Magento marketplace development agencies, ranked on published multi-vendor evidence for 2026

scandiweb takes the full 24 on the heaviest criterion, and it does it on a page almost nobody can find. Its Magento Mirakl marketplace orchestration service page opens on the problem rather than the product: running a marketplace on Magento "gets messy fast when seller onboarding, catalog ingestion, commissions, and payouts live in spreadsheets, email threads, and manual imports". It then names the mechanics one at a time. Commission: "Commission and fee rules are applied on the Mirakl side and linked to order lines, with settlement-relevant fields passed through for reporting and reconciliation." Payout: "Marketplace payment events and payout statuses from Mirakl are exchanged with Magento order and invoice references, with traceable IDs kept for audit." Finance: "We wire marketplace payment events, payout status, and reconciliation into reporting so Finance can trace every seller transaction." And the FAQ adds the line that decides whether a finance team will sign the design off: "Commission rules are calculated from Mirakl data and mapped into Magento order and invoice flows so finance can reconcile fees, refunds, and adjustments without manual rework."

The named provider and the split itself come from the other two documents, both also scandiweb's own. The Magento multi-vendor marketplace guide states the mechanism in one sentence, "the system must split the funds, deduct your commission, and route the rest to each seller on a defined schedule", says plainly that "this is where marketplaces get legally and technically serious, because you may be handling money on behalf of others", and names Stripe Connect, PayPal for Marketplaces and Mangopay as the providers that exist for exactly that. The Slow Cosmetique marketplace case study shows one of them shipped: "We fully integrated MangoPay for end-to-end payment handling between customers and vendors", and "when they pay for their orders, the sellers are immediately credited with the proceeds". So four of the five components are carried by the service page or by the named build, and one, the commission ladder with its levels, is carried only by the guide. The same rule was applied to every company here.

The seller side takes the full 18. The Mirakl page publishes the governance layer: "We design Mirakl seller onboarding with KYC steps, role-based access, and moderation trails that match your internal controls", plus catalogue plumbing that is unusually specific, "Seller catalog data is ingested from Mirakl and mapped to Magento product structures, with attribute, category, and locale-specific values aligned per mapping rules." The Slow Cosmetique build shows a vendor dashboard that actually existed: the Unirgy Vendor Portal was retrofitted as "a second admin panel for vendors", through which vendors managed their catalogue, saw their transaction history, transferred funds out of their wallet and ran their own microsite, described on the page as "an online boutique where all the products offered by a single vendor can be explored by customers". The guide supplies the approval model including the hybrid most operators end up with, "auto-approve known vendors, review new ones".

Order splitting takes the full 14. The Mirakl page covers the awkward cases rather than the happy path, "we cover the messy parts like multi-seller carts, order splitting, partial cancellations, returns, and customer service signals", and the routing itself, "Orders placed in Magento are pushed to Mirakl for order routing, splitting, and seller assignment, while order status updates are synced back to Magento." Seller offers carry "price, inventory, lead times, and shipping constraints synced and validated before publication". Slow Cosmetique shows the buyer-facing half working: customers could "add to cart products from different vendors and complete their transaction with a single checkout". That case study is also candid that fulfilment was centralised rather than split, "fulfillment is fully managed by Slow Cosmetique", which is a design decision and is scored as one.

Build or buy takes the full 12, and only Opace, Bemeir and ScienceSoft match it. The guide publishes a six-row comparison of the extension route against a custom build, covering time to launch (weeks against months), upfront cost ("roughly $99 to $700 for the module" against development time and a team), business-logic fit, payout and commission rules ("Standard, with add-ons" against "Fully bespoke"), maintenance burden and what each is best for. It shortlists the modules by name, Webkul, Amasty, Purpletree, CedCommerce and CreativeMinds, with real prices. It names the point where the shelf runs out: "Non-standard payout logic. Tiered commissions, region-specific tax handling, milestone payouts, or escrow that no module supports cleanly." And it separates the two things buyers confuse, "the Adobe Commerce Marketplace is software you consume, while a multi-vendor marketplace is a business you operate".

