What order management system (OMS) buyers wish they'd known

An order management system is kept for a decade, often by people who did not choose it. What buyers wish they'd known is what that costs: customization that decides the price of every later change, an upgrade with the same vendor that is still a migration, and a system judged in the end on whether its inventory is right. Seven lessons from the people who run one.

Based on verified interviews with the ecommerce, omnichannel, operations and technology leaders who run order management, at DTC and enterprise brands. Buyers are anonymized before publication; vendor names and views are reported as given. No vendor paid to appear or could edit this page.

OMS buyers wish they had known they were choosing a system for a decade, and that the customization that made it fit would set the cost of every change after. The established platforms in these interviews, IBM Sterling and Manhattan, are run for years and rarely replaced outright: unhappy buyers build around them or upgrade within the vendor. Several buyers choose newer systems such as Fluent Commerce, Deck Commerce, NewStore and OneStock for how they connect, and judge them on what teams can see and change once live. For what an OMS is and when a company needs one, see what is an order management system?

~6.5/10
IBM Sterling's buyer rating, the only dedicated OMS rated by enough buyers to publish a figure. Its users praise its scale and fault how hard it is to change.
6–15 yrs
How long IBM Sterling's users in these interviews say their company has run it. Several were not there when it was chosen.
Rewired
What an upgrade with the same vendor still means: a global apparel brand moving to Manhattan's cloud version is redoing every integration that touched the old OMS.

What do OMS buyers wish they'd known?

Seven things recur. The system outlasts the people who chose it. How changes get made, by the vendor, an integrator or your own developers, decides what the OMS costs over its life. Unhappy buyers rarely rip one out; they build around it. Moving to a vendor's newer version is a re-implementation. A suite's OMS is chosen for its fit with the rest of the suite, and that fit is where its complaints sit too. Newer systems are chosen for how they connect, and some users then struggle to see and change what is inside them. And several of the strongest and weakest accounts turn on inventory accuracy.

The seven lessons

01

You are choosing a system your successors will run

Buyers in these interviews describe some of the longest tenures of any system in their stack. IBM Sterling's users say their company has run it for six years, more than eight, about ten and at least fifteen; a multi-brand fashion-accessories company has run IBM's OMS for more than ten; another brand describes a long-standing relationship with Deck Commerce. Many of the people now rating these systems inherited them. Asked why Deck Commerce was chosen, that brand's digital director says “It was chosen before I started, so I don't know.” A global sportswear brand's analytics lead says of IBM Sterling that the choice “was done much before my time,” and an athletic apparel retailer's technology manager answers the same question with “I have no idea. I wasn't here then.”

The implication — ours, not a buyer's — is to write down why you chose the system, what you customized and what you deliberately left out, because the team that inherits it will be asking.

Named in this context: IBM Sterling · Deck Commerce

02

How changes get made shapes the OMS's long-term cost

The fashion-accessories company that runs IBM's OMS explains its choice, made more than a decade ago, as a trade between licence and change. As it saw the market then, Manhattan's licence was lower, but for changes “you'll have to go and knock on the door of the Manhattan team and they charge a bomb for them to implement any changes that you request,” while IBM's licence cost more and “you are given the base code.” That was the company's reading of both products when it chose, not a comparison of them today: Manhattan now describes its Active platform as configurable through low code, no code or the customer's own code (manh.com, checked November 2026). Our recommendation, not a buyer's: separate the changes your own team can make from those that need the vendor, a partner or a specialist, and price both over the system's life.

Owning the code has its own bill. An athletic apparel retailer calls its Sterling installation a dependable workhorse, but “You have to be a Sterling developer to be able to work on it,” so “we pay a premium when it comes to hiring consultants.” A footwear and accessories retailer that left Sterling for Fluent says its old system was run by an integrator and heavily customized before the buyer arrived, and “when you customize something and you don't use the out of the box functions, you create technical debt.” Smaller vendors raise the opposite problem. A consumer health company kept Tejas after reconsidering it, for low cost and flexibility, but finds “we end up paying for platform-level features that, in my opinion, sometimes should just be part of their roadmap to be done.”

