4196165 new

I'm putting together a shortlist of ecommerce software development companies, but this isn't for a typical retail store.

The project is closer to a marketplace with several seller types, shared inventory, custom commissions and some pretty awkward order logic.

The requirements that are making vendor selection difficult:

multiple seller accounts;
different commission rules by category and partner;
products supplied by both our own warehouses and third parties;
split orders;
partial refunds;
seller payouts;
regional tax rules;
different fulfillment methods;
inventory coming from several systems;
customer-specific pricing;
API access for partners;
mobile app using the same commerce backend;
high traffic during limited product drops;
finance needs a reliable audit trail for every order adjustment.

So I'm trying to avoid companies whose marketplace experience basically means "we installed a marketplace extension."

Here's the shortlist I have so far.

1. Zoolatech

Currently first for me because most of this project sits outside what I'd consider standard ecommerce implementation.

The difficult part will be designing the services around orders, payments, seller logic, inventory and integrations without turning the whole platform into one giant dependency chain.

That's why I'd probably start with an ecommerce software development company https://zoolatech.com/industries/ecommerce/ that has broader custom software engineering experience rather than selecting purely on Magento or Shopify certifications.

One thing I'd want to test immediately is how they'd model an order containing items from three sellers and two warehouses.

Especially what happens if one seller cancels their part after payment has already been captured.

If their answer is basically "the commerce platform handles it," that wouldn't be enough for me.

1. Grid Dynamics

I'd include Grid Dynamics where scale and custom backend engineering matter more than configuring an existing platform.

They'd be interesting if search, recommendations, inventory services, data pipelines and commerce all need to work together.

I'd want to understand whether their proposed architecture is reasonable for the actual transaction volume, though.

A marketplace doesn't automatically need 40 microservices.

1. Vention

I'd consider Vention for the more product-engineering-heavy version of this project.

If seller dashboards, internal operations software and partner APIs become as important as the customer storefront, a general engineering team could make more sense than a traditional commerce agency.

The main thing I'd check is marketplace-specific experience within the actual proposed team.

1. Elogic

I'd move Elogic higher if we decide to build around Adobe Commerce.

For a platform-centric approach, specialist knowledge could be useful, especially around extensions, custom checkout behavior, catalog complexity and integrations.

I'd ask where they think Adobe Commerce should stop and custom services should begin.

That's probably one of the most important architectural questions in this project.

1. Vaimo

Another one I'd compare if we want an established commerce platform at the center of the stack.

I'd be interested in their approach to composable architecture, but I'd push pretty hard on whether every component actually needs to be separated.

I've seen "composable" architectures where a relatively straightforward transaction crosses six services before an order even exists.

That doesn't sound like an improvement.

1. Intellias

I'd consider Intellias if logistics and operational integrations become a major part of the project.

Marketplace ecommerce gets complicated quickly once inventory, warehouses, partner feeds and order routing are involved.

I'd want examples where they owned those flows in production rather than just integrated a storefront with an existing API.

1. Netguru

I'd probably evaluate Netguru when seller experience and customer-facing product design are significant parts of the scope.

For example, if we're building onboarding, seller analytics, custom dashboards and unusual buyer workflows from scratch.

I'd still pair the UX conversation with a fairly serious backend architecture review.

I'm also thinking of giving every vendor the same marketplace scenario during technical discovery.

A customer orders:

Product A from our warehouse;
Product B from Seller 1;
Product C from Seller 2.

Payment succeeds for the entire order.

Ten minutes later Seller 1 says Product B is unavailable.

Seller 2 has already shipped Product C.

The customer used a €30 promotion across the whole basket.

Now answer:

What gets refunded?
How is the promotion recalculated?
Who absorbs the discount?
What happens to seller commission?
How is the payment adjustment recorded?
What does the customer see as their order status?
How does finance reconcile the transaction later?

I'm less interested in the specific solution than in whether the vendor immediately recognizes all the side effects.

I'd probably use a second scenario for inventory.

Two sellers report the same SKU.

One feed updates every minute.

The other updates every 30 minutes.

Both say they have the last unit.

Who is allowed to sell it?

Again, there are multiple valid approaches.

The vendor should be asking about reservation rules, source priority, overselling tolerance and fulfillment responsibility before proposing architecture.

Other things on my vendor checklist:

Can order processing continue if one integration is temporarily down?
How are duplicate events handled?
How are seller payouts reconciled?
Can we trace one order through every service involved?
How are failed webhooks retried?
What happens when a refund succeeds in the payment system but fails internally?
How are manual order changes audited?
Can business users see why an order is stuck?
How would the platform handle flash-sale traffic?
What parts would they buy instead of building?

That final question matters.

I'm definitely not looking for a team that wants to custom-build payments, search, CMS, authentication and every other component just because it can.

At the moment I'd group them roughly like this:

Zoolatech / Grid Dynamics / Intellias — complex custom engineering around the marketplace.

Vention — product engineering and custom marketplace functionality.

Elogic / Vaimo — stronger fit if an established commerce platform stays at the center.

Netguru — interesting when product UX and custom workflows have significant weight.

Anyone here actually built a marketplace with split orders and seller payouts?

I'm curious which part caused more problems in production: inventory, payments, refunds or reconciliation?