Platform depth takes the full 10, all of it off scandiweb's own pages. Magento services publishes "Adobe Commerce Gold Partner, Hyva Platinum Partner, 894+ Adobe certifications, 95 NPS rating", and Adobe's own Solution Partner Directory carries the Gold listing independently. The services page publishes 23+ years, 2,100+ projects, 600+ specialists and 700+ brands across 45 countries. The Mirakl page adds the marketplace-specific version, "we've delivered 2,100+ eCommerce projects and handle $4 billion+ yearly", and Slow Cosmetique is a Magento 1.9 to Magento 2 migration that came out at a page load time of no more than two seconds. In the Hyva full register scandiweb is listed once for each country it appears in, and every one of those listings is Platinum.

Three things count against all of that, and they are set out at length further down the page. The marketplace orchestration page is a Mirakl integration page, so the commission, routing and payout logic it describes executes on Mirakl and is connected to Magento rather than built natively into it. It names no client. And the named-client evidence is four years old, which is precisely where this entry gives up 11 of the 22 points on offer.

Opace publishes the most complete marketplace service page in this comparison and one of the thinnest delivery records, and both facts are visible on the same domain. Its marketplace development page, carrying a modification date of 22 June 2026 and a visible update stamp, takes the full 24 on the money criterion. Commission: "Automated commission calculation with flexible rules per vendor, category or product type", with the delivery history attached, "we've implemented flat-rate commissions, tiered percentage models, category-specific fees, subscription-based vendor pricing and hybrid models". Payouts: "Vendors receive payouts on your defined schedule (daily, weekly, monthly)." Providers: "We integrate with marketplace-capable payment providers including Stripe Connect, PayPal for Marketplaces and custom gateway solutions." And the mechanism itself, stated as the requirement rather than as a feature: "The key requirement is split-payment functionality: the ability to divide a single buyer transaction between the platform and one or more vendors automatically. We configure commission rules, payout schedules and currency handling based on your marketplace model."

It takes the full 18 on the seller side and the full 12 on build or buy. Onboarding runs from "self-service registration, product listing tools, order management, earnings reports and payout tracking" through an admin gate covering "vendor approvals, product moderation, dispute management, financial reporting and fee configuration", with an identity component most competitors omit, "vendor identity verification during onboarding", and a launch model stated plainly: "we launch marketplaces with a controlled group of founding vendors, then iterate based on real data before opening wider registration". On build or buy it names four routes, WooCommerce with Dokan, WooCommerce with WCFM, Magento with marketplace extensions, and a custom Laravel build, and prices them: "custom marketplace development typically starts from £25,000-£40,000 for a minimum viable platform" rising to "£50,000-£100,000+", against "3-5 months from initial planning to launch" and "6-9 months" for the complex version. The failure signal is its own sentence: "early marketplace decisions determine whether you can scale to 500 vendors or hit a wall at 50".

Order splitting takes 11 of 14 on one well-built sentence, "multi-vendor carts that split into separate orders per seller, each with its own shipping, tracking and fulfilment workflow. Buyers get a single checkout experience; vendors manage their own dispatch", and loses the component for returns and partial cancellations across vendors, which is not described. Platform depth takes 6 of 10 on "15+ YEARS IN BUSINESS", "500 + PROJECTS DELIVERED", a certified-team claim and a Magento 1 to Magento 2 migration. It loses the tier component: the word Adobe does not appear in the visible text of the marketplace page at all, and its Magento development page calls Opace "Your Magento Open Source and Adobe Commerce Partner" without naming a tier or a programme. One caveat on the certification component, since a reader might object to it: the certifications that page itemises are Google, HubSpot and Semrush, and the engineering claim is the broader "our certified developers and marketers are skilled in leveraging Adobe Commerce".

Named builds take 11 of 22 and this is where a buyer should slow down. The one named Magento marketplace is The Bootroom, and Opace's own description of it is unusually frank: "The Bootroom came to us in 2016 looking for an agency to provide Magento technical support. Taking on a half-finished project with a highly bespoke Magento solution using the WebKul Marketplace extension meant we had our work cut out." The page's JSON-LD publication date is 11 January 2018, it carries a 2026 update stamp, and its closing line is that the client "has a stylish, fully responsive and far more effective solution, and it's not long now until the launch date", so the page as published does not record a launched marketplace. The second marketplace it publishes, Bazar, is WordPress with the WCFM plugin rather than Magento. No seller count appears on either. So the doctrine on the service page is excellent and current; the Magento evidence behind it is one 2018 rescue of somebody else's half-finished build.

