Pricing a custom cake: the costs most bakers forget

Most quoting mistakes are not arithmetic. They are omissions: the board, the box, the dowels, the drive to the venue, the hour spent on the phone agreeing the design.

This post walks through a full cost breakdown for a single custom order, in the order the costs arrive. You do not need software to do any of it. You do need every line written down before a price is.

All figures here are ratios and proportions. Substitute your own currency and your own numbers.

1. Ingredients, at what you actually paid

The ingredient cost of a cake is not the price on the shelf. It is the price you paid for the pack, divided by how much of that pack ended up in the cake.

That division is where people lose money, for two reasons.

The unit you buy in is rarely the unit you bake in. You buy flour in a 5 kg (11 lb) sack and measure it in grams. You buy fondant in a 750 g (1.65 lb) block and use 220 g (about 8 oz) of it. Somewhere, something has to convert. Hold one cost per one unit for every item you buy: pack price, divided by pack quantity in the unit it was sold in, converted into the unit you measure in. Do that conversion once, record it, reuse it. Where two units cannot honestly be converted at all, fix the record rather than guessing a factor, because a guessed conversion is a wrong cost on every product using that ingredient.

Per-gram costs run to several decimal places, and rounding kills them. This is the single most common way a spreadsheet quietly zeroes out an ingredient. If your cost per gram is 0.000048 and your spreadsheet is formatted to two decimal places, that ingredient now costs nothing, and every cake you price with it is wrong. The fix is to vary the precision with the size of the number instead of formatting everything the same way: two decimals at or above 1, four at or above 0.01 and six below that, so a cheap per-gram cost keeps the digits that matter.

The costs people leave out of this section are the ones that are cheap per cake and expensive per year: gel colour, gold leaf, isomalt, the flavouring you use six drops of. Six drops is nearly free. The bottle is not, and you buy the bottle several times a year.

One thing to be honest about: almost nothing costs waste for you. If you trim 15% off every sponge, no sheet and no system knows that unless you enter the quantity you take out of the tub rather than the quantity that ends up in the cake. Cost the sack, not the slice.

Two checks will find most of this in an afternoon, on whatever you cost with now. Take the cheapest thing in your sheet by unit cost and see what a batch of it contributes: if that is zero, or a row of noughts, the conversion or the rounding has eaten it and every product containing it is mispriced. Then take the most expensive thing you buy and multiply your recorded per-unit cost back up to a full pack. If that total is not close to the invoice you paid, the conversion is wrong somewhere between the sack and the bowl.

2. Packaging and the board

This is the cost that is most often forgotten and most reliably underestimated, and we can say that with some confidence because we got it wrong ourselves. Before ibakepro tracked packaging properly, packaging contributed zero to every product, and every margin was overstated by exactly the packaging spend. It looked fine. It was wrong on every product.

Three rules came out of fixing it.

Packaging is its own line, never folded into ingredients. Materials cost is raw materials plus packaging, and the two are worth keeping as separate buckets all the way through. If a box price is hiding inside the ingredient total and is also declared as packaging, it gets counted twice and your margin reads worse than it is. Whichever way you track it, pick one place and keep it there.

The number of packages is not always one. This is the arithmetic that catches everybody. Packages needed is the base quantity divided by the package capacity, rounded up. A single-cavity box packing six cupcakes is six boxes, not one. A twelve-cavity box holding six cupcakes is one. A six-cavity box holding twelve cupcakes is two. Do that division deliberately every time, because the instinct is to write "1 box".

Some packaging only exists if the order leaves the building. A courier-safe outer carton, the ice pack, the crush protection: those cost on delivery and on dispatch, never on a collection. So tag each packaging item with the fulfilment methods it applies to, and count against an order only what that order needs for the way it is going out. Anything untagged should cost on all of them: over-costing a collection is a small error, and missing the carton on a courier order is not.

Boards, dowels, ribbon, cake boxes, labels and toppers belong in the same list as your flour, carrying a cost per unit in the same way. A 25 cm (10 inch) drum under a 20 cm (8 inch) cake is a real cost with a real price on the invoice: bought fifty at a time, charged one at a time.

If you have never totalled a full kit, total one rather than estimating it. Lay out everything one delivered cake leaves the building wearing: the drum, the box, the non-slip, the ribbon, the dowels, the label, the topper, the outer carton, the cold pack. Add the line prices you paid, then hold that total next to the ingredient cost of the cake that sits inside it. Most people who do this once stop arguing about whether packaging deserves its own line. The worked example later in this post carries a packaging line too, but that figure is illustrative and yours will not match it.

