KVY TECH CO LTD

KVY TECH CO LTD Senior-led engineering partner building and operating production-critical software systems.

Choosing Medusa because it is flexible is not a decision.Almost every technical stack is flexible in the brochure. The u...
16/09/2026

Choosing Medusa because it is flexible is not a decision.

Almost every technical stack is flexible in the brochure. The useful question is what that flexibility asks you to own. With Medusa, your team owns more of the commerce logic, integrations, deployment decisions, upgrades, and production accountability.

That can be an asset when transaction rules are part of the product. It can also become expensive complexity when your requirements are mostly standard retail.

Three questions usually make the answer clearer:
▪️ Is the marketplace model proven enough to justify owning infrastructure?
▪️ Are your important workflows genuinely unusual, or merely preferences a mature SaaS already handles?
▪️ Who owns the system in year two: an internal team or an accountable development partner?

Medusa is not “good” in isolation. It is good when its ownership model matches yours.

Which matters more for your next stage: speed to learn, or control over the system?

Commission and payouts get treated as one problem in most marketplace builds. They are two, and merging them is how mone...
14/09/2026

Commission and payouts get treated as one problem in most marketplace builds. They are two, and merging them is how money logic quietly rewrites its own history.

Commission is order logic. The marketplace's cut should be calculated and recorded on the order at a defined moment, usually creation or completion, using the rate that applied at that moment. Skip that and derive it later from current settings, and every rate change, every promotion, every vendor moving to a new tier silently changes what past orders earned. Finance tends to discover this during an audit, which is the worst possible time to find out.

Payouts are money movement, which is a payments infrastructure problem with a different shape. One rule keeps it sane: payout ex*****on reads the commission that was recorded and never recomputes it. Timing is a policy rather than a constant, so make it configurable from the start. On fulfilment, on delivery, after the return window closes; whichever you pick today will eventually change.

There is a regulatory layer here too. If you operate where holding other people's money is licensed activity, route funds through a licensed provider and keep your platform outside the flow.

Last week's question was who owns the customer. This is the same category of decision, and there are more of them hiding in a multi-vendor build. The vendor dashboard is the one that hurts budgets most often. Medusa's Admin serves the platform operator, so vendors need their own interface, and that interface arrives with design, authentication, permissions and support attached. Teams scope it after the backend is finished and find they budgeted a screen and built an application.

We put the whole picture in one place, no code: the three architecture patterns and the single axis separating them, order splitting and the grouping record worth borrowing from Mercur, and five pitfalls that only appear after launch. Refunds across a split order is the one most teams meet the hard way.

Full overview in the first comment 👇

You've read the Marketplace Recipe, skimmed the tutorials, looked at Mercur, and still can't picture the whole thing. Th...
11/09/2026

You've read the Marketplace Recipe, skimmed the tutorials, looked at Mercur, and still can't picture the whole thing. That isn't you. Nobody has drawn the map.

The material is genuinely scattered. The official recipe demonstrates one approach, tutorials demonstrate another, starters embody a third, and none of them explain how the approaches differ or why you would pick one over another.

What helped us most was realising that all three answer the same four questions, just differently. Who are vendors in your data model. What can they touch. How does one cart become several orders. How does money reach sellers.

And underneath those four sits a single question that quietly decides the other three: who owns the customer.

If a buyer belongs to the marketplace, they get one account and one cart spanning every vendor, and marketplace wide personalisation becomes possible. If a buyer belongs to the vendor, you get clean isolation between sellers, which is exactly right for franchises or white label networks, but a unified marketplace experience now swims upstream against your own architecture for as long as the platform exists.

Notice that neither answer is technical. It is a business model decision wearing a code decision's clothes, and it usually gets made by accident, in week two, by whoever set up the first module.

Answer it deliberately and two of the three patterns disqualify themselves before you write a line of code.

If you have built multi-vendor on any framework: did you decide that consciously, or did you inherit it from whatever example you started from?

Most "Medusa vs Shopify" comparisons quietly answer a different question than the one you asked.They compare the two as ...
09/09/2026

Most "Medusa vs Shopify" comparisons quietly answer a different question than the one you asked.

They compare the two as platforms for a store: one seller, one catalog, one checkout. That is useful if you are opening a store. If you are building a marketplace, the honest starting point is that neither one does that out of the box. Shopify was architected for a single merchant, and Medusa ships no vendor logic at all.

So the real comparison is between two extension paths. Shopify plus a multi-vendor app, where marketplace mechanics are rented on top of a single-seller platform. Or Medusa plus custom work or a starter like Mercur, where those mechanics become architecture you own. One gets you a low ceiling fast. The other gets you a high ceiling slowly.