Brainvire has the best named-build record on this page, which is exactly the criterion the page's leader is weakest on. Its multi-vendor marketplace extension page states the custom side plainly: "Beyond the extension, our Magento team builds and customizes Magento 2 multi-vendor marketplaces and marketplaces on Adobe Commerce, including vendor registration, seller approval and product approval workflows, a separate vendor dashboard, seller profiles, and ERP, shipping, and payment integrations." Behind that sit three named marketplaces. Craftsvilla, described on its Magento development page as scaling "to 2M+ products" with "moderation management, simplified vendor management and an enhanced marketplace experience across thousands of vendors". Loopify, on a Magento multi-vendor marketplace for eco goods, where "Brainvire crafted a Magento framework that brings buyers and sellers together, with an exclusive panel for orders, catalogs, shipments, and customers, and admin control over commission, smart partial shipping, returns, and custom reports". And Al Arkan, an auto-parts B2B build migrated to Adobe Commerce Cloud. A fourth marketplace carries the hardest number of all and no name at all: "A B2C multi-vendor Magento marketplace serving 2,000+ sellers and 5 lakh buyers daily." That is worth being precise about, because it is the figure a buyer would most want and the marketplace it belongs to is never identified. It is why Brainvire takes 16 of 22 rather than the full 22 here: a named marketplace, three of them, and dated evidence, but no seller count attached to any marketplace it names.

It takes 14 of 24 on the money criterion, and the gap is specific. Commission granularity is there in detail, "customize commission rates by setting different rates for different product categories, or for individual products themselves", and the delivery note behind it is better still: "For one client, we changed the admin panel, added seller-based and price-based commission, and customized the Magento import functionality." Payout and reconciliation are there through Loopify, "the admin panel shows clear payout details, so the client reviews statements before releasing seller payments". What is missing on the pages read is a named marketplace payment provider and any description of the split at checkout. No Stripe Connect, no PayPal for Marketplaces, no Mangopay, no Adyen for Platforms appears anywhere. For a marketplace that is the one integration nobody can improvise, and it costs 10 points.

Everything else is strong. Onboarding takes the full 18 on registration, seller and product approval workflows, a separate vendor dashboard, seller-data validation ("a stack that fetches and validates seller data and runs quality checks on product information") and a customised Magento import path. Order splitting takes the full 14 on Loopify's "smart partial shipping with an order split facility, so sellers view only their own orders and manage shipments effectively", plus admin control over returns. Build or buy takes 6 of 12: the extension page publishes a real decision, "a custom build suits marketplaces whose seller workflows, integrations, or scale go beyond what an extension covers", and a trigger for it, "if your vendors need a fuller dashboard, our team scopes and builds it as a customization", but it names no rival stack and gives no cost or timeline for either route. Platform depth takes the full 10 on a self-published tier and a certified-team claim in the same sentence, "a Magento Gold Partner with certified developers", plus "2,500+ Enterprise programs delivered", Adobe Commerce Cloud named as the platform on the Al Arkan migration, and scale figures on its own Magento work in "2M+ products" and "5 lakh buyers daily". One thing to check if credentials matter to you: the page says Magento Gold Partner, the retired Magento-era label, while the site footer says Adobe Solution Partner, the current one with no tier attached.

Webkul publishes more named Adobe Commerce marketplace builds than anyone else researched for this edition, and nothing at all about choosing between routes. Its Adobe Commerce marketplace development service page carries three customer stories, Camer Shopping, Kimberly Clark and Sqella, and its case-studies section carries five more, each tagged Adobe Commerce: Schneider Electric in France, IndiaHandmade in India, Tile Worx in Australia, Stockpile in Sri Lanka and Bharat eMart in India. The delivery language is unambiguous in several of them, "we helped IndiaHandmade build a scalable multi-vendor marketplace using Marketplace Multi Vendor Module for Magento 2", and the Stockpile study carries a named quote from a director of the client company. Named builds take 16 of 22: eight named marketplaces, all of them carrying a JSON-LD date of 2023 or later, and not one seller count among them.

The mechanics are evidenced through clients rather than through a specification. Commission is thinner than the field, "the admin can set the seller commission separately or can set it at a global rate for all the sellers", which is two levels and no more, and payouts are described once, "seller payments are managed through automated payout workflows that track orders, calculate commissions, and release payments via integrated payment gateways". Stripe appears as a marketplace payment module against named clients on its customer roster. What is missing is any description of the split at checkout itself, which is why it takes 19 rather than 24. Onboarding takes the full 18 on manual or automatic vendor approval, manual or automatic product approval, vendor-specific product attributes and mass upload. Order splitting takes the full 14 on Sqella, where the split order add-on let customers "purchase products from multiple vendors in a single transaction" and "generate separate order IDs, invoices, and shipments", plus per-vendor tax invoicing on Bharat eMart and vendor returns on Schneider Electric.