Named in this context: IBM Sterling · Manhattan · Fluent Commerce · Tejas

03

Unhappy buyers build around the OMS before they replace it

The lowest IBM Sterling scores in these interviews come from companies still running it. A luxury retailer gives its custom-built version one of the lowest scores here and calls it “incredibly unstable,” and is keeping it anyway, because it is replatforming the website, stores and email around it: “we can't replatform the nucleus, as I just called it.” A global sportswear brand rates Sterling 5 because “It is a very difficult and complex system to drive any changes into”; its global team is weighing alternatives, with none named yet. A membership retailer has stopped investing in Sterling customizations and is building replacement pieces outside it. A large beauty retailer that gives Sterling a 6 was told order splitting “would take two years to implement,” and has moved functions it expected an OMS to handle into other systems, while saying “I still think it's a good platform.”

The same pattern appears beyond Sterling. A wine company found reservations in Salesforce Omnichannel Inventory that never released for orders that were never completed, and built its own app to clear them (lesson 05). These are workarounds, not departures, and they belong in the cost of staying. Not every low score is about the OMS itself: a home-goods retailer rates Sterling 5 and says “A lot of it has to do with limitations of other internal systems it talks to, not necessarily the tool itself.”

Named in this context: IBM Sterling · Salesforce Omnichannel Inventory

04

An upgrade with the same vendor is still a migration

Many of the Manhattan stories here are moves from an older Manhattan system to Manhattan Active Omni, its cloud version, rather than moves to or from another vendor. They are big projects. A global apparel brand moving from a version it hosted itself to the vendor-hosted one says that in the integration work “anything that touched the OMS previously has to be redone,” naming loyalty and customer-service connections among them. In an interview the following January, a colleague at the same brand described the new version as live only in Canada, with the US and then Europe to follow, and “a couple more years” before the move is complete.

Those further along describe the gains. The same colleague credits event-based messaging that is “a lot easier to manage, a lot quicker,” and an apparel group that moved when its on-premises system reached end of life five years ago rates the cloud version 8 for orchestration and routing. An athletic footwear retailer that upgraded last year rates it 7: it has unlocked new capabilities, but “everything needs to be developed and engineered a little more intensely than we'd like.” Manhattan describes its Active platform as version-less, with feature releases every 90 days and a rolling update strategy that gives customers flexibility in which updates to implement (manh.com, checked November 2026). That is the vendor's product model; the buyers here describe the effort of moving onto it, not yet what running it costs across many releases.

Named in this context: Manhattan Active Omni · Manhattan

05

A suite's OMS is bought for the suite, and that is where its complaints sit too

Buyers of Salesforce Order Management describe choosing it for how it fits the rest of Salesforce. A specialty health company chose it for “The fact that it connected easily into Commerce Cloud and the cross-cloud compatibility,” rates it 9, and still wishes the cross-cloud setup were simpler: “some of that should be a little bit more innate when you're using a cross-cloud environment.” For a wine company it arrived with another purchase: “As we were upgrading to Service Cloud, it came with the product.” The company rates the OMS 8 and Salesforce Omnichannel Inventory beside it 5, calling its reservation process “still deeply flawed” because reservations from orders that were never completed would not release. Salesforce's developer documentation does describe methods to release, cancel and transfer inventory reservations (checked November 2026), so this describes the company's setup rather than a missing feature.

A buyer at a consumer goods manufacturer, comparing Salesforce's OMS with SAP's in 2024, warned from a previous employer that Salesforce's clouds do not simply plug together: “You would think that they would work very well together. They do not.” IBM's buyers describe a suite too. The fashion-accessories company bought IBM's middleware because it was built to connect to the IBM OMS, and runs IBM's call-center module bundled with it. A big-box specialty retailer moving onto Manhattan's OMS, where it already ran Manhattan's warehouse system, put the other side of the trade: buying a whole ecosystem “can lead to pretty sticky situations when you're relying holistically on one vendor.”

