Selling on your own site versus Amazon, eBay and Etsy

Marketplaces are very good at one thing: putting your product in front of people who are already shopping, with their card details saved and their trust given to the platform rather than to you. That is a real service and it is why they can charge for it.

Your own shop is good at a different thing: keeping the margin, owning the relationship, and letting you sell the way your business works rather than the way a template allows. It also has to find its own customers, which is the part people underestimate.

It is easy to be sniffy about marketplaces if you have never had to build demand from nothing. What they provide is worth being clear about.

We are not going to print percentages here, because they change, vary by category and differ between platforms and countries. Read the current fee schedule for the categories you actually sell in, which every major marketplace publishes, and do the arithmetic on your own products rather than an average.

What we will say is that the referral or final value fee is rarely the whole cost. Work through this list before deciding whether a channel is profitable.

This decides the strategy, and it gets far less attention than the fees. When somebody buys from you on a marketplace, in most practical senses they have bought from the marketplace. You typically do not get their email address, you are usually restricted from putting promotional material in the box, and the rules about contacting buyers directly are strict and enforced.

So a marketplace sale is a transaction rather than the start of a relationship. That is fine for some products and a serious problem for others. If your business depends on repeat purchases, refills or a customer who comes back every season, building it entirely on a channel where you cannot contact your own customers is a strategic weakness rather than a technical one.

Your listings, your reviews and your standing on the platform are not really yours either. Accounts get suspended, sometimes wrongly and usually without warning, categories get restricted, and fees change. A firm selling on a single marketplace and nowhere else is one automated decision away from having no business that week. That is not a reason to avoid marketplaces. It is a reason not to have only marketplaces.

On your own site you keep the margin apart from payment processing, you own the customer record and the email list, you decide what the product page says and how the checkout works, and you can sell things the marketplaces will not let you sell in the way you want to sell them: bundles, subscriptions, trade accounts with their own pricing, made to order items with options.

You also get the data. What people searched for, which pages they looked at, where they gave up, what they bought together. On a marketplace you get a fraction of that, filtered through the platform's reporting.

What you do not get is traffic. That is the honest trade. Nobody arrives at your shop because it exists, and the cost of bringing them there is a real line in the budget that the marketplace fee was quietly covering for you. A shop with no plan for how people will find it is not cheaper than a marketplace. It is just paying differently, and often paying for nothing.

Stop treating it as a choice and an obvious division of labour appears. That division tends to be where retailers land once they stop treating it as a choice.

The marketplaces do discovery and volume, particularly for products people search for generically rather than by brand. Your own site does margin, range, brand and repeat business, and it is where you send everybody you have already met: the person who found you on social media, the trade buyer, the one on your email list.

Traffic can flow both ways too, within the rules. A customer who first bought on a marketplace may remember the brand, search for it directly, and buy from your shop the second time at full margin. That only happens if the product and the packaging are memorable enough, and if there is a proper shop waiting when it does.

Here is where it usually goes wrong, and it has nothing to do with strategy. Adding a channel adds a copy of everything: another stock number, another product list, another order queue, another set of customer messages, another set of fees to reconcile at month end.

Run that manually and you get the same three problems every time. You oversell, because the shop and the marketplace do not know about each other's sales until somebody reconciles them. You retype orders into your accounts and your despatch process by hand. And you lose track of the true margin per channel, because the fees land somewhere separate from the sales.

The way out is one stock pool every channel draws from, product information written once and mapped out to each platform's format, orders landing in a single list wherever they came from, and marketplace fees flowing into your accounts as costs against the right sales. That is what makes a second and third channel add revenue instead of adding a person. It is worth sorting out before you add the channel rather than eighteen months afterwards.

Do this on paper before committing to anything. Take your five best selling products and work out the margin left on each after the full list of marketplace costs above, not just the commission. Then do the same for your own site, including a fair share of what it costs to bring a visitor in.

Then ask what the numbers cannot answer. Do your customers buy again, and how often. Would you survive a month with the marketplace account suspended. Do people search for your product by name or by description. How much staff time can you really give a new channel.

For a lot of firms the answer turns out to be both, weighted differently than they expected, with the joining up done properly rather than by hand. If you want to talk it through, tell us what you sell, which channels you are on now and roughly how many orders a week. That is usually enough for us to say whether the next thing to fix is a channel or the plumbing behind the ones you have.