Two things hold it at fourth. Build or buy takes zero: its own marketplace guide was read in full on 29 September 2026 and contains no extension-versus-custom comparison, no cost or timeline for either route, and no rival marketplace stack. Platform depth takes 6 of 10: it publishes a real tier on the service page cited above, "Adobe Technology and Solution Partner Silver", and it is edition-specific, stating that it "has developed the Adobe Commerce Multi-Vendor Marketplace extensions for both Adobe Commerce Enterprise Edition and Adobe Commerce Enterprise Cloud Edition", but no certification count or certified-team claim and no project count appears on the pages read. Its CMMI Level 3 appraisal is an assessment of process, not a count of certified people, and "150k customers from 150+ countries" counts customers of a company selling hundreds of extensions. Webkul also publishes, in a testimonial on its own service page, a limit a competitor would have to dig for: a client development director writes that the setup "perfectly meets our needs" and then asks for what is not there, "it will be even greater if Webkul could provide calculated fee per vendor on shipping cart and with API". Its customer roster names enterprises against the specific marketplace module each one runs, which is a licence roster rather than a build list and is scored as one. In the Hyva full register Webkul holds Partner, the lowest of the five tiers.

Bemeir publishes the best build-or-buy analysis in this market and the best payments analysis, and it publishes no named marketplace client at all. Both halves of that sentence are why it scores 72. Its multi-vendor marketplace comparison for distributors sets Magento with a Webkul or Ced extension against Mirakl, Nautical Commerce, CS-Cart Multi-Vendor and Sharetribe across ten rows, and two of those rows are the ones nearly everyone else leaves out: "Implementation timeline | 3-5 months | 6-12 months | 2-4 months | 2-3 months | 1-2 months" and "Annual cost (platform + implementation) | $70,000 - $180,000 first year | $300,000 - $700,000+ first year | $80,000 - $200,000 first year | $40,000 - $100,000 first year | $15,000 - $40,000 first year". That is the full 12 of 12 on build or buy.

The decision rule itself sits in a second article, its marketplace platform development tool review, and it is stated rather than implied: "If the marketplace is your primary business model going forward, invest in purpose-built infrastructure (Mirakl or Nautical) that scales to hundreds of sellers with full operational tooling." The same article tabulates five payment providers with what each is for, Stripe Connect ("Platform payments with managed accounts | US-focused marketplaces | Full KYC, 1099 reporting"), PayPal for Marketplaces, Adyen for Platforms ("Enterprise split payments | High-volume, multi-currency"), Payoneer and Hyperwallet, then explains what the integration buys: "The Connect APIs handle seller onboarding, identity verification, payout scheduling, and commission deduction automatically." That is the full 24 on the money criterion.

It takes zero on named builds, and not because its work is hidden. Its one marketplace case study is deliberately anonymous, "a regional distributor of specialty industrial supplies, serving 3,200 customers across the Northeast US", it publishes real numbers ("47 active sellers, 1,840 marketplace SKUs, and $4.2M in GMV"), and Bemeir's own description of its role in it is advisory rather than delivery: "Bemeir's development team consulted on the scaling architecture." That same case study is where the commission and tax detail lives: "Different product categories carried different commission rates (8-15% depending on category margin structure). The system calculated commissions per transaction, aggregated by seller, and triggered monthly payouts via ACH", and the observation no other company here publishes, that "when a single order ships from three different seller locations to one buyer, sales tax nexus rules apply independently per shipment origin". The only named client on its Magento page is a luxury retailer, not a marketplace.