3. Labour, at a rate you would accept from someone else

Labour is minutes multiplied by an hourly rate, and both halves get shaved.

The rate is the easier one to fix. Set it to what you would have to pay a competent decorator to do this work instead of you. If you are not willing to put a number on your own hour, the number you are using is zero. That is not a figure of speech: a rate left blank is a rate of zero, and a price built on one prices your time at nothing. Nothing about the resulting number looks wrong. It is a confident, complete-looking price missing the largest cost in most custom work, and nothing warns you, because a blank field is a valid number.

The minutes are the harder half, because the decorating time is the part you remember and the smallest part of the total. The minutes that go missing are:

• the consultation, and the messages afterwards
• sourcing the topper, the ribbon, the specific colour
• levelling, torting, filling and crumb-coating, which is most of the day on a tiered cake and none of the photos
• the second trip to the shop
• washing up, and the tools that cannot go in a dishwasher
• boxing, transporting and setting up

Keep the hourly rate in exactly one place, so raising it reprices labour everywhere at once. And record labour per size rather than per portion, which matches reality: a bigger tier is not proportionally more work, and a heavily detailed small cake can be far more work than a plain large one.

There is only one way to fix the minutes: measure them once. Take the next order you accept, write down your estimate before you start, then run a clock on the whole thing: the first message, the design conversation, the sourcing, the shopping trip, the bake, the fill, the decorating, the boxing, the washing up. Stop the clock when the kitchen is clean, not when the cake looks finished.

Almost nobody guesses high. What that exercise gives you is not one order's number but a correction factor, and the factor is the transferable part: apply it to the next twenty estimates and you are pricing from something measured instead of something remembered. Repeat it once a year, or after any change in how you work, because the factor moves as your process does.

4. Overhead, allocated rather than ignored

Overhead is everything you pay for whether or not this particular cake exists. It divides into two kinds that behave differently and need allocating in two different ways.

Costs that scale with the size of the job. Power for a six-hour bake, gas, water, the wear on your mixer. Charge these as a single percentage applied to materials plus labour: overhead equals (materials + labour) x your overhead rate. A bigger, longer job carries more of it, which is the point.

Fixed monthly costs that exist regardless. Rent, insurance, internet, the standing portion of your utilities. These do not scale with one cake, so a percentage is the wrong tool. Add them into a monthly total and divide by your monthly order count to get a flat amount that every order carries.

That division is where to be careful. It divides by an estimate. Estimate optimistically and every order under-carries its share, and the gap does not announce itself, because your rent does not fall in a quiet month. Set that number from what you did last quarter, not from what you hope to do.

Two consequences worth knowing. A per-order cost is charged once per order, not once per item, so a five-item order carries it once and not five times. And, like the labour rate, an overhead rate you never set is a rate of zero, and nothing warns you that it is missing.

Do that division once for yourself, on paper, before deciding whether it is worth doing at all. Add up one month of costs that arrive whether or not you bake: rent or the share of your home the kitchen occupies, insurance, licences, subscriptions, the standing portion of power and water, the accountant, the phone. Divide by the orders you completed that month, not the number you would like to. Then set that per-order figure beside the ingredient cost of one typical cake.

Whatever the two numbers turn out to be for you, the comparison is the argument for allocating overhead at all. It is also why a quiet month is expensive twice over: the monthly total does not shrink when the order count does, so every order in a thin month should be carrying more, and almost nobody adjusts.

5. Delivery, which is not profit

The delivery fee feels like income. It is not. It is a reimbursement, and it is only a reimbursement if it covers the drive.

Keep it out of the margin calculation entirely. Measure margin on the discounted item subtotal, excluding tax and excluding delivery: tax is not your money and delivery is a pass-through. A fee then cannot flatter your margin, and if it does not cover the trip the shortfall comes out of the cake, where you can see it.

Three things to check in whatever you use to charge for delivery:

Distance is driving distance, and it is one way. Price from real driving distance between your address and the customer's, not a straight line on a map, and remember the number you resolve is the distance out and not the return. A 20 km (12 mile) delivery is a 40 km (25 mile) round trip for you. If your per-distance rate only covers the outbound leg, it covers half the drive.

A free-delivery threshold does not belong on a distance charge. The drive costs the same regardless of what is in the car, and a big order at the end of a long trip is exactly where waiving the fee costs you most.

The time is labour, and no fee structure captures it automatically. An hour and a half of driving and setup is an hour and a half. Distance pricing prices the distance, not your time. If you want the time covered you have to build it into the rate or into the order. Nothing does it for you.

