A $50M gross merchandise volume (GMV) business running commercetools spends between $560,000 and $1.08M a year to operate it, against a license of $150,000 to $220,000. That gap is where backend evaluations start, and it is a staffing number before it is a software one. Whoever replaces commercetools inherits the same question in a different shape: who on your team owns search, content, payments, and extension logic once the contract is signed. This guide covers the six criteria that answer it, the vendors each criterion favors, and the two commercetools releases that changed the question in mid-2026.
Key takeaways:
Choosing between commercetools and an alternative is a staffing decision, because the license is the smallest line in a composable build's cost.
Two commercetools releases in mid-2026 made buying less of the platform a real option, and it now competes with every vendor on this list.
Alternatives fall into composable backends, all-in-one suites, and open-source backends, and each category moves the ownership line somewhere different.
Migration effort concentrates in logic, not data, because API Extensions and Custom Types have no export format and get rebuilt by hand.
A decoupled storefront in front of the existing backend returns performance sooner than a backend swap and shrinks the blast radius of a later one.
Copy link to headingWhy teams look for commercetools alternatives
Four patterns show up repeatedly in evaluations, and departures rarely trace to a single capability gap.
Copy link to headingImplementation load arrives before the platform team does
commercetools documents longer time to market than all-in-one platforms as a property of the model, and its own answer is to pair the APIs with its other products. That answer works for teams with a platform group. Mid-market teams without one discover the gap after the contract, when the first integration needs an owner and nobody has one.
Copy link to headingMerchant Center still routes routine work through developers
Merchandisers hit this before engineers do. Product search exists, but it requires indexing to be turned on for the Project before anyone can search a catalog that is already loaded. Custom objects remain developer territory. Cart-level permissions have been filling in through 2026. The July release notes carried cart editing to general availability alongside an override for line item price, and a separate delete permission landed in August. That gap is narrowing without being closed.
Copy link to headingThe license is the smallest line in the cost
For a $50M GMV business, annual cost of ownership reaches $560,000 to $1.08M once four to six engineers, infrastructure, integration upkeep, and third-party services are counted. License fees account for $150,000 to $220,000 of that, which lines up with the $151,200 median contract recorded across 2025. Everything above the license is roughly three quarters of the bill, and in the $10M to $100M GMV band that three quarters is what stops adding up.
Copy link to headingThree chief executives in under 16 months belongs in the procurement file
After the October 2025 leadership change, Doug McNary became the third chief executive in fewer than 16 months, following layoffs in February 2025 that the company attributed to missed revenue growth targets. Vendor stability belongs alongside roadmap and support terms as a procurement input, and it cuts both ways. commercetools has since reported a $100B GMV run-rate in the first quarter of 2026. Knowing why teams leave still does not settle which alternative fits, and in 2026 it no longer settles whether leaving is the right move.
Copy link to headingWhat commercetools shipped in 2026 changes the alternatives question
Most alternatives research assumes commercetools is an all-or-nothing composable commitment. Two releases in mid-2026 made that assumption unsafe, and a third option now sits alongside stay and leave.
Copy link to headingCore Commerce and Product Catalog now sell separately
In June the company introduced Sphere, an API-native platform that is now the umbrella its other products sit on. A month later it packaged that platform into two offerings sold separately. Core Commerce covers cart, order management, checkout, and customer management. Product Catalog covers product modeling, pricing, inventory, and search. Both are pitched as modernizing without a full replatform, which is the objection an alternatives evaluation usually raises first.
Copy link to headingThe catalog model is mid-migration
Catalog limits moved in parallel. A product caps at 100 variants under the Classic model, and the Modular catalog model, in public beta since June 2026, raises that ceiling to 10,000. Pricing there runs exclusively through Standalone Prices, so a team carrying Embedded Prices rewrites its pricing integration to make the move. A staged migration path between the two models entered public beta in September, letting a Project change catalog models incrementally instead of in one cut.
Buying Core Commerce or Product Catalog on its own, instead of the whole platform, narrows the contract without a migration, and that option competes directly with every vendor below. Teams whose pain is catalog ceilings or checkout flexibility should price that path first. Where the pain is the operating cost of running a distributed system, modular packaging changes nothing, because the integration surface stays exactly where it was.