Named in this context: Salesforce Order Management · Salesforce Omnichannel Inventory · IBM Sterling · Manhattan

That's the average buyer. How does your stack measure up?A free, peer-powered read of your tech stack, with a shortlist and your next steps.
Get my stack read →
06

Newer systems are chosen for how they connect; some users then struggle with visibility and control

Several buyers of cloud-first systems name connectivity as the reason they chose one. An apparel brand building a composable site says of Deck Commerce: “It works with our composable architecture, which is the main strength, so we connect with the API.” A luxury retail group chose Fluent for scalability and says “It's easy to integrate with.” A footwear brand chose NewStore for its integration with commercetools, and a multi-country retailer that has OneStock live on some of its sites praises its orchestration rules: “From a customizability perspective, it's very flexible.” Fluent describes integrations it builds and maintains for customers, which it calls Fluent Connect (fluentcommerce.com, checked November 2026).

The lower scores here concern what people can see and change once the system is live, and they differ by product and setup. The footwear brand rates NewStore 6 because it is “difficult to export data” and “sometimes orders can get stuck,” leaving staff to push them through again by hand. A luxury fashion house rates Fluent 6 because “We need to check order by order to find the information,” and adds “but maybe that's because I didn't have enough training.” A sportswear brand's Latin American business, newly on Fluent, rates it 6 for configuration it cannot change itself, “For our local configuration, it's quite stiff,” and blames most of its problems on a new ERP and invoicing integration, “not directly because of them.” Fluent's individual ratings here run from 6 to 8.

Named in this context: Fluent Commerce · Deck Commerce · NewStore · OneStock

07

The OMS is judged on whether the inventory is right

Several of the highest and lowest scores here come down to inventory: which stock can actually fill an order, and whether the OMS agrees with the systems around it. A large footwear retailer gives Manhattan DOM a perfect score after five months, having moved to it “To get better accuracy of inventory records and have better protections against false availability”; its main strength is that “it does not let you duplicate demand for a product.” The global sportswear brand traces its cancellations to inventory moving between IBM Sterling, SAP and its warehouse system without syncing in real time: “Sometimes we get overstock or understock, and it leads to cancellations.” The wine company's unreleased reservations held back stock it actually had.

Integration is one place accuracy breaks, and the fault can sit in the OMS, the systems around it or how they are connected. In a 2023 interview, a footwear chain described pickup orders going to the wrong store, which took weeks to trace to “an API from Manhattan that was timing out” combined with a fallback condition in its own website code. That describes one company's setup at the time, not a fault found in the product. The footwear retailer that gives Manhattan DOM a perfect score still names “the difficulty in integrating the data that we use” as its biggest challenge.

Named in this context: Manhattan · IBM Sterling · Salesforce Omnichannel Inventory

Which OMS platforms do buyers run and rate?

The dedicated systems named most often are IBM Sterling and Manhattan, followed by Fluent Commerce and Salesforce Order Management, with NewStore, Deck Commerce, OneStock, Radial and Tejas among the others. Many companies instead run order management in their commerce platform, an ERP such as NetSuite or Oracle, or their own code, which is the subject of what is an order management system? Only IBM Sterling has enough ratings to publish a figure.

