Marketplace or your own storefront? The honest scorecard

Every few months a founder sits across from us with a spreadsheet, a marketplace that pays their rent, and a dream of burning it down. "We want our own store. We want to own the customer." Sometimes they're right. Sometimes they're about to trade a loud, profitable room they dislike for a quiet, expensive house they can't fill. The decision deserves better than vibes — so we score it.

Our first piece on this topic did the fee arithmetic in detail; if you haven't run those numbers, run them first. This article is what we do next in an e-commerce engagement: a weighted scorecard that prices the things the fee table misses — brand, data, discovery, ops — and forces the hybrid question that's usually the real answer.

Why a scorecard and not a spreadsheet

The fee maths answers "what does a sale cost on each channel?" That's necessary and dramatically insufficient, because the channels differ most on things that don't per-order well: who owns the customer relationship, whose brand is being built by each transaction, and what happens to you when the platform changes its mind.

So we score seven criteria, weight them for your business, and make every score defendable out loud. If a stakeholder can't argue a number out of evidence, the number is a mood.

The seven criteria

Score each 1–5, marketplace versus owned. The weights below are our default for an early-stage consumer brand; yours will differ, and the weighting conversation is half the value.

1. Contribution margin per order (weight: high). Not revenue — contribution after fees, fulfilment, and the acquisition cost of that order. Marketplaces include discovery in the fee; owned stores must buy it. Include the amortised cost of the content and paid media it takes to fill your own funnel, or you're comparing a rent to a fantasy.

2. Data ownership (weight: high). On a marketplace you get the order; the platform keeps the relationship. On your own store, every session, cart and repeat purchase is yours — the raw material of email, lifecycle and growth work. Score this by what the data is worth to your category: a replenishable product (coffee, supplements, pet food) lives on repeat purchase, so data is a high score. A buy-once product (a suitcase) lives on discovery, so data matters less than you think.

3. Brand surface (weight: medium). On your own domain the brand does the talking: art direction, pacing, story, tone, the bundles and rituals a marketplace grid can't represent. On a marketplace you are a row with a thumbnail, one filter from invisibility. Score high for owned if your positioning is the product; score low if you're selling a commodity on specs and price.

4. Demand discovery (weight: high). Marketplace shoppers arrive with intent and a search box; you didn't have to build that audience, and for a young brand that's worth more than any feature. Owned discovery is rented attention you re-rent monthly. The honest question: do you have a working organic engine — search, social, email, press — or is "our customers will find us" a hope wearing a strategy's clothes?

5. Operational load (weight: medium). An owned store is a business inside your business: site, analytics, email, CRO, customer service, returns. If the founding team can't staff that well, scoring operations honestly sometimes answers the whole question — we've recommended staying on the marketplace specifically because the team's strength was product, not plumbing.

6. Platform risk (weight: medium). Fees creep, algorithms reshuffle, categories get saturated with sponsored listings, accounts get suspended by bots and restored by prayers. On your own domain the risks are different — ad costs rise, SEO redraws the map — but nobody can repo your storefront. Score marketplace lower here if you're in a category the platform itself is entering.

7. Speed and optionality (weight: low, but real). A marketplace listing can be live in a week; an owned store worth building takes two to four months. But optionality runs the other way too: an owned store can add subscriptions, bundles, gifting and content a marketplace structurally cannot. Low weight, until one of those options is your growth plan.

Running it: a worked example

A fictional candle brand — call them Half Past Six — does $900k a year, 80% through a marketplace, with a loyal repeat base they've never been allowed to email.

Criterion	Marketplace	Owned store
Contribution margin / order	3	3
Data ownership	1	5
Brand surface	1	5
Demand discovery	5	2
Operational load	5	2
Platform risk	2	4
Speed & optionality	4	3

Weighted, this lands almost exactly even — which is the correct answer, and the founders hate it. "Almost even" means both. The scorecard's real output isn't a verdict; it's the shape of the hybrid. Candles are replenishable, gift-driven and brand-heavy, so the recommendation wrote itself: keep the marketplace as a discovery channel for the hero scents, build the owned store around what the marketplace can't do — subscriptions, gift sets, seasonal drops — and harvest the repeat base the marketplace kept anonymous.