Onboarding takes the full 18 on tiered seller enablement, approval before publication ("the distributor's merchandising team reviewed listings for technical accuracy before publication") and three named catalogue ingestion routes from manual through CSV to full API. Order splitting takes the full 14 on split-cart handling with "separate shipping calculations, and independent fulfillment tracking per seller" and a dispute case-management workflow. Platform depth takes 4 of 10: Magento since 2010 and "multiple certified team members", but no tier, no project count and no performance figure, on a site whose footer still reads 2022.

eSparks publishes the clearest account of how a split payment actually reaches a seller, and it publishes it on a recent named build rather than in a services brochure. Its AI prompt marketplace on Magento 2 case study, dated 7 April 2026, explains the choice and then the wiring: "Stripe handles the full payment lifecycle and has mature support for marketplace routing through Stripe Connect, where a single buyer payment gets split automatically between the platform operator and the seller." Then the settlement: "Stripe splits each payment between the platform commission and the seller's share automatically. The seller receives their payout; the platform retains the configured fee. No manual reconciliation needed." And the detail that shows somebody debugged it: "Magento order status wired directly to Stripe payment events, orders move to processing only when Stripe confirms a successful charge, not before." It takes 19 of 24 and loses only the commission-granularity component, because the platform fee it describes is a single configured rate rather than a ladder.

Two named marketplaces sit behind it, both dated 2026. Wezzomart, where "independent sellers manage products, inventory, and orders through a dedicated vendor system" on "a hybrid architecture combining Magento and Laravel", with a custom wallet where "wallet balances are automatically adjusted once the order is completed, ensuring accurate financial reconciliation". And MySpareParts.com, "a European online marketplace built to simplify sourcing of machinery and heavy equipment spare parts across 16 countries", running "500,000+ SKUs across 16 country storefronts" from a single Magento installation, with CedCommerce's multi-vendor extension, per-category custom attribute sets, a validated CSV import pipeline and "vendor-specific tax rules configured for compliance across 16 European markets". That takes 11 of 22: two named marketplaces rather than three, and the seller figure is qualitative, "dozens of spare parts suppliers across Europe", rather than a count.

Onboarding takes 14 of 18 on a per-vendor dashboard ("vendors upload and manage products, pricing, and inventory directly within their own panel"), central product approval ("the platform owner maintains central governance, approving products, managing commissions, and overseeing compliance"), seller permissions and the catalogue work above, and loses the component for a named approval mode. Order splitting takes 7 of 14, the weakest of any entry with two named builds: order routing to the right vendor is there ("order data is immediately transmitted to the vendor panel") and shipping is there through DHL Express with live rates and tracking, but a multi-vendor cart on a single checkout is never stated and neither are returns across vendors. Build or buy takes 9 of 12 on three stacks named, CedCommerce, Amasty and a stated custom Laravel build, a criterion for going custom ("the platform's modular structure makes this feasible without modifying core files, which means marketplaces built this way can still receive Magento security and platform updates") and a precise limit, "Magento's out-of-the-box GraphQL layer lacked the endpoints required for a multi-vendor mobile platform". Platform depth takes 4 of 10 on a certified-team claim and Magento version work, with no tier and no project count on the pages read.

Sutunam publishes the deepest engineering on this page about the hardest thing a marketplace order has to do, and its Magento marketplace is its own property rather than a client's. Its custom marketplace development page states the problem without softening it. Its heading is "One order, multiple sellers, multiple product types", and the paragraph under it reads "On a multi-vendor marketplace with heterogeneous products, the basket is the easy part. The complexity is everything that happens behind it: packaging, logistics, timeframes and costs all calculated differently for each product type, within one single order." Then it shows the answer on a live site: "We solved this on Sevellia.com: a single order can contain fresh products (insulated packaging, next-day delivery, cold-chain certified carrier) and dry goods (standard carton, three-day delivery)." Shipping is "calculated per seller and per type", with "different carriers per segment" and a "Cold-chain certified carrier for fresh items". Nobody else researched for this edition publishes that at all. Order splitting still takes only 11 of 14, because returns and partial cancellations across vendors are not described.

The money criterion takes the full 24 and it is the most regulation-aware treatment in the set. "Split payment: centralised collection, automatic redistribution, commission rules by seller, category or volume." Three providers are named with their real constraints rather than as logos: "Stripe Connect, HiPay and PayPal, integrated on real marketplace projects. Each solution has its own constraints: Stripe Connect imposes its connected-account logic (Standard, Express, Custom), HiPay requires fine-grained wallet and seller KYC management." And the legal frame is stated: "The buyer pays once, and funds are automatically redistributed to each seller, minus the platform's commissions. Split payment is governed by PSD2 regulation and requires a partnership with a licensed payment service provider." Onboarding takes 14 of 18 on a dedicated seller space, seller moderation and an unusually explicit compliance component, "KYC and compliance: seller identity verification, PSD2 compliance, management of documents required by PSPs".