Buyer ratings for dedicated order management systems, Alium verified interviews through November 2026. Systems rated by fewer than eight buyers are marked not rated; between eight and about twenty-five, figures round to the nearest half.
System Rating What buyers describe
IBM Sterling around 6.5 Praised as scalable and proven at high volume; faulted for how hard and slow changes are, unstable custom builds, and the scarce skills it needs. Some buyers still call it Sterling Commerce, the company IBM acquired in 2010.
Manhattan Active Omni not rated Too few buyers rated it to publish a number. Praised for omnichannel orchestration and inventory accuracy; companies on Manhattan's older system describe the move to it as a migration. Some buyers call it expensive.
Fluent Commerce not rated Too few buyers rated it to publish a number. Chosen for scalability, ease of integration and lower cost of ownership; lower raters cite order lookup and configuration they cannot change themselves.
Salesforce Order Management not rated Too few buyers rated it to publish a number. Chosen to fit Commerce Cloud and Service Cloud, and praised for showing order status across them; complaints concern cross-cloud setup and the Omnichannel Inventory product beside it.
NewStore not rated Too few buyers rated it to publish a number. Praised for bringing store and online orders together and for its store app; one buyer reports stuck orders and slow data exports.
Deck Commerce not rated Too few buyers rated it to publish a number. Described as simple, modular and API-first, a fit for composable sites; one buyer describes a long-standing relationship with it.

On shortlists, Manhattan and IBM Sterling are the default pair. A specialty retailer weighing the two names its criteria as the “Ability to integrate across e-com, Warehouse Management, and our distributed store network, total cost to serve, data accuracy.” A beverage company built its shortlist from “online research, analyst reports.” Challengers appear beside them: a pet specialty retailer's team was looking at Fluent, Kibo and Nextuple ahead of a planned RFP, and a regional retailer replacing a partly homegrown system named Kibo its top contender for ease of use and reporting, with contracts not yet signed. The footwear and accessories retailer that chose Fluent over Manhattan and Sterling did so because its own data and engineering teams could build what it would otherwise have paid either vendor for.

How OMS moves break down

Completed moves between two different named OMS vendors are rare in these interviews. What these interviews show instead is companies leaving an older or homegrown system, moving to a newer product from the same vendor, moving some functions into other systems, or still deciding. The table keeps them apart by each company's latest reported status; it lists individual company accounts, not a churn rate or a market share.

OMS moves described in Alium verified interviews through November 2026, by status. Each row is one or more individual company accounts.
Status Move What buyers describe
Completed IBM Sterling to Fluent Commerce, one brand first A footwear and accessories retailer moved one brand like-for-like before building anything new, citing technical debt and cost of ownership.
Completed Older Manhattan system to Manhattan Active Omni An apparel group at its old system's end of life; an athletic footwear retailer, described as an upgrade under way in one interview and finished by the next.
Completed Earlier system to Fluent, Deck Commerce or Manhattan DOM A global sportswear brand and a global beauty company to Fluent; an apparel brand to Deck Commerce; a large footwear retailer from an older system to Manhattan DOM.
Completed, vendor-forced OrderWave to Salesforce Order Management A consumer brand moved because its old supplier was going out of business, not because of a competitive choice.
Partly live Older Manhattan system to Manhattan Active Omni, country by country A global apparel brand, live in Canada with the US and then Europe to follow; not yet a company-wide move.
Under way, in parts Selected IBM Sterling functions to the company's own software A membership retailer building replacements for Sterling functions outside it; Sterling still runs its post-purchase fulfillment.
Planned Aptos to Manhattan A big-box specialty retailer doing integration work for a launch planned for the following year; no later interview confirms it.
Selected, not live Fluent Commerce; Salesforce Order Management A sporting goods retailer whose group had chosen Fluent, with go-live planned; a sporting goods brand adopting Salesforce's OMS as its first.
Evaluating Manhattan vs. IBM Sterling; packaged vs. custom A specialty retailer shortlisting the two; a consumer products brand weighing Salesforce's OMS against building its own; a global sportswear brand reviewing alternatives to Sterling.
Reconsidered and kept Tejas; IBM Sterling A consumer health company stayed with Tejas on cost and flexibility; a luxury retailer is keeping Sterling while it replaces almost everything else.

One evaluation shows how slowly these decisions can move. A pet specialty retailer whose team was looking at replacement systems in early 2025 was still on Oracle's order management when interviewed again in October 2026, rating it 5 for downtime and rigidity and saying the OMS “might need a reconsideration.”

