The real cost of building on rented land

Every marketplace pitch is the same: we bring the traffic, you bring the product. And early on, it’s a good trade — distribution you couldn’t buy, trust you haven’t earned yet, checkout you don’t have to build.

The trade quietly changes as you grow. This is a breakdown of what marketplace dependency actually costs a brand with real revenue — not as an argument to leave marketplaces, but to stop living there.

Cost #1 — The fees you see

Marketplace commissions stack: admin fees, category fees, payment processing, program fees for visibility, campaign participation that’s “optional” the way rent is optional. Across Southeast Asian marketplaces, effective take rates for an active seller commonly land somewhere between 5% and 20%+ of gross once programs are included — and the direction of travel has been up, year after year.

Run the number for your own store: take last month’s marketplace revenue, multiply by your effective rate, multiply by twelve. For a brand doing Rp 200 million a month at an effective 10%, that’s Rp 240 million a year — roughly the cost of building a serious custom platform, every single year, forever.

Cost #2 — The algorithm you don’t control

Your marketplace ranking is an asset you can’t own. It reprices overnight: a search algorithm update, a new “sponsored” tier above organic results, a competitor buying placement against your bestseller. None of this is malicious — the marketplace optimizes for the marketplace. But it means your revenue sits downstream of decisions made in someone else’s boardroom.

The practical symptom: brands that feel busy but fragile. Sales are fine this month, and nobody can say why they’ll be fine next month.

Cost #3 — The customers you never met

This is the expensive one, and the least visible. On a marketplace, the buyer is their customer, not yours:

• You usually can’t email them.
• You can’t retarget them.
• You can’t see their history across purchases.
• You can’t follow up a first order into a second.

Repeat purchase is the cheapest margin in e-commerce — no acquisition cost, higher basket, better conversion. Marketplace dependency amputates it. A brand that has sold to fifty thousand people but owns zero customer relationships has been building someone else’s asset.

What “your own platform” actually changes

A store on your own domain flips each cost:

• Fees become flat. Hosting and payment processing replace percentage-of-everything. The bill stops scaling with your success.
• Traffic becomes cumulative. SEO compounds; every article, every product page, every backlink is equity you keep. (AI assistants are becoming a real referral channel too — AI-referred orders on storefronts grew roughly 13× year-over-year per Shopify’s 2026 data — and they cite websites, not marketplace listings.)
• Customers become yours. Email, WhatsApp follow-up, abandoned-cart recovery, repeat campaigns — the whole retention playbook unlocks.

When it’s not time yet

Honesty matters more than the pitch: if you’re pre-revenue, still validating product-market fit, or doing occasional sales to friends-of-friends, stay on the marketplace. A custom platform amplifies a working business; it doesn’t create one. The signal that it’s time is uncomfortable success: fees that sting because volume is real, support that drowns because customers are real, and a growth ceiling you can feel.

The sane migration path

Nobody should flip a switch. The pattern that works:

1. Keep the marketplace as a channel — it’s distribution, and distribution is good.
2. Launch your own platform as the home base — where margins are full and data is yours. (In my pricing, a core custom store starts at $2,500 / Rp 15 million, live in 4–6 weeks.)
3. Route repeat buyers home. Package inserts, WhatsApp receipts, loyalty pricing on your domain — every marketplace order becomes an invitation to the direct relationship.
4. Automate the back office from day one so the second channel doesn’t double the workload — that’s what workflow automation is for.

The goal isn’t leaving the mall. It’s owning a house, so the mall becomes optional.

I build custom e-commerce platforms for brands making exactly this move — here’s what one looks like running, and here’s what it costs.