It takes only 6 of 22 on named builds, for a reason worth weighing rather than dismissing. The Magento marketplace is Sevellia, which the page describes as "a marketplace for organic, natural and artisanal products, acquired and operated by Sutunam", so it is Sutunam's own business and not a client engagement, and the case-study link for it resolves to the portfolio index rather than to a case study. Its two client-named marketplaces, ActivCorner and LS Occasions, are Symfony and Sylius rather than Magento. Build or buy takes 6 of 12 on a published platform-choice rule, "Symfony and Sylius for complex custom architectures, Magento 2 when catalogue richness is at the core of the project", with no marketplace module named and no priced route. Platform depth takes 2 of 10: Magento 2 is named, and no Adobe tier, certification count, project count or Magento performance figure appears on the pages read.

ScienceSoft publishes the most complete functional specification for a Magento marketplace of any consultancy researched, and no marketplace it has built. Its Adobe Commerce as a marketplace platform page and its deeper marketplace software page together cover the seller side in full: "Manual and automated mechanisms for a seller approval", "a separate admin panel for each seller to manage product listings, payments, orders and sales analytics", "role-based access to a seller's profile", "restricted product categories, to protect consumers from fraudulent products and require from sellers licensing confirmation before giving them an approval to sell", and catalogue ingestion by "bulk uploading products via CSV/XLS/XML files" with sellers able "to create their own product attributes". That is a full 18. Order splitting takes the full 14 on "automated splitting of orders with products from several sellers", "seller-specific configuration of allowed shipping methods", "seller-specific flat rate shipping" and return merchandise authorisation.

Build or buy takes the full 12 and the pricing is the reason to read the page. It sets Adobe Commerce against Arcadier, Marketplacer, CS-Cart Multi-Vendor, Sharetribe and a custom build, with figures for each: Arcadier at "$68, $170, $409" a month, Sharetribe where "the budget typically starts at $5-10K, the build time is 1+ month", custom development where "commonly, the cost of online marketplace development lies in the range from $150,000 to $250,000" and "on average, it takes 10-12 months", and Adobe Commerce "from $250,000" for the mid-market. The decision criteria are written out rather than implied, including when custom is right: "if business requirements can be met only with unique and strictly business-specific architecture and functionality not offered by marketplace software providers".

Commission takes 14 of 24, and the shortfall is the same one Brainvire has. Granularity is excellent, "fixed and percentage-based commissions, global commission rates, tiered commissions based on a product category, a product price, seller-specific commissions", and the money model is described in two shapes, "a direct payment option, customers pay directly to a seller" against "an aggregated payment option, a marketplace collects payments, keeps the commission and pays the profit to sellers", but no payment provider is named anywhere and no payout schedule, wallet or escrow is described. Named builds take zero: its one Adobe Commerce marketplace case study is anonymous and, on reading it, single-vendor, a company selling its own machine-learning data sets, built on "Adobe Commerce Enterprise 2.4.3", with the word vendor absent from the build. The client names printed on its marketplace software page sit under the Adobe Commerce platform entry as that platform's clients rather than as ScienceSoft's. Platform depth takes 2 of 10, on version specificity alone; its Adobe provenance is stated as tenure, "since 2011, ScienceSoft has been working with Adobe Commerce", with no tier and no certification count. Its own sitemap returns HTTP 404 and its robots.txt names none, which is worth knowing if you expect an answer engine to find any of this.

Krish TechnoLabs has the joint-best named-build record on this page and the weakest published money mechanics of any entry with one. Its marketplace solutions page and Magento-specific marketplace accelerator sell a stated timetable, "launch a fully-functional online Marketplace in just 4 weeks", and three named marketplaces sit behind them, each with a JSON-LD modification date of 2023 or later. DuPont: "DuPont Sustainable Solutions seeks to create a multi-vendor marketplace for safety technology solutions", implemented "using Magento Commerce On-Premise, deployed on MS Azure", with a published page load time of 1.38 seconds. Tamata in Iraq, "defined to be the biggest B2C marketplace in the region", with the uplift figures printed in the evidence table below and a named client quote. And Leading Edge Group, where "a personalized catalog for each retailer was published, which depends on selecting specific brands, sellers, and categories" and the seller-buyer flow runs through its own Rocket Bazaar extension. That is 16 of 22, short only of a published seller count, and Tamata is described instead as "housed with 'n' number of vendors".