Want your own stack read the same way?

Get mine →

The stories behind the lessons

A multi-brand fashion-accessories company has run IBM's OMS for more than ten years and rates it 8. Its operations lead describes the sourcing logic as the core: for an order with items spread across a nearby store, a distant one and another country, “It considers all these factors and understands the right fulfillment location from where the order should be fulfilled.” The company chose IBM's OMS because it received the base code and could build on it, and it runs IBM's middleware and call-center module beside the OMS. The same reasoning carried into its AI tooling, where it chose IBM because “we are already leveraging the IBM OMS platform.” Senior operations manager · fashion-accessories company
A global apparel brand is part-way through moving from an older Manhattan version to the new one, and its DTC technology manager rates it 9 on what the team has seen so far. Inventory used to move as “a lot of big files that were sent back and forth”; now it is event-based messaging between Manhattan and the other systems. Manhattan manages the full order lifecycle, store inventory allocated to online orders, pickup in store, and returns. The rollout is phased: “That's currently live only in Canada. The next step is the US, and then we move to Europe.” The buyer says Manhattan was the clear winner of an RFP two or three years earlier. Senior manager, DTC technology · global apparel brand
A consumer health company reconsidered its OMS, Tejas, and kept it. The product director, who was not there when it was first chosen, gives the reason: “the cost was fairly low compared to other Order Management Systems. They have been really flexible and easy to work with.” The same openness is the drawback. The vendor is “not that opinionated on how Order Management works,” and lacks “a really solid prioritized roadmap,” so the company pays for features it thinks should come as standard. The company has a dedicated order management product manager who works across its direct-to-consumer site, direct fulfillment and TikTok Shop. Director of product, DTC · consumer health company

The counter-current: a long-lived OMS is not a failed one

For all the complaints above, several buyers describe old systems they are content to keep. The athletic apparel retailer calls its Sterling installation “a bulletproof, proven, high-scale, high-volume solution. So it's a workhorse.” A global consumer brand's partner-channel lead, on IBM's OMS for more than eight years, values “a robust backbone” for peak periods. A specialty retailer rates Sterling 8 after six years running both ecommerce and store orders on it, and a long-standing Deck Commerce customer calls it “very simple, easy to use.” Our reading, not a buyer's: tenure alone is not a reason to replace an OMS. The reasons buyers give for leaving are the cost and speed of change, instability and inventory errors.

What this means for your OMS decision

These recommendations are ours, drawn from what the buyers above describe, not a process any of them laid out. Before shortlisting, decide who will make changes after go-live, whether that is the vendor, an integrator or your own developers, and ask each vendor how a typical change request is scoped and priced. Price three things separately: the licence, the implementation, and the ongoing cost of change and support over several years, including scarce specialist skills. Run the evaluation on your hardest real orders through the integrations you will actually use: a split shipment from a store and a warehouse, a pickup reservation the customer abandons, a marketplace order, a return to a store. Test inventory sync with your ERP and warehouse system at peak volume, and have the people who look up orders, such as customer service and finance, try to find what they need. If the OMS comes with a suite, test the flows across its products rather than assuming them. If you are moving to a newer product from the same vendor, list every interface that touches the current OMS, plan the rollout by brand or country, and budget it as a migration until the vendor shows otherwise.

For the ERP side of the same decision, see what is an ERP?; for building rather than buying, build or buy; for the commerce platform the OMS sits behind, what buyers wish they'd known before replatforming and how long ecommerce replatforming takes; and for the returns side of the same flows, what returns and post-purchase buyers wish they'd known.

What these interviews do not show

The ratings measure how satisfied each system's own users are, many of whom inherited it, not a head-to-head test, and only one system has enough ratings for a figure. The individual scores quoted for other systems are single companies' views. The moves table is a set of company accounts, not a churn rate, a win rate or a market share, and a company that moved from an unnamed earlier system cannot be counted as leaving any particular vendor. Where a company was interviewed twice, the change between interviews shows the interval between them, not how long a project took. Complaints about a system often describe that company's customizations and integrations, as several buyers say themselves. Vendor descriptions come from vendor sites and can change.

