buns and 3 event days sell 288. A plain average of all 60 days is 105.6 buns, which is 10% above an ordinary day. An ordinary Saturday forecast built from that average would run 10% high, and an event forecast built from it would run far low. Any average over mixed days has this problem, so separate the two kinds.

What if the event is rained out?

Check how long each item keeps before the event. Frozen patties carry over to the next trading day, so 100 unsold patties cost only the cash tied up. Fresh buns spoil, so 100 unsold buns cost ₱650 at ₱6.50 each. For an item that keeps, set the overage cost near zero in the calculator and the order moves to the top of the range. For an item that spoils, enter its full purchase price.

ForeFlux

See the next order before you need it

Know what you need, before you need it.

The demo needs no signup and no card. It is the real product running on demo data.

Join the waitlist Open the order list

Demo data · SM Mall of Asia kiosk, BGC weekend pop-up, Cebu IT Park

We reply within 48 hours. Send a message.