Can you post what you bake?

Sooner or later someone far away asks whether you post. It is a flattering question, and it is usually answered too fast in either direction. "No" gives away a market that costs you no driving. "Yes" turns one order into a fortnight of arguments about a smashed box.

The honest answer depends far more on what you make than on which courier you pick. So start there.

What survives a parcel network, and what does not

A parcel network is not gentle transport. Between your hands and the customer's, a box goes down chutes, along conveyors, into cages and under other people's parcels. It gets tipped on every face. It sits in a van that reaches oven temperatures in summer and freezing in winter. None of that is unusual or a sign of a bad carrier. It is the design.

Things that travel well share four traits. They are dense, so there is no structure to collapse. They are low in moisture, so they neither dry out nor go sticky in a few days. They are stable at room temperature, because you control the temperature at no point in the journey. And the decisive one: no part of them carries the weight of another part.

That last one rules out most celebration cakes. A tiered cake is a building. It stands because every layer is loaded from directly above, in one orientation, in a vehicle that stays level. Turn it on its side for ten minutes and it is not a cake with a dent, it is a bag of crumbs and buttercream.

Reliable in a parcel:

• Biscuits, cookies, shortbread, brittle and bark
• Brownies, blondies, traybakes cut and wrapped individually
• Fruitcake, gingerbread, dense loaf cakes without a soft filling
• Fudge, caramels, honeycomb, most confectionery
• Sealed jars, dry mixes, granola, decorated cookie kits

Unreliable, and a source of refund arguments:

• Anything with fresh cream, custard, cream cheese or an unstabilised filling
• Tiered or stacked cakes, and tall thin cakes of any kind
• Macarons and meringue, which fail from vibration rather than impact
• Untempered or thinly tempered chocolate in anything above mild weather
• Anything that depends on refrigeration to be safe rather than merely nice

A middle group is worth testing rather than assuming: single-layer buttercream cakes, cupcakes in a fitted insert, anything covered in fondant. Bakers post these successfully every week, with packaging designed for that specific product rather than a cake box inside a carton.

The box is not a car seat

When you drive an order yourself, the packaging only has to stop the cake sliding on one journey you are watching. In a parcel network it does the whole job unsupervised.

The enemy is not a single dramatic impact. It is movement. A product that cannot move cannot build up speed inside its own box, and almost everything that arrives broken arrived broken because it was allowed to travel a couple of centimetres, about an inch, several thousand times.

So the working structure is two boxes: an inner container holding the product immobile, and an outer carton with a cushioning layer of at least 5 cm, or 2 inches, on every face including top and bottom. Void fill that compresses to nothing under load is worse than none, because it gives you confidence you have not earned. The inner box should not shift when you shake the outer one, and the product should not shift when you shake the inner one.

Three tests before you sell a single posted order. Shake the sealed parcel next to your ear: you should hear nothing. Turn it upside down for a minute, then open it. Drop it from about waist height, roughly 75 cm or 30 inches, onto a hard floor, then open it. If the product survives all three, you have packaging. If you find yourself writing "this way up" on the box, you do not, because nobody in the network is reading it.

Two more things earn their place: a moisture barrier, so the product does not share humidity with its own packaging for three days, and a note inside telling the customer how to open it, because a surprising share of damage happens in the recipient's kitchen with a knife.

A courier is not a driver you control

This is the part most bakers underestimate, and it is not about reliability.

With your own driver you have discretion. You can leave later, call ahead, wait while someone finds the key, carry a repair kit, fix a smudge on the doorstep. When something goes wrong you are standing there, and almost every delivery problem you have ever had was solved by a person being present.

With a courier you have none of that. You cannot intercept a parcel in a depot, you cannot explain the fridge situation to the driver and you find out about a problem after the customer does. A carrier's delivery estimate is a probability, not a promise, and repeating it as a promise in your own marketing is how bakers end up refunding orders that arrived on time.

What you get in exchange is real: reach beyond your driving radius, no return leg, no Saturday morning and a cost that does not climb with distance the way your own vehicle does. The arithmetic for your own driving is in what to charge for delivery, and if the answer turns out to be your own van after all, plan the day with an optimised route.

Decide the failure policy before the first failure

Late and broken parcels are not edge cases, they are a running rate. Decide these in advance, in writing, on the product page:

• Do you replace, refund or neither, and on what evidence? Most bakers land on "photograph it and we will replace or refund", which is simple and generous and has a real cost that has to be priced in.
• What is your dispatch cutoff? Posting late in the week risks a weekend in a depot: two extra days of ageing and a temperature swing you cannot see. Many bakers post early in the week only, and say so.
• What do you photograph before sealing? A picture of the packed inner box takes four seconds and settles arguments about whether it left you intact.

Then carry the cost. If a fraction of parcels fail and each failure is replaced and reposted, the number of shipments per successfully delivered order is:

At 1 failure in 20 that is 1.053, so every posted order should carry about 5.3 percent more than one shipment's cost. But unlike a failed doorstep delivery, a failed parcel usually writes off the goods too, so add the production and packaging cost at the same rate. That second part is normally the larger one.

Is shipping worth offering?

Reduce it to one comparison you can make. Take a product you already sell at the counter and cost the posted version honestly: production, plus shipping-only packaging, plus postage at the real packed weight and distance, plus the replacement provision above, plus the time it takes to pack and label one parcel. Then compare the margin per hour of your time against the same product sold locally.

Both answers are legitimate. What is not is offering it without doing the sum, discovering the postage costs more than the product and either eating that silently or raising prices on your local customers to cover it.

If it does clear the bar, narrow it. Sell a postal range rather than your whole menu: three or four products chosen because they travel, in one or two box sizes you have weighed and measured once. That makes every quote accurate, every pack identical and every failure investigable.

When the volume arrives, the mechanics are what start to hurt: quoting a rate against the actual parcel rather than a guess, getting a label onto the box without retyping the address and knowing where something is when a customer asks. That is worth solving with software once you are posting regularly.

The parcel network will still do what parcel networks do. Choosing the product, the packaging and the policy first is what turns that from a surprise you argue about into a cost you already priced.