Small industrial buyers need credit at the point of purchase, and the producers who supply them already hold the data, documents and digital channels to underwrite it.
A small maker of plastic kitchenware buys polymer every week. Polymer is its largest cost, it has to pay for it long before its own retail customers pay, and every order comes with the same negotiation about when payment is due. Multiply that by hundreds of thousands of small converters and farm businesses, and buying on credit from an industrial supplier turns out to be one of the least digitised journeys in B2B commerce.
In commoditised industries such as polymers, steel, cement and fertilisers, products are close to interchangeable and prices follow the market, so what wins a small business is often the terms. Aleksei Plekhanov, a London-based fintech product and strategy leader specialising in trade finance, payments and risk infrastructure for SMEs in global trade, argues that those terms should be treated as a digital product. “Payment deferral is the main example, but the same approach applies to logistics and other services around an industrial sale,” he says.
The users and what hurts
The users are owners and finance managers of small and mid-sized businesses whose main input comes from a large producer. They pay for raw material weeks or months before their own customers pay them, so working capital is their core problem. They clearly want credit. Atradius’s 2025 surveys found that about half of B2B sales in chemicals are made on credit in markets from Switzerland to Türkiye and India, with steel and machinery at similar levels. The ECB’s late-2025 survey showed firms’ need for trade credit rising while its availability stayed flat.
“What goes wrong is the way credit reaches them, and in my experience the same complaints come up again and again,” says Plekhanov. “Requesting deferral is slow and manual, done through calls, emails and paper and negotiated order by order with a sales manager. Bank finance means a second application that duplicates paperwork the supplier already holds. Banks do not understand the business, because seasonal volumes and thin accounts meet conservative, generic risk models, and the answer is a rejection or a limit too small to use.”
“And there is no transparency,” he continues. “Users do not know how much they can buy on credit until they ask, and asking is the slow part. There is no single record of orders, deliveries and payments either.”
“All four are product problems. The money is there, but the way to it is broken.”
Credit at the checkout
In consumer e-commerce, Klarna moved credit from a separate bank application to the checkout page. More and more of the small clients of industrial producers already buy through the producers’ own digital platforms, and in Plekhanov’s view the producers can now do the same.
He describes how such a flow would work. “The order screen shows a pre-approved limit and the available terms, such as 30, 60 or 90 days, with the price of each option calculated in real time. Onboarding happens once and digitally, and after that buying on credit takes one button. Delivery is booked and tracked in the same flow. Every order, delivery, invoice and repayment sits in one record that the user, the producer and the lender all see.”
“The decision then takes minutes instead of days, and the user does the paperwork once,” he says. “The limit is based on the user’s own trading history, and the user can see it before ordering.” Once the flow exists, he adds, other services such as inventory finance for seasonal peaks, early-payment discounts for cash-rich users, cargo insurance and FX can join it at low cost.
India provides an example of this model at scale. According to JSW One Platforms, the company recorded gross merchandise value of ₹12,567 crore in FY2024–25 and facilitated approximately ₹3,800 crore in credit disbursements through banking and non-banking financial company partners. The platform also reported more than 84,000 registered MSMEs. These company-reported figures illustrate the growing integration of credit facilities into digital B2B procurement platforms.
Lessons from building it
Plekhanov has worked on this problem from three sides. At a top-10 global petrochemical producer he led the portfolio of commercial digital products, and payment deferral was the service that mattered most to small and mid-sized users. Around 20 to 40 per cent of purchases of the core product came with deferral attached. At a UK trade finance scale-up he led the launch of a financing product for buyers in different markets. At a UK payments business he rebuilt the risk scoring and checks behind client onboarding, which cut onboarding costs by 40 per cent.
“The speed and predictability of the decision matter as much as price,” he says. “A limit the user can see before ordering changes behaviour. An application that takes a week does not.”
He argues that the product needs product metrics next to the loan-book ones. “The core metrics are activation (the share of eligible users who accept a limit), usage (the share of their orders placed on credit), time from request to decision, repeat use, and incremental volume per user against a control group, tracked together with overdue rates.”
A pilot, in his view, should be small and fully instrumented. “A handful of real transactions with real users shows more about onboarding friction and document gaps than months of design.”
Why Producers May Have an Advantage in Credit Risk Assessment
“The product works only if the credit decision is fast and sound, and here the producer has an edge no bank can match,” Plekhanov says.
