Dropshipping gets described as a sales channel, which hides what it actually is: a financing arrangement. You take the money first and buy the goods afterwards, and the supplier holds the stock in between. That is the whole trick, and it is a good trick — it lets someone with a laptop and no capital sell heavy objects stored in another country.
It also rests on three assumptions that most write-ups never state, because for mass-produced goods they are true by construction. Carved stone breaks two of them, and it is worth knowing exactly which two before you build a listing strategy on a feed that cannot deliver.
What the model actually promises
Strip away the software and a dropship arrangement is four promises:
- you will not hold inventory;
- the supplier holds enough depth that an order is always fillable;
- an order placed today can be packed and dispatched without you touching it;
- the thing that arrives is the thing in the photograph, because the photograph is one of thousands of identical units.
The first promise is the one people want. The second and fourth are the ones that make it safe, and they are the ones that fail when every piece is singular. A carver would rather sell you a piece you have seen than promise you a copy of it — how lead times vary explains why a copy takes as long as the first one did.
Depth is not a setting
Ask a supplier how many units of a design they hold and you will usually get a number. For a carved stone range the honest number is one, and it is not a stock-management failure you can ask them to fix by ordering more.
Depth exists because a mould, a die or a line can produce unit after unit. A carving has neither. Each design is cut from its own block, and the block runs out. What looks like thin inventory is the production method showing through.
This has a specific consequence for a feed: quantity is not a field you can top up. A number that goes to zero and never returns is not inventory, it is a sale. Any listing model that treats zero as "restocking soon" will be permanently wrong about the design.
The listing problem: a SKU that does not come back
Dropship catalogues are built on repeatable SKUs. The same code, the same photograph, the same specification, sold over and over, with a quantity that moves up and down. That structure is what makes a feed possible: your store and the supplier's warehouse can both hold the same row and agree about it.
A one-of-a-kind range inverts it. One SKU maps to one object, and when that object sells the row is finished — not out of stock, finished. So the correct lifecycle for a listing is different, and it is worth writing down before you import anything:
- Listed. One piece, photographed, with its own weight and dimensions. Quantity is one.
- Sold. The listing is retired, not paused. It is a record of a piece that has been sold, and the photographs stay useful as a portfolio.
- Replaced, if ever. A new carving of the same design appears as a new listing with new photographs, because it is a different object with a different weight and different colour. Reusing the old photographs would be the single most damaging thing you could do to your own returns rate — why a second batch differs is the same problem from the buying side.
Done that way the catalogue stops being a shop and becomes an archive with a shop attached, which is roughly how galleries work and is a perfectly good way to sell singular objects. What it cannot be is a feed.
The failure that costs the most: selling one piece twice
The expensive failure in this model is not "out of stock". It is two buyers, in two places, buying the same piece in the same hour.
That happens whenever the quantity in your listing came from anywhere other than the person who can actually put hands on the object. A cached number, a number you updated last week, a number the supplier emailed you: all of them are guesses by the time a customer clicks. With fungible stock a stale number means a short delay and a substitute. With a single carving it means two refunds, a chargeback risk and a customer who has now told people you sold something that did not exist.
The fix is a rule rather than a tool: availability for a specific piece is confirmed per order, not published as a standing number. That sounds slower than a feed and it is. It is also the only version of the promise you can keep.
Your label on the outside
The part of dropshipping that does still work is presentation. A customer who buys from you should receive a parcel that looks like it came from you: your name on the outer carton, your paperwork, no supplier invoice and no supplier branding for the buyer to read.
Ask for that explicitly rather than assuming it, because the default export carton often carries the factory's name and a commercial invoice taped to the outside showing what you paid. On a gift that is an unpleasant surprise; on a resale it is your margin, visible.
The inside of the parcel is a separate question and one you should settle in writing before an order, because presentation packaging is a decision each buyer makes differently — agreed per order, confirmed on the quote, not assumed from photographs.
Returns get worse, and you should price that in
A dropship return is the worst version of a return: the piece has already travelled once, it comes back to you rather than to a warehouse you control, and — because there is no second unit — it cannot be restocked and resold as new.
So three things follow. Returns have to be built on refunds rather than exchanges, because there is nothing to exchange against. The piece needs to survive two journeys, not one, which is a reason to care about transit protection more than a normal seller does. And a damage claim is not a return at all: it is a claim against a carrier, on its own clock, and it needs photographs taken at the moment of unpacking rather than a week later.
Returns and damage when every piece is unique works through the policy side in detail, and our own returns page states the version we commit to: fourteen days, buyer pays the return freight, and a damaged piece refunded rather than swapped.
Three arrangements that do work
None of these is a feed. All three are honest about where the piece physically is.
Hold a little stock yourself. Buy ten pieces across three or four designs, ship from your own shelf. This is the arrangement that most closely resembles dropshipping from the customer's side — fast dispatch, your packaging, your label — and it costs the price of ten carvings rather than the price of a warehouse. It is also the only one where you have personally handled what you are selling.
Sell, then have it carved. Works for custom and OEM work, where nothing exists until the order arrives, and it is genuinely zero-inventory. The cost is time: the piece is made after it is sold, so the lead time you quote has to include the carving and not just the freight. See the MOQ guide for how a quantity order is built.
Catalogue plus confirmation. You show a range, the buyer picks, you confirm the specific piece and its weight and colour, and it ships. Slower per order, and it suits higher-value pieces where the buyer wants to know exactly what is coming. It is the arrangement our own stock range is built around: each design is a single carving, photographed, and one of each is held.
What to ask before you wire anything up
If you are evaluating a carving supplier for a no-inventory model, these five questions separate the ones who can deliver from the ones who cannot:
- Is the quantity in your feed per object, or per design? Those are different numbers and only one of them is safe.
- Who confirms availability, and when? If the answer is "the feed", the feed will eventually be wrong.
- Can you ship without your invoice and without your branding on the outer carton?
- What lead time applies to a repeat of the same design, and will the new piece match the photographs? Ask them to say plainly that it will not, and to tell you what will differ.
- Who pays when a piece arrives damaged on a journey you arranged? Get this in writing before the first order, not after the first claim.
A supplier who answers all five without hesitating is describing a process. A supplier who answers two of them is describing a hope.
Tell us what you are planning to sell and where the parcel would go, and we will tell you which of the three arrangements fits — including when the answer is that a feed is the wrong tool for a range like this.