The familiar pitch is persuasive: “We pay the marketplace too much. On our own site, we would keep that money.” Part of the fee may disappear. The job of acquiring, reassuring and serving the customer does not. A route can have a cheaper transaction and a more expensive sale.
Compare one product on the same basis
Fictional planning example: Loomark sells an $80 desk kit. The brand, demand, costs and scenarios are invented teaching inputs, not Aright results or a named platform's current rates.
Compare one market, the same product and price, and an initial paid, delivered one-unit order. Assume goods cost $30 and outbound fulfilment $8. There are no discounts, failed deliveries, cancellations or partial refunds in this simplified model. Costs exclude tax/duty effects, fixed overhead and other unlisted items.
The hypothetical marketplace route costs $12 per order in combined transaction/payment and other stated route fees, plus $6 in media acquisition cost per delivered order. Do not add another payment charge to that $12. The owned route has a $3 payment cost and initially needs $20 of acquisition spending per delivered order, including spending on visits that do not buy. All those case fees and acquisition costs remain paid after a refund; actual provider refund rules must be checked separately.
Both routes use the same 5% full-return assumption over the original delivered orders. A return adds $6 freight and $2 inspection, offset by $8 assumed net recoverable value after future recovery costs. The extra reverse costs net to zero in this example; the $80 sales reversal remains. Thus the expected return effect is 5% × $80 = $4 per initial delivered order. This is a planning expectation, not an additional charge on every actual order, not a denominator of retained orders, and not cash already recovered.
| Expected contribution per initial paid, delivered order | Marketplace | Owned checkout |
|---|---|---|
| Selling price before expected refund effect | $80 | $80 |
| Goods cost | −$30 | −$30 |
| Outbound fulfilment | −$8 | −$8 |
| Combined marketplace route fees / owned payment fee | −$12 | −$3 |
| Acquisition spending per delivered order | −$6 | −$20 |
| Expected full-return effect under identical assumptions | −$4 | −$4 |
| Expected variable contribution | $20 | $15 |
The owned route saves $9 in transaction-related charges but spends $14 more acquiring the order. It finishes $5 behind. The mistake is not building a website. It is claiming that fee savings alone establish a stronger business.
Etsy's fee policy illustrates why real marketplace bills need their definitions: transaction, payment, listing and advertising services have distinct terms. None of the amounts above represents an Etsy fee quote, and the fictional refund treatment is not its policy. Use the actual provider, country, account and service definitions before applying the calculation.
The customer-acquisition threshold is useful, but incomplete
On these assumptions, the owned route matches the marketplace's $20 variable contribution when its acquisition cost is $15:
$80 − $30 − $8 − $3 − $15 − $4 = $20.
That is variable-contribution parity, not company break-even. It says nothing about the website's build cost, monthly maintenance, additional customer-service work or the cash needed before receipts arrive.
Suppose an owned-site plan introduces $600 a month in additional recurring software and maintenance costs. At an acquisition cost of $10, the owned route would contribute $25 per initial delivered order, or $5 more than the marketplace route. Under the same assumptions, $600 ÷ $5 = 120 comparable shifted orders per month would cover that recurring cost difference.
At exactly 120, the incremental benefit is zero before one-off build costs and other omissions. It is not evidence that 120 customers will migrate, that acquisition can remain at $10, or that the marketplace orders would otherwise have occurred unchanged. If these are genuinely additional orders rather than substituted ones, the comparison needs a different counterfactual. Do not use an attractive threshold as a demand forecast.
Control may still justify a portal with a different job
Loomark has choices beyond “leave the marketplace” and “have no owned presence.”
| First investment | What it can accomplish | What must justify the cost? |
|---|---|---|
| Keep the current marketplace acquisition route | Continue a route whose actual economics can be inspected | Its contribution, customer outcomes and dependence risks remain acceptable |
| Build a focused owned product/service destination | Explain configuration, provide product evidence and support existing customers | Specific buyer/service tasks are not being served adequately elsewhere; the team can maintain the destination |
| Add owned checkout for a bounded offer | Test whether a defined audience can buy and be served economically | A credible acquisition route, complete transaction path, customer-remedy funding and comparison method |
| Move the business primarily to owned checkout | Concentrate control and investment in a proven route | Evidence that the route can support the needed volume, service and whole-business economics—not only a lower fee |
A product explanation may be valuable even when the marketplace remains the buying route. An existing customer may need care information or a precise replacement part. A trade prospect may need a quotation rather than a consumer cart. These are real reasons to own a destination, provided the work has an owner. They should not be dressed up as evidence that paid consumer acquisition will be cheaper.
Nor does “owning the customer” mean unrestricted access to or use of personal data. The actual customer relationship, consent, channel contracts and permitted communications must be respected. Do not make a migration plan depend on copying marketplace customers into a new marketing list or directing transactions in ways the applicable platform terms prohibit.
Make the first investment answer a real uncertainty
For Loomark, keep the inspected marketplace route while building only the owned explanation and service capability that buyers need. If there is a credible defined audience for owned buying, authorize a limited checkout test with an explicit cost ceiling, operating limit and remedy backing. A budget authorization is not a provider-enforced spending cap; actual controls must be checked.
Record acquisition spend, initial delivered orders, mature refunds, service costs and actual route charges on compatible boundaries. Separate repeat and new customers; a cheap repeat purchase does not prove the first customer was inexpensive to acquire. An unpaid visit still has content, staffing or partnership costs somewhere in the business.
The recommendation would change if repeat demand is established, marketplace restrictions make a required service impossible, a different offer is economically viable on the owned route, or actual marketplace costs rise. It could also change the other way if maintenance overwhelms the team or the owned site attracts the wrong customer task.
Do the calculation for one product and one audience before accepting the promise of saved fees. Then write down what the portal is for. Control is useful when it lets the business do something valuable and sustainable; it is not a substitute for explaining where the next customer will come from.