The money criterion takes 5 of 24, the lowest of any ranked entry. The accelerator page lists "Commission Management", "Auto Vendor Payout", "Financial Reporting" and "Tax Management" as feature tokens, and only the payout component survives a check, because no commission rate levels are published, no marketplace payment provider is named, no split at checkout is described and no reconciliation or settlement run is named. The only payment provider named on a marketplace build is Braintree, for subscription payments on DuPont, which is an ordinary gateway. There is one more thing a reader should know: the commission documentation this company does publish, with rates per vendor, category and product and a configurable precedence rule, sits on a separate product domain for its Rocket Bazaar extension, and `krishtechnolabs.com/rocket-bazaar/` is a soft 404 with no Rocket Bazaar path anywhere in its sitemaps. This page scores each company on its own primary domain, so that evidence is noted here and not counted.

Onboarding takes 11 of 18 on a vendor portal with bulk product upload, mass product action, vendor microsites and vendor reports, user role management, and the per-retailer catalogue work on Leading Edge Group, losing the components for a named approval mode and for listing moderation before publication. Order splitting takes the full 14 on "Split Order & Auto PO Generation", "Multivendor Checkout Option", "Real time Shipping Carriers" and "Partially Shipment Receivables" on the accelerator page, plus "Multi-Shipping" and "Return Management" on the marketplace solutions page. Build or buy takes 3 of 12: two stacks are named, its own accelerator and the Rocket Bazaar extension, but its page offers a labelled pair of approaches with no criteria, one timeline for one route and an explicit refusal on cost, "you can contact our experts for more information". Platform depth takes the full 10, the only entry other than scandiweb and Brainvire to do so, and it comes from two pages rather than one: its Adobe Commerce partner page publishes "a preferred Adobe specialized gold solution partner", "100 + Certifications" and "350 + Storefronts Launched", while the version specificity and the performance figure are on the DuPont case study, "using Magento Commerce On-Premise, deployed on MS Azure" and a 1.38 second average load time. In the Hyva full register it holds Silver.

CreativeMinds is a module vendor rather than an agency, and it is on this page because its documentation answers the money question better than most agencies do. Its multi-vendor marketplace extension for Magento 2 publishes commission at two levels, "charge a commission fee based on product category or set commissions for each individual vendor", and then does something almost nobody does: it publishes what the base product cannot do. "Is there an automated marketplace system for delivering payments to suppliers? Not in the standard plan. The extension includes an admin billing system that lets you track pending payments to vendors, but the actual payment is done manually based on the agreed fee." The paid PayPal vendor payments module closes it with three named payout timings, automatically after the order is placed, automatically after the order is completed, or manually for bookkeeping, with "admin fees deducted" and PayPal's Payouts API used to "make PayPal payments to multiple PayPal accounts in a single API call". That is a full 24, earned on precision rather than breadth.

Order splitting takes the full 14 and is the clearest worked example in the set. Its order split module publishes the rule, "Smart Split, only orders involving different vendors are split" and "Automatic Split, orders are split automatically after checkout", and then works an example through to the outcome. Per-vendor shipping is there, "each vendor can add multiple shipping methods" with flat, table or free rates and "a unique tracking code granted to each order", and returns run through a separate RMA module that will "allow admin and vendors to manage return requests for products" and lets an operator "create status such as possible fraud or broken good to streamline your RMA system". Onboarding takes 14 of 18 on manual or automatic vendor approval, product approval including for CSV imports, a vendor dashboard explicitly walled off from the admin account, and a documented SOAP API for product attributes; no identity or document check appears on the pages read.

Named builds take zero, on a distinction this page applies to every module vendor. Two live Magento multi-vendor marketplaces are named on its customer stories page, and in both cases the customer states they did the building: "I used CreativeMinds Marketplace Extension for Magento to build our multi vendor store corporategift.com." That is a licence reference, not a delivery record. Build or buy takes 3 of 12 on one component only, a published list of what the module does not do, including no auction feature and no seller vacation mode. Platform depth takes 2 of 10 on version coverage tested through Community and Enterprise 2.1 to 2.4; no Adobe tier, certification count, project count or performance figure appears on the pages read. It is priced at $299 and $499 with add-ons at $99, plus a $59 installation service, and its founder's profile says he "founded CreativeMinds in 2006".