Copy link to headingSix dimensions to price every commercetools alternative against
Most of these surface after the contract is signed, which is the argument for pricing all six during the evaluation.
Copy link to headingCommerce feature depth out of the box
commercetools provides commerce and search APIs; teams still implement the storefront and connect the content, payment, and other services their architecture requires. The suites invert that. Shopify Plus ships checkout and Shop Pay. Agentforce Commerce, the name Salesforce Commerce Cloud is now marketed under, ships AI merchandising, order management, and payments. Adobe Commerce ships seven native workflows for business-to-business (B2B) selling, including negotiable quotes. Built-in capabilities can reduce custom development and maintenance.
Copy link to headingAPI model and extensibility
commercetools covers both REST and GraphQL, with synchronous API Extensions across 11 resource types and a ceiling of 25 per Project. Saleor is GraphQL only, while Medusa v2 is REST and routes custom logic through Workflows. Shopify runs custom logic in Functions, capped at 11 million instructions and 20 kB of output for carts up to 200 line items, with the allowance rising as carts grow past that. Each of these caps is a statement about where your business logic is allowed to live.
Copy link to headingTotal cost of ownership
License price is the number every vendor publishes and the wrong number to compare on. The commercetools figures above put licensing at roughly a quarter of the annual total, and the other three quarters are engineers, infrastructure, integration upkeep, and third-party services. Every alternative should be priced against those same four lines. A suite absorbs most of the integration upkeep and shrinks the engineer count. An open-source backend removes the license and converts it straight into engineering time. A composable peer barely moves the number at all.
Copy link to headingTime to first launch
The published figures here are vendor-supplied, and they still beat an estimate because each one names a scope. A Forrester composite Salesforce customer implemented a composable storefront in nine months, and J.Lindeberg reached Shopify in 16 weeks. Those are different jobs, a storefront build against a full platform move, so treat them as endpoints and not as a range. Accelerators claim 30% to 50% off implementation timelines, which is a discount on a number you should still budget in full.
Copy link to headingCatalog and multi-market ceilings
Ceilings are the cheapest thing to check and the most expensive thing to discover late. Shopify raised its variant limit to 2,048 per product in October 2025. Its B2B catalogs run on two separate axes, with three catalogs active at once below Plus and up to 25 assignable to any one company location. Salesforce enforces 100 sites per instance. A catalog carrying 150 or more variants per product, or a storefront count approaching that Salesforce ceiling, needs modeling against these numbers before anyone signs.
Copy link to headingMigration cost off commercetools
commercetools publishes an Import API and no export counterpart, so resources leave through the Merchant Center or through HTTP API queries you write yourself. Customer passwords typically cannot transfer at all, which means every account holder resets a password on launch day and the launch plan has to absorb the support volume.
Copy link to headingFive commercetools alternatives compared
The five below span the three categories teams shortlist. Four of the six dimensions fit into columns, while launch time and migration cost stay in the notes because they are per-team numbers.
None of these is a drop-in replacement, and the notes below cover what the table cannot.
Copy link to headingShopify Plus
Shopify Plus reaches launch fastest of the suites and absorbs the most operational surface, which is the whole argument for it. Checkout customization is the part teams check first, and checkout UI extensions for the information, shipping, and payment steps are available only on a Plus plan. The cost is that custom logic lives inside Functions limits instead of in your own services, and the two platforms divide control and speed along exactly that line.
Copy link to headingAgentforce Commerce
Existing Salesforce customers face a forced API migration whether or not they replatform. Salesforce deprecated the Open Commerce API in April 2026 and now routes all new work through the Salesforce Commerce API, with security updates on the old one for roughly two more years. That deadline is why teams planning their exit tend to run both evaluations together. Its B2C Commerce release line keeps the older name, so both appear across Salesforce's own documentation.
Copy link to headingElastic Path
Elastic Path sits closest to commercetools architecturally. Its November 2025 expansion added native search, a content management system, and a hosted frontend, absorbing three integrations a composable build usually owns. Composer, its integration platform, takes the connector work behind them. The API is REST-first, with GraphQL available only as a self-hosted open-source wrapper, so a frontend team standardized on GraphQL inherits that wrapper as something to run.