Choosing between a flat fee, distance bands and a hybrid is a decision of its own, and it turns on figures you have to measure rather than adopt: your all-in running cost per distance, your real door-to-door time and how many drops fit into one run. What should you actually charge for delivery? works that method through end to end.

For the quote in front of you, only one thing has to be true: the fee covers the round trip and the time, or you have decided deliberately to carry the difference. A shortfall you chose is a marketing cost. A shortfall you never noticed is a smaller cake margin than the one on your screen.

6. The margin that sits on top

Everything above is cost. Margin is the part that makes the business a business rather than a job that pays for its own ingredients.

Margin is not markup, and confusing the two is expensive. A margin is a share of the selling price. A markup is a multiple of cost. They are not the same number and they do not produce the same price.

To hit a target margin, the formula is:

price = cost / (1 - margin)

A 40% margin therefore means multiplying cost by about 1.67, not by 1.4. If you multiply cost by 1.4 and call it a 40% margin, your actual margin is 28.6%, and you have given away roughly 40% of the profit you meant to make. Check that formula against whatever you price with now.

Discounts come out of profit, not out of price. There is a clean rule here: the share of your profit a discount consumes is the discount percentage divided by the margin percentage. At a 40% margin, a 10% discount gives away a quarter of your profit. At a 20% margin, that same 10% discount gives away half of it. Nothing about the conversation with the customer makes that feel like what is happening.

Compare margin excluding tax. Where prices are displayed with tax included, the price on your website contains money that is not yours. Calculating margin against a tax-inclusive price flatters you by the tax rate. Strip the contained tax out before comparing revenue against cost, line by line, so lines in different tax categories back out at their own rate rather than one blended one.

A cost change should not silently reprice your work. When a supplier price moves, the recorded cost should move with it and the price you publish should stay where you put it. What you want is a flag: this product's realised margin has drifted off target, and here is the price that would restore it. You decide whether to move it. The one case that needs no thought is a product that has fallen below its own cost, which loses money on every single sale. Which of the rest need a price move after an input cost changes, and which can absorb it, is worked through in Your flour just went up 20 percent. Now what?.

A worked example, in units rather than currency

Here is the shape of the arithmetic. These proportions are illustrative, not a benchmark. They are here to show the order of operations, and you should replace every one of them with your own.

Call the ingredient cost of one cake 100 units.

Line	Amount
Ingredients	100
Packaging, board, box	25
Labour (4 hours at 30 per hour)	120
Overhead at 15% of materials + labour	36.75
Fixed monthly costs, per-order share	20
Total cost	301.75

At a 40% target margin, the price is 301.75 / 0.6, which is about 503.

Price it by multiplying cost by 1.4 instead and you get about 422. That is a 28.6% margin, and about 80 units of profit gone from the 201 you meant to make, on one cake.

Read the shape of that table, not its numbers. Labour sits above ingredients in this example because most custom decorating work is time-heavy, but "labour is 120 against 100 of ingredients" is not a rule, a benchmark or an average of anything, and nothing in this post is claiming it is typical. A plain sheet cake and a hand-piped tiered wedding cake sit at opposite ends of that ratio. So do two decorators of different speed working from the same sketch. Every one of the six lines above will land somewhere else on your own orders, and several of them will land somewhere else on two different orders in the same week.

The arithmetic is the part that transfers: cost every line, add them all, then divide by one minus your margin. The proportions are yours to measure.

What to check on your own quote today

• Is any ingredient in your sheet rounded to two decimal places when it costs fractions of a cent per gram?
• Is the board on the quote? Is the box? Is the outer carton, on the ones that get driven?
• Is the number of boxes correct, or is it 1?
• What hourly rate are you actually charging yourself, and would you accept it from an employer?
• Does anything in your price cover the rent?
• Does the delivery fee cover the drive back?
• Is your margin a margin, or a markup you have been calling a margin?

You do not have to fix all seven at once. Pick the one you have never costed, cost it properly on the next quote you send and compare that price against the one you would have sent yesterday. The gap is not a hypothetical. It is what the omission has already cost you on every order you have taken so far, and it is the most persuasive argument for the other six.

Every line above is arithmetic you can do on paper, and doing it once on paper is the only way to believe the answer. What software is good for is the repetition: ibakepro holds one cost per unit for everything you buy, keeps packaging on its own line and works the price back from the margin you set rather than a markup. What your hour is worth, and how long the job really took, are the two inputs only you can supply, and they move the price most.

Costing is not the interesting part of this work. It is the part that decides whether you get to keep doing the interesting part.