A procurement manager needs forty industrial pumps. She knows the spec, she has bought from this supplier before, and her company has a negotiated contract. Ten years ago this meant an email to a named rep, a wait for a quote, a phone call to confirm stock, and a back-and-forth on payment terms. Today she would rather log into the supplier's site, see her contract pricing, check live availability, drop the order into an approval workflow, and be done before lunch. If the supplier cannot offer that, she has a competitor's tab already open.
That preference is not a niche. Gartner's most recent sales survey, run across August and September 2025, found that 67 percent of B2B buyers prefer a rep-free buying experience, up from 61 percent the year before. Buyers want to research, configure, and place orders, including very large ones, without a salesperson in the loop. The consequence for anyone running a B2B business is concrete and often underestimated: the commerce platform now has to carry work that used to be a person's job. The catalog has to know who is logged in. The pricing has to reflect a contract. The checkout has to handle approvals, credit, and net terms. The salesperson did all of that by hand. Now the software does, or the sale does not happen.
Where B2B ecommerce came from
Businesses have been buying from businesses electronically for a long time, but for decades it was not what anyone would call self-service. Electronic Data Interchange, the structured machine-to-machine exchange of purchase orders and invoices, dates to the 1960s and became the back-office plumbing of large-scale procurement. It worked, but it was a fixed pipe between two companies, set up by IT, invisible to the person actually doing the buying.
The web changed the front of that process before it changed the back. Grainger, the industrial supplies distributor, launched its ecommerce site in 1995, an early bet that buyers of maintenance and repair parts would order online. In the late 1990s, Ariba and other procurement vendors created cXML, an XML-based protocol that let a buyer's purchasing system connect out to a supplier's catalog. That gave us punchout, a mechanism still central to B2B today, where a buyer inside their own procurement tool jumps into the supplier's store, shops, and returns a filled cart for internal approval.
For a long time these systems served the largest buyers and the largest suppliers, and everyone else still phoned a rep. The self-service expectation that now reaches the mid-market is recent. It was accelerated by the pandemic, when in-person selling stopped working overnight, and it has been cemented by a generational handover. Millennials and Gen Z now make up the large majority of B2B buyers, and they bring consumer habits with them. They grew up buying everything else on a screen, comparing options without help, and treating a phone call as a last resort. They expect their work purchasing to behave the same way.
Where it stands now: the buyer has changed
The headline number is the rep-free preference, but the more revealing data is about size and behavior.
Start with size, because it breaks the old assumption that self-service is only for small, repeat orders. McKinsey's B2B Pulse survey, which polled 3,942 decision-makers across thirteen countries, found that 73 percent of B2B buyers are willing to spend more than 50,000 dollars on a single online order, up from 59 percent two years earlier. At the top end, 39 percent would place an order above 500,000 dollars, and 20 percent would go past a million. Buyers are not just reordering paper through a portal. They are signing off on six- and seven-figure purchases without wanting a sales call to do it.
Then look at how they research. The modern B2B buying journey is long, non-linear, and mostly conducted in private. Gartner's work on buying groups has consistently shown that buyers spend only about 17 percent of their total journey meeting with potential suppliers at all, and when they are comparing several, the share with any one supplier can fall to roughly 5 or 6 percent. The rest is independent research, peer input, and internal discussion. Gartner's 2025 survey found buyers consulting an average of seven information sources during a single purchase. Generative AI is now one of them: 45 percent of buyers said they used GenAI during a recent purchase, mostly to gather information on vendors and products. By the time a buyer is ready to act, they have usually done most of the thinking already. They do not want education from a rep. They want to transact.
The market backdrop matches the behavior. Global B2B ecommerce is large and still growing fast, with the total market projected near 37 trillion dollars in gross merchandise value in 2026, several times the size of global B2C ecommerce. Most B2B firms now operate a storefront or buyer portal. The question has moved past whether to sell online. It is whether the online channel can actually do the job the rep used to do.
What real B2B self-service requires
This is the part that gets underestimated. A B2C storefront shows one catalog at one set of prices to everyone. B2B self-service has to do something much harder: reproduce, in software, the entire commercial relationship a salesperson used to hold in their head and their inbox. That breaks down into a specific list of capabilities.
Account-specific catalogs and pricing. Two customers logging into the same site should often see different things. One has a contract that covers 800 SKUs at negotiated rates; another sees list price on a narrower range; a third has volume tiers that change the unit cost at 100, 500, and 1,000 units. The platform has to resolve all of that at the moment of login, per account, per contract. Show the wrong price and you have either lost money or lost trust.
Quotes and configuration. Plenty of B2B orders are not a fixed SKU at a fixed price. They are configured, negotiated, or both. Self-service does not remove the quote; it absorbs it. Request-for-quote tools let a buyer assemble a complex order and ask for a price without a phone call. Configure-price-quote, or CPQ, logic lets the buyer build a valid configuration of a complex product and get an accurate price on the spot. The rep used to be the CPQ engine. Now the software has to be.
Approvals and buyer hierarchies. A company is not one buyer. It is a junior purchaser who can spend up to a limit, a manager who signs off above it, and a finance owner who controls the budget. A real B2B platform models that org chart, routes an order through the right approval chain, and supports requisition lists so a team can build a shared order over days. Get this wrong and the buyer's own internal process breaks against your checkout.
Credit and payment terms. Consumers pay with a card. Businesses frequently buy on account, with net 30, net 60, or a negotiated credit line. In one survey of UK and European B2B buyers by Hokodo and the B2B Ecommerce Association, 83 percent said they would abandon a purchase if no payment terms were offered at checkout. Bank transfer and direct debit matter more than cards. Self-service checkout has to carry credit limits, terms, and purchase-order payment, not just a Stripe field.