Copy link to headingMedusa v2
Medusa reaches a first deploy faster than anything else here for a small team, and Partbase launched on it in under five months with two technical founders. The MIT-licensed core carries no license fee, and a Medusa storefront starter deploys to Vercel from the template gallery. B2B is custom work, so the licensing saving converts into engineering time the moment requirements go past business-to-consumer (B2C) selling.
Copy link to headingSaleor
Saleor is GraphQL end to end, with more than 180 webhook events whose payloads are shaped by subscription queries. Integration teams get precise control over what each consumer receives. The BSD-3 license permits self-hosting, and Saleor Cloud starts at $1,599 a month. Native B2B is thin against the suites, so the same caveat as Medusa applies.
Four more options sit outside the table and still belong on a shortlist. Spryker ships the deepest native B2B of the composable peers as a platform-as-a-service, with recurring orders reaching general availability in release 202608.0. It suits teams whose requirement is procurement workflow more than catalog scale. Emporix and Kibo appear on the MACH Alliance certified ISV list, as does commercetools, and both pair their APIs with process depth a generalist backend leaves to you. That listing moves without announcement, so check current status for any vendor whose architecture standards name the certification. BigCommerce sits among the all-in-one suites and lands closest to Shopify Plus on operational surface, so teams weighing the two are choosing between checkout models and app ecosystems.
Open-source backends carry a different risk profile. Vendure moved to GPLv3 in 2024 with a commercial license alongside it, and Mage-OS forks Magento Open Source without the Adobe B2B module. A fork inherits the upstream security surface and the upstream patch lag, and both land on your on-call rotation. Mage-OS shipped an emergency release in September 2026 for an upstream flaw rated 10.0, which is the cost of that inheritance made visible.
Start the shortlist with the requirement that is hardest to satisfy. Where one of these is the binding constraint, the shortlist narrows on its own:
Merchant tooling and launch date: A suite absorbs the most operational surface, which points at Shopify Plus or Agentforce Commerce before any composable peer.
Catalog ceilings or market structure: Check published limits first, because a suite ceiling can bind harder than the one you are leaving and no migration loosens it.
License cost with engineering capacity to spare: An open-source backend removes the fee outright, and Medusa reaches a first deploy fastest of the two here.
Extensibility past what a suite allows: A composable peer keeps the architecture and changes only the vendor, which is the narrowest move on this list.
Whichever criterion binds hardest should decide, since the others are usually negotiable once the first one is satisfied.
Copy link to headingWhat leaving commercetools costs in engineering time
Order history is the cheap part, because only orders still inside the return window need to move at all. Everything encoding a business rule is the opposite. Every Subscription consumer, covering order management sync, inventory updates, and email triggers, gets rewired to the destination's event system. Each of up to 25 API Extensions and each Custom Type's field definitions gets rebuilt by hand, with no automated tooling to carry them across. That work scales with how much logic you pushed into the platform, which is why the teams who customized most pay most to leave.
Sequencing decides the risk. commercetools documents a strangler pattern that runs low-hanging fruit first, then a deep dive, then harvest, keeping both systems live through the transition. A mid-market B2B move to Shopify takes 12 to 20 weeks by one agency's delivery data, depending on catalog complexity and integration count. Across that window the routing layer splits traffic and owns every request including checkout, so it needs the monitoring, rollback, and on-call coverage of a production system. That makes the replatforming sequence the thing to plan around, not the launch date.
Copy link to headingReplacing commercetools will not make your storefront faster
Backend choice and storefront performance are separate problems, and teams conflate them expensively. The common failure is approving a replatform to fix Core Web Vitals, then discovering months later that the storefront renders the same way against a different API. Caching is what moves the number. Vercel's engineering team documented caching commerce API responses alongside Incremental Static Regeneration in front of Salesforce Commerce Cloud. A global sportswear brand cut time to first byte from 2.0s to 0.8s there without touching the backend. The same pattern applies against commercetools, Shopify, or Medusa, because it works on the rate limits every commerce API enforces and not on anything specific to one vendor. Partial Prerendering serves the static shell from cache and streams the cart from whichever commerce API sits behind it.
Copy link to headingHow Vercel helps teams evaluating commercetools alternatives
The storefront is the tier that stays yours whichever backend wins, which makes it the cheapest place to get performance and the safest place to absorb migration risk.
