Bakery & Food Retail (Multi-Outlet Chain) · Business Performance

Costing a cake to the gram, then pricing it for the market

A bakery chain priced its cakes on instinct and a rough ingredient cost. Full costing showed which cakes paid for themselves and which the customer was being subsidised to buy.

Illustrative scenario. This describes the kind of problem we work on and how we approach it. It is representative of the work, not a record of a specific client engagement, and no figures here relate to an identifiable business.

← All case studies

The challenge

A multi-outlet bakery chain sold around forty cake varieties across its stores. Prices had been set by taking the cost of ingredients, adding a mark-up, and then adjusting to whatever the nearest competitor charged. Nobody could say what a cake actually cost to make and sell. Management knew the signature cakes sold well and assumed they carried the business; the plain everyday cakes were treated as volume drivers. Wastage was accepted as unavoidable and was never costed to a product. Outlet staff discounted end-of-day stock without any rule about what the floor price should be.

The diagnosis

Costed properly, the range looked nothing like management expected. Ingredient cost was the smaller part of the picture. Once labour by production stage, oven and mixer time, refrigeration, packaging, delivery to outlets, outlet occupancy and unsold wastage were allocated to each product, several results reversed. The premium signature cakes carried heavy decoration labour and high spoilage because they were made to a forecast rather than to order, so their real margin was far below the assumed one. The plain everyday cakes, dismissed as low value, produced strong contribution per oven hour because they used the cheapest production window and sold out. Wastage was concentrated in a small number of products and was the single largest hidden cost in the range. Two seasonal varieties lost money on every unit sold once the write-off on unsold stock was allocated. Because pricing anchored to a competitor whose cost base and outlet mix were different, several prices bore no relationship to what the product cost to produce.

The response

We built a product-level costing model for the range. Recipes were costed by weight to the gram, with yield loss at each stage measured rather than estimated. Labour was timed by production step: mixing, baking, cooling, decoration and packing. Oven and refrigeration capacity was treated as the constraint it actually is, so contribution was measured per oven hour rather than per unit alone. Delivery, outlet occupancy and card charges were allocated to products on a usage basis. Actual wastage was tracked by product and outlet over a full cycle and charged to the products that caused it. On that base we produced contribution per unit and per oven hour by product, a break-even volume for each variety, a price sensitivity table showing how much volume each cake could lose before a price rise stopped paying, and a floor price below which end-of-day discounting destroys value. We then modelled a repriced range, a shortened product list, and a production plan weighted toward the varieties that earn most per oven hour.

The outcome

Management could see, for the first time, contribution by cake rather than a single blended margin. Pricing decisions moved from matching a competitor to reflecting cost, capacity and demand, and the discounting rule at outlet level was set with a number behind it rather than left to staff judgement. Two loss-making varieties were retired and production shifted toward the products that used constrained oven time best. Conclusion: in food retail the ingredient cost is rarely the reason a product loses money. Labour, capacity, spoilage and outlet cost decide the outcome, and none of them appear on a recipe card. Costing a product to the gram is not an accounting exercise; it is what allows a business to price with confidence instead of copying the shop across the road.

Recognise the problem?

Every business is different, but the pattern often is not. Tell us what you're seeing.

Start a Conversation
CallWhatsAppFree Consultation