For a kitchenware maker polymer is the largest cost line, so its polymer purchases track its production. “Its payment behaviour on that input is the most honest signal of its health,” he says. “A bank sees annual accounts that are months old. The producer sees weekly orders, deliveries, disputes and the exact days each invoice took to settle, often over many years.” A large producer also sees the whole industry and can benchmark. “A converter whose volumes fall while its segment grows is a different risk from one following a seasonal dip that every peer shares.”
“Turning this into decisions is an engineering job,” he continues. “It needs entity resolution (so that one owner trading through several companies counts as one exposure), an order-to-cash ledger joined from the ERP, registry and bureau enrichment, and a scoring model with dynamic limits and early warnings. The data plumbing is usually harder than the model, and so is the governance that lets a lender trust the output, which means backtesting, drift monitoring and a reason code for every decision.”
Documents are the second edge. The invoice, waybill and proof of delivery come from the producer’s own systems, so the lender funds against documents issued by the party that shipped the goods. “That largely removes the invoice fraud and double financing that plague independent receivables finance,” Plekhanov says. Although producer-generated transaction records can strengthen verification and reduce certain fraud risks, they do not eliminate the possibility of duplicate financing, document manipulation or operational failures. Effective controls, lender coordination and independent verification remain important.
In the UK, the Electronic Trade Documents Act 2023 gives such documents legal standing.
In the model he describes, the lender receives a scored, verified transaction through an API, funds it and is repaid at maturity. “The bank becomes an essential but interchangeable component, and the producer can add or swap funders without the user noticing.”
The business model
Between becoming a bank and handing its clients to one, a producer has several options.
| Model | Brand the user sees | Who lends | Who bears credit risk | Producer’s control |
| Marketplace referral | Bank | Bank | Bank | Low |
| White-label | Producer | Bank | Bank | High |
| White-label with credit support | Producer | Bank | Shared, producer takes first loss | High |
| Captive lender | Producer | Producer’s own finance arm | Producer | Full |
The appropriate financing structure will depend on applicable lending regulations, the allocation of credit risk and the responsibilities of participating institutions. Even where producers supply transaction data and manage the customer experience, regulated lenders must maintain appropriate oversight of underwriting, affordability or creditworthiness assessments where applicable, data protection and ongoing risk management. These requirements can vary considerably across jurisdictions.
“For most producers the middle rows work best,” says Plekhanov. “The user sees one brand and one checkout, and the producer sets the experience, the limits and the pricing logic while the capital comes from banks. A first-loss guarantee is often what persuades a bank to accept a producer-built risk model.” JSW One’s platform terms provide for exactly this kind of credit support.
White-labelling also allows selective subsidy. “The producer can pay part of the financing cost to enter a new region or win a competitor’s key account,” he says. Car and farm machinery makers have used subsidised finance to move their core product for decades. “Now a subsidy can be aimed at a single client segment, and the volume it actually buys can be measured.”
According to Plekhanov, the producer earns from higher and steadier volumes of the core product, a share of financing fees and lower overdues on its own book, and it builds a data asset that grows with every transaction. He sees a captive lender as a later step for producers that find the credit business attractive in its own right. OfBusiness, which sells steel, polymers and chemicals to Indian SMEs, built its lending arm Oxyzo into a profitable finance company, with ₹339 crore (approximately £30 million) of net profit in FY25.
Where the model can go wrong
Plekhanov also points to the risks. “The first-loss guarantee that brings banks in has a cost,” he says. “Much of the credit risk can stay on the producer’s balance sheet for accounting purposes, so external capital is only a partial transfer.”
“Subsidies are price cuts,” he adds. “They invite competitors to match, and without controlled tests by region or segment it is impossible to tell new volume from discounted purchases that would have happened anyway.”
Competition law is another constraint. “The industry-wide view that makes a large producer a strong underwriter is also its biggest regulatory exposure under competition law,” Plekhanov says. “User consent for data sharing, an information barrier between sales and credit, and published eligibility criteria have to be part of the design from the start.”
“The rollout also has to decide whether distributors become partners on the platform or competitors to it,” he says. In chemicals and metals, distributors often carry credit today.
References:
Atradius — 2025 B2B payment practices surveys.
European Central Bank (ECB) — Survey on the Access to Finance of Enterprises, Q4 2025.
JSW One Platforms — FY2024–25 business performance and credit disbursements.
CRISIL — Oxyzo Financial Services rating report, January 2026.
UK Government — Electronic Trade Documents Act 2023.