Copy link to headingCommerce API rate limits throttle the storefront before the backend does
Traffic spikes expose the API budget first. A category page that fans out to product, pricing, and inventory calls multiplies every visitor into several backend requests, and the rate limit binds long before the backend runs out of capacity. Caching those responses with ISR collapses repeat requests to a cached read. That absorption is what let Vercel's content delivery network (CDN) take 518,027 requests per second at peak over Black Friday Cyber Monday 2025, with the backends behind it never seeing that volume. Cache tags then invalidate only the affected category and product pages when inventory moves, so restocking never requires a full deployment.
Copy link to headingA big-bang cutover puts every risk on one launch date
Replatforms concentrate risk because the traffic switch is a single event, and rolling back means restoring a backend that has been taking writes. Routing Middleware moves one path at a time with the legacy backend still live. Category pages can run against the new backend while checkout stays where it is, so a failure takes out one route instead of the storefront.
Copy link to headingThe backend choice should not force a frontend rewrite
Storefronts coupled to a backend SDK make the next backend decision more expensive than it needs to be, and that is how teams end up renewing a contract they wanted to leave. The Next.js Commerce template isolates the integration behind a lib/ directory, so swapping the commerce layer leaves the rest of the template mostly unchanged. That seam keeps the storefront decision reversible while the backend one is not.
Copy link to headingFrontend-first works on suites as well as composable backends
Teams on a suite often assume the performance ceiling is the platform's. Helly Hansen started its move to Next.js at checkout instead of the homepage. Core Web Vitals went from all red to all green within five months, and the brand posted 80% year-over-year Black Friday growth. What transfers is the sequencing, since proving the pattern on the highest-value path makes the rest of the storefront an easier argument internally.
Copy link to headingMove the frontend while the backend decision is open
Teams get this wrong by treating one decision as two, or two as one. Your backend is a durable, expensive, staffing-shaped commitment, and in 2026 it includes the option of buying less of the platform you already run. A storefront moves in weeks, reverses cheaply, and carries most of the performance the business is asking for. Running the storefront move first buys the months you need to make the backend call on evidence instead of on a renewal date. Vercel covers the storefront side of that split:
Routing Middleware: Shifts production traffic one path at a time, so a replatform runs as a sequence of reversible steps with the legacy backend still serving.
ISR and cache tags: Caches commerce API responses at the CDN and invalidates them per category or product, keeping a rate-limited backend off the critical path at peak.
Partial Prerendering: Serves the static shell from cache while personalized data streams in, so backend latency stops setting the floor for every page.
Next.js Commerce: Isolates the commerce integration behind a
lib/seam, keeping the storefront portable across whichever backend the evaluation picks.Vercel CDN: Absorbs repeat traffic across 126 points of presence in 51 countries, so peak load lands on cache instead of on a commerce API with a rate limit.
Start a new project to put a decoupled storefront in front of the backend you already run, or browse the commerce templates to see how each integration is structured before committing to one.
Copy link to headingFrequently asked questions about commercetools alternatives
Copy link to headingDoes commercetools charge based on GMV?
No. commercetools prices on order volume, and its pricing page states that the order-based approach is unlike GMV-based models. SKU count, API usage, and project complexity adjust the quote, so two businesses at the same order volume can land on different numbers.
Copy link to headingWhat is the difference between headless and composable commerce?
Headless separates the storefront from a backend that still owns every commerce function itself. Composable replaces that backend with independent services you select and integrate. commercetools is composable, while a Shopify Plus store with a custom frontend is headless without being composable.
Copy link to headingDoes Vercel replace commercetools?
No. Vercel runs the storefront layer in front of whichever commerce backend you pick, with no cart, catalog, or order management of its own. That separation is what lets the frontend move and the backend decision run on different timelines.
Copy link to headingWhich commercetools alternatives have an official Vercel integration?
The Vercel Marketplace lists BigCommerce, Saleor, Salesforce Commerce Cloud, Shopify, Sitecore OrderCloud, Swell, and Wix. commercetools has no listing. The vercel/commerce repository carries community-maintained integrations for Medusa, Shopware, Umbraco, and others.