Reordering and self-management. Much B2B buying is repeat buying. Fast reorder from order history, saved lists, and subscription-style replenishment turn the platform into the thing a buyer reaches for by habit. Alongside that, buyers want to manage their own account: track shipments, pull invoices, manage users, see contract status, all without opening a support ticket.
Punchout and ERP integration. Large buyers do not want to leave their own procurement system. Punchout, built on cXML or OCI, lets them shop the supplier catalog from inside SAP Ariba, Coupa, or Jaggaer and pull a cart back for internal approval. And none of the above works on stale data. The catalog, pricing, inventory, and order status have to stay synchronized with the seller's ERP in close to real time. The old separation, with ecommerce at the front and EDI in the back office, is collapsing into a single connected flow, because a buyer placing a six-figure order self-service needs to trust that the stock number on the screen is true.
Miss any one of these and self-service quietly fails. The buyer hits the gap, cannot finish the order alone, and either calls a rep, which defeats the point, or leaves.
Where self-service stops and humans still matter
Here is the part the rep-free statistic hides, and it is the most important point in this piece. Buyers wanting to avoid reps does not mean reps add no value. Gartner's research on rep-free buying has also found a sharp downside to going it alone. Purchase regret runs about 23 percent higher among buyers who choose a rep-free experience. And buyers are 1.8 times more likely to close what Gartner calls a high-quality, low-regret deal when they combine supplier digital tools with a sales rep, rather than using either on its own.
The AI angle sharpens this further. In a Gartner survey reported in May 2026, 69 percent of B2B buyers said they turn to a sales rep to validate AI-generated insights. Buyers will happily use ChatGPT to research a vendor, then want a human to confirm that what the machine told them is actually right for their situation. Gartner has gone as far as to project that by 2030, 75 percent of B2B buyers will prefer sales experiences that prioritize human interaction over AI. The preference for rep-free is real, but it is a preference for self-service on the buyer's terms, not a wish for the rep to vanish.
So the line is not self-service versus people. It is self-service for the routine and the known, people for the complex and the uncertain. A buyer reordering a known SKU on a known contract wants zero friction and zero conversation. A buyer specifying a first-of-its-kind system, negotiating a multi-year supply agreement, or making a high-stakes switch wants a human to pressure-test the decision. The platforms that get this right do not pick a side. They make the handoff between self-service and a rep frictionless in both directions: a buyer can escalate a self-service cart into a negotiated quote, and a rep can hand a finished quote back for the buyer to check out themselves. The reps who survive this shift are not order-takers. They move to later-stage, higher-value work, complex deal support and helping buyers feel confident about a hard decision, while the platform handles the transaction.
What platforms get this right, and where it is heading
B2B capability is now a real axis of competition between commerce platforms, and they are not equal on it. Adobe Commerce has the deepest native B2B feature set of the major platforms, with request-for-quote, negotiable quotes, multi-level approval chains, requisition lists, buyer-specific shared catalogs, and credit management available without custom development, which is a large part of why it remains common in complex manufacturing and distribution. BigCommerce ships a B2B Edition aimed at strong mid-market value, with a buyer portal, quoting, and price lists. Shopify has invested heavily in B2B on Shopify, folding B2B and direct-to-consumer into one store with company accounts, catalogs, price lists, and payment terms, and it is the fastest-improving option, though it still leans on apps for some advanced procurement workflows. SAP Commerce Cloud and commercetools serve the largest enterprises, the former where the buyer is already deep in SAP, the latter where a composable, API-first build is justified. For punchout and procurement-system connections, specialist integrators bridge these platforms into Ariba, Coupa, and Jaggaer.
Two shifts will shape the next few years. The first is that agents start to act on the buyer's side. Buyers already research with AI; the next step is an agent that places and reorders against a contract, or negotiates a quote with the seller's system. That raises the bar on machine-readable catalogs, clean pricing data, and well-documented APIs, because an agent cannot read a PDF price list or phone a rep. The second is consolidation of the buying journey itself. The buyer using seven information sources and a chatbot before they ever reach your site means the storefront is no longer where the decision is made. It is where a decision already taken gets executed. That puts the weight on product data, accurate availability, and a checkout that can handle the buyer's real commercial terms.
For an enterprise selling B2B, the practical work is to audit the self-service path against the capability list above, find where it still forces a phone call, and close those gaps, then keep the catalog and pricing data clean enough for an AI assistant to read. Perform Digital works with enterprises on exactly that: making the commerce platform carry the work the rep used to do.
The buyer has already changed. They will research without you, shortlist without you, and place a large order without you, if you let them. The platform is no longer a brochure with a cart bolted on. It is the salesperson, the price book, the credit desk, and the order desk, running without a lunch break. The B2B businesses that win the next few years are the ones that build it that way on purpose.
Council summary
This post argues that B2B self-service has outgrown small reorders: buyers now want to place six- and seven-figure orders without a sales call, which forces the commerce platform to reproduce the contract pricing, quoting, approvals, credit, and ERP sync a rep once handled by hand. The council verified the core figures against primary sources. Gartner's 67 percent rep-free preference (646 buyers, August to September 2025, up from 61 percent) and the McKinsey B2B Pulse spend data both hold. Three claims were corrected: the 36 trillion dollar figure is the total B2B ecommerce market, not cross-border; the cart-abandonment statistic is 83 percent abandoning when no payment terms are offered, not two-thirds; and the 17 percent journey figure is total time across all suppliers, with any one supplier closer to 5 or 6 percent. The takeaway for a B2B seller is to find every point in the self-service path that still forces a phone call, close it, and keep catalog and pricing data clean enough for the AI tools buyers now use.
Comments