Marketplace or your own store? Run the real numbers

"We need to get off the marketplace." It's the sentence we hear most often from founders who are doing well on one — which is exactly why it deserves suspicion. Marketplaces feel like a landlord: the fees creep, the rules change, and some stranger's review can tank your week. Being mad at your landlord is not a strategy. Running the numbers is.

This is the framework we walk clients through at the start of our e-commerce engagements. It's arithmetic, not ideology — and it ends, roughly a third of the time, with us advising the client to stay where they are.

First, run the actual fee maths

Founders quote the referral fee offhand — "they take fifteen percent!" — and stop counting there. Marketplaces take more, and your own store takes more too. Here's the honest comparison for a fictional homewares brand selling a $62 stoneware set (all numbers illustrative):

Line	Marketplace	Owned store
Referral / platform fee	15% ($9.30)	~3% platform + apps ($1.90)
Payment processing	included	2.9% + $0.30 ($2.10)
Fulfilment / pick-pack	$8.40 (network fulfilment)	$7.80 (3PL, slower)
Customer acquisition	$0–4 (organic rank)	$11–19 (paid + content amortised)
Returns & CS overhead	~$2.10	~$1.40
Channel cost per order	~$21.80 (35%)	~$23.20 (37%)

That last row shocks people. On a per-order basis, an owned store with paid acquisition is often more expensive than a marketplace — especially in year one, before organic and email kick in. The marketplace's dirty secret is that its fee buys you real, working demand generation. Your own store's dirty secret is that demand generation is now your job, forever.

The crossover comes later. Owned channels get cheaper per order as repeat purchase, email and organic search compound; marketplace costs stay flat or ratchet upward. A five-year model usually favours owned — but only if you survive the expensive years.

What the marketplace actually sells you

Strip it down and a marketplace sells three things:

Discovery. Shoppers arrive with purchase intent and a search box. You didn't have to build that audience. For Fernleigh Wines, the wine marketplace listings kept working as a top-of-funnel long after the DTC site launched — people discovered the label where they already shopped, then hunted out the cellar-door club directly.
Borrowed trust. Prime-era delivery promises and A-to-Z guarantees are trust infrastructure you would spend years earning. A new brand on its own domain starts from zero credibility; on a marketplace it starts from the platform's.
Logistics leverage. Network fulfilment at SMB volumes genuinely beats most 3PLs on two-day delivery maths in metro areas.

What it rents you is the customer. You get the order; they keep the relationship. No email address, no browsing data, no retargeting, no voice in how you're presented — you're a row in a grid, one filter away from invisibility.

What your own store buys you

The mirror image:

Margin control. You set prices without a fee floor, run bundles and subscriptions a marketplace can't represent, and keep the spread. See bundles and kits for why this alone can be worth points of margin.
The data. Every session, cart and repeat purchase is yours — which powers email, lifecycle and real growth work. An owned email list of 20,000 past buyers is an annuity; a marketplace review count is a deposit you can't withdraw.
Brand equity. On your own site the brand does the talking: art direction, storytelling, unboxing, tone. On a marketplace the platform is the brand and you're interchangeable inventory. If your positioning matters — and if you're reading agency blogs, it does — that's the whole ballgame over a decade.

The hybrid plays that actually work

Pure plays are rare and mostly ideological. The patterns we see working:

Marketplace for discovery, owned for margin. List your bestsellers and your simple SKUs on the marketplace — the things people search for generically. Keep bundles, subscriptions, limited runs and gifting (high margin, high story) exclusive to the owned store. The marketplace becomes paid acquisition with a revenue line attached.

Marketplace as liquidation valve. End-of-line and seconds move through marketplace clearance without poisoning your owned store's price integrity or brand tone.

Owned store as the destination, marketplace for logistics. Some brands fulfil marketplace orders through the network but invest all brand-building spend in their own domain, accepting the marketplace as a cost-of-goods channel for strangers.

The common thread: decide what each channel is for, then measure it against that job. The mistake is judging your owned store's first-year ROAS against the marketplace's seventh-year organic rank. Compare like with like across a five-year horizon or you'll make a romantic decision dressed as a financial one.

When you should honestly stay on the marketplace

We advise staying put more often than our own interests would suggest. Stay if:

Your product is a commodity bought on price and delivery speed, with no story to tell (replacement filters, basic cables, industrial consumables). An owned store adds overhead and converts worse.
Your repeat rate is structurally low — nobody develops a relationship with a mattress or a tap washer. Owned-channel economics depend on repeat and email; single-purchase categories belong where the searchers are.
You're pre-product-market-fit. Marketplaces are the cheapest demand-testing rig ever built. Fifty B-listings teach you more about pricing and positioning than fifty grand of site design. Come back when the reviews tell you who you are.
Your category's shoppers genuinely never leave the platform. Some don't. Meet them there.

The modelling worksheet

Before any platform decision, we build a one-tab model with six inputs: average order value, contribution margin per order, realistic CAC for owned paid channels, expected repeat rate at 12 months, marketplace fee stack, and email capture assumption. The outputs that matter: five-year contribution per acquired customer, by channel, and the month owned-channel economics overtake marketplace economics. If you've never written your repeat rate assumption down, that's the tell — you're deciding by vibes.

If the crossover month is beyond month 30 and your cash position won't carry it, the honest move is a hybrid play, not a leap.

Key takeaways

Marketplace fees buy discovery, trust and logistics — price them as a channel, not a tax.
Owned stores are usually more expensive per order in year one; they win on five-year compounding (repeat, email, organic).
The marketplace rents you customers; your store and email list own them. Long-run brand equity lives on land you own.
Hybrid plays beat ideology: marketplace for discovery, owned for margin, story and subscriptions.
Sometimes the right answer is staying put — commodities, low-repeat categories and pre-PMF products included.
Model five-year contribution per acquired customer before you move anything.

FAQ

Should we launch on both at once? Usually no. Launch where your customers already search, prove the product and the pricing, then add the owned store when you can fund its acquisition honestly. Two half-fed channels are worse than one well-fed one.

How do we move marketplace customers to our own store? You don't — at least not by breaking marketplace terms, which can get you delisted. You win the next purchase by being memorable: packaging worth keeping, inserts that teach rather than beg, and products (bundles, subscriptions, exclusives) the marketplace structurally can't offer.

What margin do we need before an owned store makes sense? As a rule of thumb, contribution margins under ~45% make owned paid acquisition brutal. Fix the product economics first — our PDP design and checkout audit guides cover the conversion side, but no page rescues a margin that isn't there.

Is headless architecture relevant to this decision? Indirectly. If you go owned and the brand experience is the point, the flexibility matters — we covered the honest costs in headless commerce trade-offs.