Contrast with a fictional parts supplier whose customers buy a bracket once and never again: their scorecard tilts hard to the marketplace, and we told them to invest the store budget in better listings and photography instead. Same framework, opposite answer. That's the point.

The hybrid plays the scorecard points to

Pure plays are ideological. The scorecard usually lands you in one of these patterns:

Marketplace for discovery, owned for margin. Searchable hero SKUs stay where strangers shop; subscriptions, bundles, limited runs and gifting live exclusively on the owned store. This is the pattern behind the Fernleigh Wines storefront — marketplace listings kept working as a top-of-funnel while the cellar club, which needed data and brand surface the marketplace can't offer, lived on the domain.
Marketplace as the clearance valve. End-of-line and seconds move through marketplace channels without poisoning the owned store's tone or price integrity.
Owned store as flagship, marketplace as logistics. All brand investment goes to the domain; the marketplace simply fulfils the demand it generates, treated as a wholesale line.

What we push back on: marketplace-and-owned with identical catalogues. Same product, two prices, no differentiation — you haven't built a strategy, you've built a price-comparison race against yourself, and the channel with the borrowed trust will win it.

What changes the score over time

The scorecard isn't a one-time ritual; it's an annual one. Three things move the numbers:

Your owned traffic compounds. Email lists, organic search and repeat purchase lower the real cost of owned-channel orders every year. A scorecard run in year one and never revisited will keep you a tenant long past the point you could afford to own.
Marketplace fees ratchet. Referral fees, ad-auction pressure inside the marketplace, fulfilment surcharges — the trend on every platform is one direction. Re-run the margin row annually with current numbers.
Regulation shifts the data row. Every privacy change makes owned first-party data more valuable and platform-rented targeting more expensive. The data criterion's score has risen every year we've run this framework, and we expect that to continue.

Key takeaways

Run the fee maths first, then score what the maths misses: data, brand, discovery, ops, risk, optionality.
Weight the criteria for your category. Replenishable products live on data; buy-once products live on discovery.
"Even score" is a real answer — it means hybrid, and the shape of the hybrid follows from which criteria diverged hardest.
Never run identical catalogues on both channels. Differentiate by format: marketplace for hero SKUs, owned for bundles, subscriptions and gifting.
Re-run the scorecard annually with current fee and traffic numbers. Owned economics compound; marketplace economics ratchet.
Sometimes the honest recommendation is "stay put", and a good agency will say so.

FAQ

Can't we just launch a store cheaply and see what happens? You can, and everyone does; the result is a beautiful empty store that costs you $2,000 a month in apps and attention to keep on life support. A store is a channel you must feed — content, email, paid, SEO. Score whether you can feed it before you build it, or you'll spend a year learning what a spreadsheet knew in an afternoon.

Does owning the customer data really matter that much? For replenishable categories it's the whole game: repeat purchase driven by email and lifecycle flows is the difference between a brand and a listing. For one-purchase categories, honestly, much less — the marketplace's discovery is what you were renting, and that's fine. The score exists precisely so you don't pay for ownership you'll never use.

What about social commerce and retail? Score them with the same framework — they're just more channels. Social commerce is discovery-heavy and data-poor; retail is brand-surface-rich and data-hostile. A brand with three healthy channels usually has each one doing a different job on the scorecard, on purpose.

How long until an owned store beats marketplace margins? Typically year two to three, if — and only if — organic and email reach escape velocity before paid acquisition eats you. The technical SEO groundwork and lifecycle email have to ship with the store, not after it; they're the compounding the whole model depends on. If that timeline makes you queasy, let's talk about what a staged build costs and looks like before you commit.