Acro Commerce

Decision guides

What a B2B storefront actually costs over three years

Year one is the quote you were given. Years two and three are the licence increases you didn’t model, the connector maintenance nobody owns, and the ERP tier you cross when orders grow. On the projects we see, the build is roughly 40 to 55 per cent of the three-year number, and the line that surprises people most is the cost of keeping two systems agreeing with each other.

Why the quote and the bill disagree

A build quote answers one question honestly: what will it cost to get this live. It isn’t designed to answer the one your CFO asks eighteen months later, which is what this thing costs to own. The gap between them isn’t dishonesty. It’s that most of the recurring lines belong to somebody else, so nobody puts them on one page.

The platform licence is on the platform’s price list. The ERP tier is on the ERP’s. The connector licence is a separate line from your ERP partner. The hosting is a fourth invoice, and the work of keeping the two systems agreeing with each other is a person, not an invoice at all. Five suppliers, one bill, and only you can see the whole of it.

So the useful exercise is not comparing build quotes. It’s building the three-year number once, for each shortlisted option, using the same seven lines. The differences between platforms usually show up in years two and three rather than in the quote.

The seven lines

These are the lines we can put a shape to. The dollar figures depend on your volume and your pricing complexity, so what we’ve given you here is what drives each one and when it moves, which is the part vendors don’t explain.

The seven lines in a three-year B2B commerce number
CriterionWhat drives itWhen it moves
The buildThe line everyone negotiates, and usually 40 to 55 per cent of the three-year total.Scope, data condition, and how far your pricing sits from the platform’s model of pricing.Once, plus whatever the first six months of corrections costs.
Platform licenceShopware’s B2B Components need Evolve rather than Rise. BigCommerce price list export from Acumatica needs Enterprise. Check the plan, not the product.Often a plan tier, sometimes gross merchandise value, sometimes both.When you grow, and when you need a B2B module that sits on a higher plan than the one you budgeted.
ERP tierThe line most often left out entirely, because the ERP is treated as a fixed cost that already exists.On Acumatica, your busiest single document type in a month, not the sum of them.On your peak month, not your average one. A November spike is priced as November.
Connector licenceOn Acumatica no connector is included in any edition, and at the entry tier you can’t buy a second storefront at all.Usually one licence per storefront, with B2B, tax, and point of sale as separate items.When you add a second storefront, a second brand, or a second region.
Hosting and infrastructureTraffic, catalogue size, and whether you self-host.Slowly, unless you run a headless front end, in which case it’s a real engineering line.
Owning the integrationThis is the one that gets discovered rather than budgeted. See the connector guide.How many mappings you’ve, and how often either vendor ships a release.Twice a year on Acumatica, continuously on a hosted platform.
Internal timeNot free just because it doesn’t leave the building.Content, data cleanup, sales team change, and answering the questions the storefront now raises.Heaviest before launch and in the first quarter after it.

What changes the answer most

Two variables move a three-year number more than everything else combined, and neither is on a price list.

The first is where price gets calculated. If your ERP holds negotiated prices and the storefront can carry them through a native connector, you’ve a configuration project. If it can’t, you’re buying middleware or a build, and that decision changes both the year-one number and the person you need on staff afterward. It’s worth settling before you shortlist platforms, not after.

The second is how many storefronts you’ll end up running. Sellers plan one and arrive at three: a brand, a region, and a dealer channel. Connector licensing is usually per storefront, some ERP tiers cap you at one, and a second storefront on a synced architecture roughly doubles the reconciliation work. Ask what the second one costs while you still have leverage.

The seven questions to put in the RFP

Ask these of every vendor and every implementation partner, and ask for the answer in writing. None of them can be answered with a yes.

  1. Which system calculates the price a logged-in customer sees, and what happens on the day a contract price changes?

  2. What does a second storefront cost, in platform licence, connector licence, and implementation?

  3. Which plan tier do we need for the B2B features in this demo, and is that the tier you quoted?

  4. When you ship a release, what breaks, and who fixes it?

  5. What does support cost after the warranty period, and what does it cover?

  6. What happens to the storefront when the ERP is unavailable?

  7. What would you expect us to spend internally in the first six months after launch?

A rough shape, and why we won’t give you a number

We could publish a table of dollar figures and it would be wrong for most readers, because the range on a mid-market B2B build is wide enough that an average is misleading rather than useful. Two sellers with the same revenue and the same ERP can differ by a factor of four on the strength of their item master alone.

What we’ll say is the shape. Build lands somewhere near half the three-year total. Recurring licences across platform, ERP, and connector are the next largest block and they only go up. Integration ownership is the smallest line on paper and the one that most often gets abandoned, at which point it becomes a rebuild. If your three-year model has no line for it, the model is wrong regardless of the numbers in it.

If you want the shape of your own number rather than the shape of a generic one, the diagnostic asks about the mechanics that actually move it.

The wider picture

This page answers one narrow question. Acro Commerce covers the strategy around it.

Common questions

Is a hosted platform cheaper than a self-hosted one over three years?
Often, and not always, and the crossover depends on your pricing complexity rather than your traffic. Hosted platforms trade a licence for engineering time, which is a good trade until the thing you need sits outside what the platform models. At that point you pay the licence and the engineering. Sellers with straightforward pricing are usually better off hosted; sellers whose pricing is the reason they’re replatforming should model both.
How much should we budget for year one after launch?
Plan for real spend rather than a maintenance retainer. The first quarter after launch surfaces the data problems that survived testing, the pricing edge cases nobody documented, and the sales team’s workarounds. Sellers who budget nothing for that quarter tend to spend it anyway, unplanned, and lose the goodwill that comes with a launch.
Does an ERP-first architecture cost more?
It costs more to build and less to argue about. Reading prices from the ERP at request time removes a whole category of reconciliation work, and it makes the storefront depend on the ERP being available. Syncing is cheaper to stand up and moves the cost into someone’s ongoing attention. Both are defensible. What isn’t defensible is not knowing which one you bought.

Last updated 2026-08-24. Facts on this page last checked against source 2026-08-24.