Can your catalogue survive the move? Five things you can check on your own listing, today.
Whether a store of your own is the right next spend is answerable, not a matter of opinion.
Amazon owns the relationship. You own the product.
Every order you fill on Amazon is Amazon's customer. Not as a technicality — the email, the purchase history, the right to contact them again, and the terms under which any of it continues are all decided by someone else.
That is a workable arrangement, and for many sellers it is the correct one. It has real advantages you do not get elsewhere: traffic that arrives without being bought, and a fulfilment network you did not build.
What it costs is structural rather than visible. You cannot email the person who bought twice last year, because you do not have their address. You cannot change how your product is presented beyond what the template allows. And if the category rules change, they change for you.
The question is not whether Amazon is bad. It is whether the parts you cannot control are parts you now need. That is a question about your catalogue and your customers — and it is answerable.
Five things you can check on your own listing, today, without me
Every one runs on the public page. No seller-central access, no call, no form.
- Count your variants and how they relate. Open your listing and write down every size, colour, pack-size and bundle. Then check whether any option combination is unavailable, or priced differently for a reason the page does not explain. A catalogue whose variant logic only makes sense inside Amazon's structure is the single most expensive thing to move — it is the part that becomes bespoke work rather than an import.
- Read your own product copy as a stranger. Marketplace copy is written for a marketplace search box, in a fixed template, against character limits. Open your main listing and ask: if this text were on a page of its own, with no Amazon around it, would it still sell the product? Usually about half of it does, and the other half was doing a job the marketplace was doing too.
- Count your reviews, then check what would transfer. Assume none of it transfers automatically, because it does not. Look at how many reviews you have, how many are recent, and how much of your listing's persuasion is doing work those reviews do. If the reviews are the reason people buy, a new store starts without the reason. That is solvable — and it is a real cost, so count it before you move.
- Check what your product data actually contains. Do you have your own photography, or the supplier's? Full dimensions, materials, care instructions, compliance details — held somewhere you control, or only typed into a listing form? A store needs more product data than a marketplace listing does, and the gap between the two is usually the longest part of a migration.
- Search your own brand name on Google, not on Amazon. If people are searching for you by name and landing on Amazon, that demand already exists and you are renting the destination. If nobody is searching for you by name, a store will not create that demand — it is a place to send traffic you already have or intend to buy.
What the answers mean
Simple variants, copy that stands alone, product data you hold, and people searching your name — the move is mostly mechanical, and the work is in the build rather than the migration.
Complex variant logic, copy written for the search box, reviews carrying the persuasion, and no brand search — the move is a project, and the honest first step is deciding what you are actually buying with it.
The parts that are genuinely harder
These are not reasons to stay. They are the things that surprise people, and they are cheaper to know now. Each one has a shape to its answer.
- Setup is a real project, not a weekend.
- Products, variants, payments, tax settings, shipping rules and a theme that holds them. The catalogue import is the part people budget for; the shipping and tax configuration is the part that takes the time. Handled by doing it once, deliberately, at build time — not discovered per order afterwards.
- Shipping stops being solved for you.
- If you use FBA, fulfilment, returns and delivery expectations are currently Amazon's problem. Off Amazon they are yours — rates, carriers, packaging, and what you promise on the product page. And they become yours to set: free shipping over a threshold, your own returns window, a delivery promise you decide. Configuration rather than difficulty — and many sellers keep FBA running alongside.
- Tax gets more complicated before it gets simpler.
- The marketplace handles collection in most jurisdictions; your own store means that is a decision you make and maintain. Set up per jurisdiction at build time, with the platform's own tax service doing the collection.
- Traffic does not come with the store.
- This is the big one. Amazon's traffic is what you are leaving, and a new store starts at zero unless you already have brand search, an audience, or budget to buy attention. This is the one to answer before building, not after — check 5 above. If none of the three is true today, the honest answer is not yet.
- Reviews start again.
- See check 3. Handled by importing what can be imported, and by building the review capture in from launch rather than adding it a year later.
- Marketing changes shape.
