What does a mid-market B2B ecommerce project actually cost?

Almost every B2B commerce budget that goes wrong goes wrong the same way. Somebody prices the platform licence, adds a build number from a proposal, and treats the rest as noise. The rest is usually the majority of the money.

There are five separate lines in a mid-market B2B commerce project. They are paid to different people, they arrive at different times, and only the first one has a price you can look up. Two of them, data preparation and internal time, are usually paid in your own payroll rather than in an invoice, which is exactly why they get left out of the business case and then blow up the schedule.

Read the table as a checklist for any proposal you are handed. If a quote covers line two and is silent on lines three, four and five, it is not wrong, it is incomplete, and you are the one carrying the difference.

Every figure in this section was read from the vendor’s own pricing page on 20 August 2026. Treat the date as part of the fact, because all four vendors revise these.

Shopify lists Basic at $39 USD a month billed monthly or $29 billed annually, Grow at $105 or $79, and Advanced at $399 or $299. Shopify Plus is listed separately as "Starting at $2,500 USD/month" on a one-year term or "$2,300 USD/month" on a three-year term, with additional stores beyond the included allowance at "an extra $300 USD per month per store, or a revenue share across all stores". Shopify also states that "More complex, higher volume businesses switch to a variable platform fee based on their revenue and business model", which means the listed floor is a floor and not a ceiling.

BigCommerce lists Core at $39 USD a month monthly or $29 annually, Growth at $105 or $79, Scale at $399 or $299, and Performance starting at $1,499 USD a month billed annually. BigCommerce ties the plans to trailing twelve-month gross merchandise value and upgrades you automatically when you cross a threshold, so the licence is a function of sales rather than a fixed number. BigCommerce does not publish a price for B2B Edition on its plans page, so if B2B Edition is on your shortlist the licence line is a quote, not a lookup.

Shopware publishes a free Community Edition and lists Rise from €600 a month, Evolve from €2,400 a month and Beyond as custom, all net of VAT, and states that "The pricing for our plans is based on your gross merchandise value (GMV) and further individual factors". Drupal Commerce carries no licence fee at all, which moves the entire cost of that option into lines two through five rather than removing it.

Acumatica does not price by seat. Acumatica states "Pay only for the functionality you need", and says explicitly that you are not paying for user seats. Acumatica lists the licence as a function of the applications you implement, the transaction volume and data storage you consume, and your deployment choice. The practical consequence for a commerce project is that adding a storefront can move your ERP licence through consumption even though it adds no named users, so ask your VAR to model the resource tier before you sign the commerce contract, not after.

The integration build is the line that decides whether a B2B commerce project comes in near the estimate, and it is the line nobody can price from a requirements document. The reason is specific rather than evasive: the cost is driven by the shape of your data and your selling rules, and two distributors of identical size on identical Acumatica versions can differ by a large multiple on this line alone.

Take contract pricing as the clearest example. If your customers all sit in a handful of price classes and prices change twice a year, the effective price can be pre-computed and pushed into a platform price list, which is what the native connectorA native connector is integration software published by the ERP or platform vendor themselves, configured rather than built. Acumatica ships native connectors for Shopify and BigCommerce. is built to do. If instead you hold thousands of customer-item prices with effective and expiry dates, the storefront has to ask Acumatica or Cin7 for the price at the moment the page renders, and that is a development project with caching, failure handling, and a written answer for what the page shows when the ERP is mid-upgrade. Same requirement on paper. Completely different build.

The list below is what an honest estimator is actually pricing. Every item on it is checkable inside your own ERP, which means you can answer them before you ever speak to a vendor, and a vendor who does not ask them is guessing.

Data preparation is the line that surprises people, because it feels like it should be included and it never is. An agency can build you an import, a mapping and a validation report. It cannot write a product description for a part it has never seen, decide which of four spellings of a customer name is the real account, or tell you whether a discontinued item should still be visible to the two dealers who still buy it. Those are your decisions, made by your people, and the calendar time they take is real.

The size of this line is set by how far your ERP data sits from publishable. A distributor whose Acumatica item descriptions were written for a picking slip has a bigger job than one whose descriptions came from a manufacturer feed. A company that has run two acquisitions through the same customer master has a bigger job than one that has not. Neither of those facts appears in a requirements document, and both are visible in an afternoon of looking.

There is a scoping decision inside this line that materially changes the number, and it is worth taking deliberately rather than by default. You do not have to clean everything. You have to clean the range you are launching with. A phase one that publishes the top-selling items your dealers reorder is a fraction of the data work of a phase one that publishes the whole catalogue, and it puts the storefront in front of real buyers sooner, which is where you find out what the descriptions actually need to say.

Run the readiness checks before you accept a data number from anybody, including from us. The checks are free, they take an afternoon, and they turn this line from a guess into an estimate.

Internal time is the line that never gets a budget code and routinely decides the schedule. We are not going to invent an hours figure, because it depends on how many of the roles below one person is covering. What we can tell you is which roles the project cannot proceed without, and what specifically each one has to produce.

The failure mode is consistent and it is worth naming plainly. On most mid-market projects the person who knows how the ERP is really configured, as opposed to how it was documented at go-live, is also the person who runs month-end, owns the EDI mapping, and is the only one who can approve a price change. That person becomes the bottleneck in week two, the project waits on them, and the agency bills for the waiting. Deciding up front who backs them up is cheaper than any change order.

The second failure mode is authority rather than availability. A B2B commerce project generates a stream of questions that are genuinely business decisions, not technical ones: does a dealer see stock at a warehouse they never ship from, does an expired contract price fall back to list or block the line, can a buyer without approval rights see the price at all. If nobody in the room can answer those in the meeting, every one of them becomes a two-week loop.

The integration is not a capital purchase that sits still. It is a live dependency between two systems that both change on their own release schedules, and it costs money every year whether or not anyone has budgeted for it.

The clearest example is an ERP upgrade. Acumatica users have reported commerce sync breakage tied to specific releases, including a template sync error on Shopify Plus after 25R1 and metafield data behaviour changing across an upgrade. None of that is unusual and none of it is a criticism of Acumatica; it is what happens when two products with independent roadmaps are joined at the field level. The question for your budget is not whether it happens but who is on the hook when it does, and whether that is written down.

The licence side moves too, and it moves in the direction of success. BigCommerce upgrades your plan automatically when trailing twelve-month gross merchandise value crosses a threshold. Shopify describes a switch to a variable platform fee for higher-volume merchants, and charges roughly $300 USD a month for each additional store beyond the included allowance. Shopware prices on gross merchandise value as well. A storefront that works therefore raises its own licence line, which is a good problem and still a problem if the three-year business case assumed a flat fee.

The line most often left out entirely is the failed-order queue. Orders will fail to post: a customer will be missing a tax ID, an item will be inactive, a price list row will not resolve. Somebody has to look at that queue every working day and fix the underlying record. If that person is not named before launch, the queue is discovered a month later with a few hundred orders in it, and the sales team stops trusting the channel. Name them in the project plan, not after.