Restaurants & Food · Scenario 03

Bakery & Pastry

Bake too much and it’s tomorrow’s bin; bake too little and the 4pm shelf — and the till — sit empty. We forecast demand per item per day so the oven matches reality.

Method · Forecasting + newsvendor optimisation

The situation

The bakery runs on a baker’s intuition built over years — impressive, but blind to patterns it can’t hold in its head: weather, paydays, the school calendar, which croissants vanish by 10am while the rye sits until close. Some days end with trays of waste; others turn customers away by mid-afternoon.

Every wasted item is ingredients, labour and energy thrown out; every stockout is a sale handed to the café next door. Neither shows up clearly in the books, so the leak goes unmeasured.

~14%
daily output wasted
9 items
sell out before noon
By feel
how production is set
31
SKUs baked daily

Where we dig for the truth

We turn two years of sales into a per-item, per-day demand forecast, then bake to the quantity that maximises profit given the cost of waste and the cost of a stockout.

Item-level sales (2 yrs)Waste & markdown logsWeather historyLocal calendar & eventsIngredient & labour costShelf-life by product
Croissants: baked vs soldFlat daily production against a swinging demand04385128170MTWTFSSMTWTFSSBaked (flat)Actually sold

Flat batches mean waste on slow days and stockouts on busy ones. The demand pattern is predictable — it tracks weekday and weather.

Our approach — Demand Forecasting & the Newsvendor Model

We fit a demand model per product from weekday, weather, season and local events, then apply a newsvendor calculation: the profit-maximising bake quantity given each item’s margin, ingredient cost and shelf life. High-margin, long-shelf items get a safety buffer; perishable low-margin items get trimmed.

The baker keeps the craft; the model just sets the morning’s numbers. A one-page sheet says how many of each to bake today, updated for tomorrow’s forecast.

From a daily guess to a daily forecast1Assemble historyCombine two years ofitem sales with weatherand the local calendar.2Forecast demandModel expected salesper item per day, withits uncertainty band.3Solve the bakeNewsvendor quantity peritem from margin, costand shelf life.4Print the sheetA daily bake list theteam follows; it learnsfrom each day.
Waste rate by item — before vs afterShare of daily output discarded0%6%13%19%26%16%6%Croissant9%4%Sourdough18%7%Muffin22%9%Fruit tart11%5%BagelBeforeAfter

Forecast-led baking roughly halves waste while cutting the stockouts that were quietly sending customers elsewhere.

What changes

Same recipes, same oven — a number on a sheet instead of a guess. Representative for a single-location bakery.

Representative 90-day movementDaily waste14%6%▼ -8 ptsStockouts / wk216▼ -71%Gross margin62%69%▲ +7 ptsMonthly profit$12k$19k▲ +58%
Where the profit improvement comes from+$7kmonthly profitLess waste46%Fewer stockouts34%Better product mix20%
Why this is not "social media management"
No rebrand, no ad campaign — just the right number of croissants. We replaced a daily guess with a forecast, and the profit that used to go in the bin stayed in the business. That's the unglamorous, data-driven work most agencies never touch.

Frequently asked questions

How can a bakery cut waste and stockouts at the same time?
We forecast demand for each item per day from two years of sales, weather and the local calendar, then apply the newsvendor model to set the profit-maximising bake quantity. This roughly halves waste while cutting the stockouts that send customers elsewhere.
What is the newsvendor model?
It is a classic operations-research formula for perishable goods that finds the production quantity balancing the cost of waste against the cost of a stockout, given each item's margin and shelf life.
Do we have to change our recipes or branding?
No. The baker keeps the craft; we replace the daily guess with a forecast and a one-page bake sheet. It is data work behind the scenes, not a rebrand. Book a marketing audit.

Want this run on your numbers?

Send two years of daily sales and we’ll forecast your waste and stockouts away.

Book a marketing audit
Back to Restaurants & Food