Picture a contractor at 6 a.m., standing on a job site with a crew waiting. He needs a two-and-a-half-inch copper fitting, today. He opens his distributor's website, types exactly that, and gets zero results. So he calls the branch, waits on hold, and while he waits, a competitor's site shows him the part, his price and a delivery window in three clicks.
That distributor did not lose the order on price. It lost it on flexibility.
I spent this week at B2B eCommerce World 2026 in Indianapolis, listening to manufacturers, distributors and researchers from companies like 3M, Emerson, Toyota Material Handling, QXO, Kimball Midwest and Deloitte. Whatever the session was about, the same message kept coming back: in B2B, the webshop that adapts to how each customer buys is winning, and a rigid one is quietly costing companies business they never see.
Ease beats price
One finding from Deloitte's research stopped me: B2B buyers choose the channel that is easiest to buy on, even ahead of price. Most of them are not browsing. They have a job to finish, and they will go wherever that is fastest.
At the same time, suppliers tend to overrate themselves. In the same research, 72% of suppliers described their buying process as automated, but only 47% of buyers agreed. Negative digital experiences were estimated to cost around 12% of sales. Without a real feedback loop from customers, it is very easy to believe your webshop is better than it is.
And the channel itself keeps multiplying. Emerson shared that about 70% of its customers now start their journey online, then move back and forth between online and offline. "B2B commerce" today means a website, punchout into procurement systems, EDI, dealer portals, quotes, marketplaces and, increasingly, the buyer's own AI agent.
What a flexible webshop really means
Flexibility is not a feature list. It is the ability to fit the customer instead of forcing the customer to fit you. In practice, it comes down to a handful of capabilities:
Many ways to buy. Toyota Material Handling serves small customers through its website and enterprise fleets through punchout, catalog integrations, approval workflows and EDI invoicing, all from the same foundation.
Custom catalogs and pricing. Each account sees its own assortment, contract prices and approval rules.
Experiences shaped by role. QXO made a point I loved: out of hundreds of thousands of SKUs, any one customer needs about 20. The estimator, the foreman on site and the accounts team each need something different, so show each of them only that.
Self-service beyond ordering. Order history, invoices, returns, delivery tracking, spec sheets and troubleshooting videos. QXO launched its new mobile app without checkout at all; customers' first wish was simply to see where their delivery truck was.
An open, composable architecture. Emerson was spending two thirds of its time maintaining its old platform. It moved each business group onto its own storefront under one shared commerce core, eliminated search confusion between overlapping product lines, doubled search and configure click-through and cut configuration drop-off by 20%.
Flexibility also means knowing what not to copy from consumer sites. Pop-ups, deals of the day and twelve recommendations add work for a buyer who already knows what he needs.
Flexibility starts with product data
No webshop is more flexible than the data underneath it. Search, filters, configurators, sales tools and AI all read the same product records. If those records are inconsistent, every channel fails in the same way.
Kimball Midwest's e-commerce director won approval for a PIM and an online merchandising team with one slide: the company spelled "one inch" eleven different ways on its website. Another speaker held up a water bottle whose "size" could truthfully be 16.9 ounces, 500 milliliters, eight inches, small or single-serve. Every answer was a fact. None of them agreed.
The simplest test I heard all week: if you gave the definition of an attribute to ten people on your team, would all ten give the same answer? "Capacity" fails that test for a water heater. "Nominal storage volume in U.S. gallons, as stated by the manufacturer" passes. If people cannot agree on a definition, AI will not either; it will just fill the field inconsistently, at scale.
Search is a revenue problem
If customers cannot find it, they cannot buy it. Many B2B sites stall at around 15% online adoption because buyers search in their own words, get nothing back, and pick up the phone. Every one of those calls is a more expensive order and time your best salespeople are not spending on complex sales.