- On Amazon you reach shoppers already there. Off it, you are marketing to a list you build — slower to start, and the thing that compounds. Which is the reason to own an address in the first place.
A store built to carry a catalogue from a marketplace runs against the same published checks as anything else I build — the same performance target, the same accessibility conformance, the same conversion-path walk, the same theme-update resilience.
Nine of the twelve you can run yourself, on any store, including one I did not build. The full list.
What the work involves
Getting the catalogue out with its structure intact — variants, options and the relationships between them, in a shape a store can use rather than a flattened export.
The theme, built rather than installed. A section architecture your team can rearrange without a developer, a type and colour system that holds, and the performance budget set at the start rather than recovered later. This is the part that decides what the store costs to own in year two.
Building the product page the listing could not be — your own photography, the full product data, and the comparison, bundle or configurator logic the template refused.
Campaign landing pages, built into the theme — so a page you send paid traffic to inherits the same performance work and the same checks as everything else, instead of loading a second rendering system on top. What that involves.
The customer data, in your hands — accounts, list capture, post-purchase email and the segmentation that follows. Set up at build time, because retro-fitting it means a year of orders with no addresses attached.
The shipping, tax and fulfilment configuration, set up deliberately rather than by default.
The measurement, before launch — so the store can answer whether it worked. And the published checks, before it ships.
What this is not
Not a switch, and not an urgent one. If the check above says your demand is Amazon's, moving early makes the store look like a failure it is not. Not an Amazon exit — most stores run alongside the marketplace.
The questions this decision actually raises
What do I actually get that I don't have now?
An address you own — your customer list lives in your database, so you can export it, email it and segment it. Check what Amazon lets you export today and compare the two; that difference is the answer to this question. A page that can do what your listing template refuses: a bundle mechanic, a comparison, a size guide that behaves, a configurator. Name the one thing your listing cannot do, because that is usually the first thing built. Somewhere to send paid traffic, in the form of campaign landing pages built into the theme rather than bolted on. And the second sale — post-purchase email, replacements, reorders, the launch you tell people about, all of which need an address you own.
Do I have to leave Amazon?
No, and most people do not. The common arrangement is both: the marketplace for reach, the store for margin, repeat customers and the things the listing template will not allow. That changes what the store has to be good at. A store that is your only channel must win traffic. A store running alongside Amazon mostly has to convert people who already know you, which is a smaller job and a cheaper build.
How do I get traffic if I'm not on Amazon's search?
This is the real question, and it deserves the honest version: the store does not come with traffic. There are three places it comes from and you can check which you already have. Search your brand name on Google — if people look for you by name, that demand exists and currently lands on Amazon. Your existing customers, if you can reach them. And paid traffic, which is a budget decision and the reason the landing-page work matters. If none of the three is true today, a store is a cost before it is a channel.
What about shipping and tax — Amazon handles that now?
It becomes yours, and it is configuration rather than difficulty: carrier rates, zones, packaging, what you promise on the product page, and tax collection set up per jurisdiction. It takes longer than the catalogue import, which is the part people under-budget. It is also the part that stops being someone else's policy — free shipping over a threshold, your own returns window, a delivery promise you decide, none of which a marketplace lets you set.
Isn't managing a platform more work?
Yes. Some of it is work you are already doing and some is new. New: the store itself, the checkout, the apps you add, and keeping it current through platform updates. That last one is check 12 of the standard and it is a real recurring cost. What offsets it is that the work is yours — a listing change you cannot make is not less work, it is work you are not allowed to do.
What if my catalogue is complicated?
Then it is a project rather than an import, and check 1 above tells you which. Complex variant logic is not a blocker — it is the thing worth building properly, because it is the part a marketplace template flattened. What it is not is a reason to move faster. A complicated catalogue moved badly is more expensive than the listing it replaced.
Send me the listing
I will run the five checks above against your Amazon store and write up what I find — what moves cleanly, what becomes bespoke work, and what a store would need to be good at given where your demand actually comes from.
What this is not: a sales document with findings attached. If the answer is "stay where you are for now", that is what it will say.