Path A is genuinely good at something, and it deserves saying plainly. You inherit Shopify's checkout, hosting and security, and you can be onboarding vendors within weeks without an engineering team. For proving that buyers and sellers actually want to meet, it is the cheapest credible experiment available. We recommend it regularly, including to teams who come to us asking for a custom build.

What is worth understanding is the foundation underneath. Shopify's checkout has one merchant of record, and that is you. Multi-vendor apps do not change that. They calculate each seller's share afterwards, through their own logic. Real time splitting is possible through Stripe Connect, though it arrives as a custom payment method that routes around the native checkout, with its own constraints. The workaround works. It is also the clearest picture of what is being worked around.

At ten vendors on a standard commission, none of that is your problem. The comparison is really about the point where it becomes one, and what each path costs on the way there.

One more thing worth knowing before you search: Shopify's app called Marketplace Connect is for selling your products on Amazon and eBay. It has nothing to do with building your own marketplace. The naming confuses almost everyone.

Full comparison, including the switching threshold, in the first comment 👇

A Shopify multi-vendor app isn't cheaper. It just moves the cost somewhere you can't see yet.To be clear, choosing that ...
26/08/2026

A Shopify multi-vendor app isn't cheaper. It just moves the cost somewhere you can't see yet.

To be clear, choosing that app is often the right call. For about fifty dollars a month you get a working marketplace in weeks, and for a founder who still needs to prove that buyers and sellers even want to meet, that speed is worth more than any architecture diagram.

The part that deserves more attention is what happens after it works.

As the marketplace grows, small gaps open up between how the app works and how your business works. The commission structure it almost supports becomes a spreadsheet someone reconciles every month. The vendor onboarding flow you want to improve turns out to belong to the app's roadmap, not yours. And your data settles quietly into someone else's schema, which is perfectly fine right up until the day you need to take it somewhere else.

None of these costs ever appear on an invoice, which is exactly why the app keeps looking cheap. They get paid in hours, in patience, and in options you no longer have.

The custom route has real costs too, and they arrive in the opposite order: visible, upfront, and staffed. You pay for a development budget and a team that owns the code long before you see the benefits. Nothing is hidden on that side, but nothing is rented either.

So neither path is cheaper. They charge you differently, and at different times. Which leaves the only question worth asking: are you paying for convenience today, or for control tomorrow?

Build vs buy a marketplace: who should actually build custom, and who shouldn't. An honest framework.Before reading any ...
19/08/2026

Build vs buy a marketplace: who should actually build custom, and who shouldn't. An honest framework.

Before reading any advice on this, run one experiment. Search "build vs buy marketplace" and skim the top ten results. A pattern shows up fast:

Every article that says BUY is written by a company selling marketplace software. Every article that says BUY THE BACKEND is written by a company selling backends. Every article that says HYBRID WITH EXPERT HELP is written by an agency selling expert help.

The advice isn't wrong, exactly. The conclusion was just written before the article was.

Full disclosure: a dev agency wrote this guide too. Which is why it's structured to be useful even if you never talk to us. For a meaningful share of readers, its honest conclusion is "don't build custom, at least not yet."

The core reframe inside: build vs buy is the wrong question, because a marketplace isn't one system. It's payments, search, catalog, vendor operations, commission logic, buyer experience. Each component gets its own answer.

Payments? Buy. Regulated, solved, zero differentiation.

Search? Buy. Hard to build well, and buyers only notice it when it's bad.

Commission and payout logic? Flat rate, buy it. Tiered, negotiated, category-dependent? That IS your business model. Own it.

Custom marketplace development never meant building everything. It means owning the three or four systems where your model differs from everyone else, and renting the rest.

The full component-by-component framework is in the first comment 👇

Most marketplace builds don't fail because the team can't code. They fail because the data model was wrong on day one.A ...
14/08/2026

Most marketplace builds don't fail because the team can't code. They fail because the data model was wrong on day one.

A multi-vendor platform in SEA came to us in exactly that state.

On the surface: bugs, slow releases, an ops team drowning in manual fixes. Underneath: payments, vendors, and orders had been modeled as one tangled structure. Every feature touched all three. Every deploy risked breaking checkout. The team had stopped shipping anything near the payment flow out of fear.

Here is the part nobody enjoys. The fix they asked for was more features. The fix they needed was to stop building features entirely.