Common questions

Which OMS platforms do retailers use?

The dedicated order management systems named most in Alium's verified interviews are IBM Sterling Order Management and Manhattan, followed by Fluent Commerce and Salesforce Order Management, with NewStore, Deck Commerce, OneStock, Radial and Tejas among the others. Kibo and Nextuple appear on shortlists. Many companies instead run order management inside their commerce platform or ERP, such as NetSuite or Oracle, or in software they built themselves.

Which order management system do buyers rate highest?

Only IBM Sterling is rated by enough buyers to publish a figure: around 6.5 out of 10, below what buyers give software generally. Manhattan, Fluent Commerce, Salesforce Order Management, NewStore and Deck Commerce are each rated by too few buyers for a number, and their individual scores range widely. A rating reflects how satisfied each system's own users are, many of whom inherited it, not a head-to-head test.

What is the difference between an OMS and an ERP?

The ERP books orders in finance and inventory; the OMS coordinates each order across channels, stock locations and fulfillment options. Some companies in these interviews run order management in the ERP itself, such as NetSuite, Oracle or Microsoft Dynamics 365, and value the simpler financial reconciliation. Buyers with dedicated systems describe the ERP connection as one of the hardest integrations, and inventory that does not sync between the OMS, ERP and warehouse as a cause of cancellations.

Is upgrading to a newer version of the same OMS easier than switching vendors?

It avoids a new vendor, not a migration. A global apparel brand moving from Manhattan's older order management to its cloud version, Manhattan Active Omni, describes redoing every integration that touched the old system and rolling out one country at a time. Buyers further along credit the new version with more flexibility and lighter integrations, and one says changes still take more engineering than it would like.

Should you build your own OMS or buy one?

Both routes appear in these interviews, and each costs something later. Companies that built their own describe control and a close fit, and also changes that wait on their own engineers and systems few people understand. Buyers of packaged systems describe the same trade inside the product: the more they customized it, the harder and dearer every later change became. Some companies are building around or replacing a packaged OMS with their own code, while others are leaving homegrown systems for a packaged one.

What should a retail OMS evaluation test?

The criteria buyers name are integration with ecommerce, warehouse management and the store network, total cost to serve, and data accuracy. Our recommendation, not a practice buyers describe, is to test the hardest orders you actually run, such as a split shipment from a store and a warehouse or an abandoned pickup reservation, through the integrations you will use, and to price a typical change request as well as the licence and the implementation.

Your stack, read by the buyers who've run it

Peer-powered: built from hundreds of verified buyer interviews, and fitted to the tools you already run. You get a shortlist for your stack and your next steps.

Get my stack read

Free · no password · your answers are anonymized before they inform research like this.

Want to come back to this?We'll email you a link to this page, so it's easy to find later or forward to your team.
Methodology. Alium conducts verified interviews with software buyers — the ecommerce, omnichannel, operations and technology leaders who select and operate these systems. This page draws on every published interview through November 2026 in which a buyer names the system that runs their order management, describes how it was chosen, or discusses changing it, read in the interviewee's own words. Ratings are averages across verified buyer interviews on published transcripts; IBM Sterling combines the product records the corpus keeps for it under its current and former names, a system rated by fewer than eight buyers is named without a figure, and one rated by fewer than about twenty-five is rounded rather than given a decimal. Moves between systems are read one company at a time and kept apart by status: completed, under way, selected but not live, being evaluated, or reconsidered and kept. Where a company was interviewed more than once, the change between interviews is reported as an interval, not a measured duration. Vendor descriptions come from each vendor's own site and documentation as of November 2026. Buyer identities are verified at interview time and anonymized before publication; vendor names and ratings are reported as given. No vendor paid to appear or was able to edit this page.