Search matures in stages: exact part-number match, then fuzzy matching, semantic search that understands what the buyer means, conversational search that asks clarifying questions, and finally context-aware search that knows the buyer's trade, site and order history. Direct Supply shared that it built 2,000 to 3,000 manual search rules and still only covered its top hundred terms, because most searches on its site happen exactly once. Only machine learning scales to that.
Make the webshop your sales team's best partner
In B2B, customers follow their sales rep. A webshop the sales team does not trust will not be adopted by customers either.
Kimball Midwest pays its 1,500 reps the same commission no matter which channel an order comes through, and reps now wake up to online orders they did not have to take. The pattern I heard again and again: let the website farm the routine reorders, delivery questions and invoice lookups, so reps can hunt for new business. Give reps the same site their customers use, with spec sheets, manuals and videos on every product page, and measure self-service and digitally influenced sales, not just online revenue.
Ready for the buyer's AI agent
The next new channel is not a website at all. In Deloitte's research, 38% of buyers already use agentic AI in procurement, compared with 4% of suppliers selling with it. Procurement teams are moving first.
Agents need machine-readable product data, open connections such as the Model Context Protocol, and answers that can be traced to a source. 3M designed its Ask3M assistant to serve other companies' agents, not only people on its own site. The good news: this does not require starting over. Agentic commerce can be added as a layer on top of an existing platform, one use case at a time, provided that platform is open enough to accept it.
The international lens
This is where my own work comes in. For a company entering a new market, a flexible webshop is often the fastest and cheapest way in, and a rigid one is a real barrier.
Every market buys differently. Emerson sells through channel partners for some businesses in North America and through a mix of partners and direct sales in Europe. A platform that assumes one route to market cannot follow you abroad. The model that works is one shared core with local storefronts, each carrying its own language, currency, tax rules, units of measure, assortment, compliance data and local channels. A European supplier's 189 liters and a U.S. buyer's 50 gallons must land in the same, correctly converted field.
A new-market storefront is also free market research. Before you invest in branches, warehouses or a local sales team, it shows you what buyers search for, what they cannot find and which accounts come back.
Five mistakes I see companies make
1. Building without the sales team. If reps fear for their commission or the warehouse calls it "another e-commerce order", adoption stalls, no matter how good the technology is.
2. Turning adoption into a quota. Mandates and rebates produce reps keying phone orders into the website. When the incentive ends, the numbers fall.
3. Leaving no one in charge end to end. When one team owns the PIM, another the site and a third the ERP feed, the customer falls into the gaps between them.
4. Treating data cleanup as a one-off project. New SKUs, suppliers and markets arrive every week. Governance has to be a practice, not a project.
5. Expecting AI to fix bad data. AI fills fields faster, but without clear definitions it fills them faster in the wrong way.
Where to start
If you want to make your webshop more flexible, here is the approach I recommend:
Listen first. Talk to customers and reps and pick the one to three jobs buyers most want to do online.
Start with one category. Choose a high-selling category, define its top attributes until they pass the ten-out-of-ten test, then let AI help fill them.
Measure before and after. Track search success, conversion and self-service share, so you have a result to show leadership.
Name an owner. One person responsible for each customer journey and one for product data.
Build open. Choose architecture that lets you add the next channel, market or AI capability without a rebuild.
The bottom line
A webshop is no longer a catalog with a checkout. It is how your customers, your sales team and soon your customers' AI agents experience your company. The suppliers who win will not necessarily have the lowest prices. They will be the easiest to buy from, in every channel and every market they serve.
The question is not whether we have a webshop. It is whether our webshop can bend to how our customers buy.
At PLA International, I help companies enter new markets, find the right partners and channels, and build the processes, including digital sales channels, that make international growth work across North America and Europe. If you are rethinking your webshop or planning your next market, I would love to hear from you. Book a conversation or connect directly.
#B2BCommerce #Ecommerce #DigitalTransformation #ProductData #AgenticCommerce #MarketEntry #InternationalExpansion #BusinessDevelopment #Distribution #Manufacturing