We told them that. Separate the three domains first. Vendors, orders, and payment flows, each owning its own model. Weeks of work that ships nothing a customer can see.

It feels like going backward. For a while, it looks like going backward.

Then the structure settles, and something changes: new vendor features ship without touching payment code. Deploys stop being scary. The roadmap that was frozen starts moving again.

That is the entire point of getting the model right. Not elegance. Safety to change things.

If every feature on your marketplace feels risky to ship, the code is probably not the problem. And if you want a second pair of eyes on your architecture, request a marketplace technical audit. Our inbox is open.

Medusa core ships no multi-vendor anything. No vendor entity. No order splitting. No commission engine. No vendor dashbo...
12/08/2026

Medusa core ships no multi-vendor anything. No vendor entity. No order splitting. No commission engine. No vendor dashboard.

That surprises more founders than it should because "Medusa supports marketplaces" is technically true. The official docs even provide a Marketplace Recipe. But it's an example of an approach, not a feature you switch on.

So the real question was never whether Medusa can run your marketplace. It's which of the three routes gets you there and what each one quietly costs.

The 5 slides below map all three: custom build, Mercur, community plugins. Who each route fits, what you trade away, and who ends up owning your roadmap.

One thing you won't find in them: a push toward the route we'd get hired for. We build on both custom and Mercur no horse in this race except fit.

The full decision guide including the honest cases where you shouldn't choose Medusa at all is in the first comment 👇

Medusa's docs answer every question except the one that costs the most: whether you should use it at all.That's not a cr...
10/08/2026

Medusa's docs answer every question except the one that costs the most: whether you should use it at all.

That's not a criticism of the docs. Documentation is written for people who already decided.

The problem is that almost all Medusa content works the same way. Tutorials, recipes, walkthroughs. How to build the vendor module. How to split orders. How to wire payouts.

But the expensive mistakes in marketplace builds don't happen in the code. They happen in the decision before the code and that decision has three questions no tutorial will ever ask you:

Is your model proven enough to justify owning a platform? A no-code SaaS will teach you whether buyers and sellers actually show up for a fraction of the cost of finding out on custom infrastructure.

Is your model standard enough for a starter or unusual enough that custom wins? If a Shopify-style setup already covers your flows, you'd be paying in complexity for flexibility you'll never use.

Who owns this system in year 2 in-house or contracted? Medusa without engineering ownership isn't an asset. It's a liability with a git history.

Answer those three honestly and the stack decision mostly makes itself.

Engineers and founders what's the question you wish someone had asked before your current stack got chosen?

The riskiest payment decision in a Singapore marketplace isn't which methods to support. It's whose account the money si...
03/08/2026

The riskiest payment decision in a Singapore marketplace isn't which methods to support. It's whose account the money sits in.

Cards, PayNow, GrabPay that part is solved. Stripe covers the local rails. Checkout friction in Singapore is among the lowest in the region.

The decision founders actually miss: a marketplace isn't a shop. The buyer pays once then the money must split between your sellers and your platform fee. Two ways to structure that:

Route it through a licensed provider (Stripe Connect or equivalent) the money moves through THEIR licensed infrastructure, never yours.

Or collect the full amount yourself, hold it, and pay sellers out later.

The second feels simpler and cheaper. It's also where founders can walk straight into Singapore's Payment Services Act collecting and holding sellers' money can resemble exactly the activity MAS licenses. Discovering that question after launch is the expensive version.

And it's not the only obligation that lands on the operator: under IRAS's digital-economy rules, a marketplace operator can be treated as THE SUPPLIER for certain B2C transactions. You account for the GST not your seller. Your checkout has to be designed for that from day one.

One pattern across both: compliance is cheaper as a design input than as a retrofit.
The full founder's guide covers the rest build vs SaaS by stage, what actually moves the cost, and three questions that expose whether a dev partner has ever really built a marketplace.

Guide in the first comment 👇

Address

50 No. 4 Street, Ward Hanh Thong, Go Vap District
Ho Chi Minh City
700000

Opening Hours

Monday 09:00 - 18:00
Tuesday 09:00 - 17:00
Wednesday 09:00 - 17:00
Thursday 09:00 - 17:00
Friday 09:00 - 17:00
Saturday 09:00 - 17:00
Sunday 09:00 - 17:00

Telephone

+84902261879

Website

Alerts

Be the first to know and let us send you an email when KVY TECH CO LTD posts news and promotions. Your email address will not be used for any other purpose, and you can unsubscribe at any time.

Contact The Business

Send a message to KVY TECH CO LTD:

Shortcuts

Share

Category