Your Best Customer Is Also Your Lender

Your biggest customer is lending you money. You have been booking it as revenue.

When a bank, a telco or a ministry signs a large contract and then pays in 90 to 180 days, the gap between the work and the cash is a loan. You are the lender. The rate is zero, the term is set by them, and you underwrote none of it. So the real decision sits earlier than any collections process. Price and structure the contract as a financing deal before you sign it, the same way a lender would size and cost a loan it was about to write.

Accounts receivable are a proxy for how much a firm lends to its customers.¹ In African B2B and B2G, that loan is frequently larger than the round you are raising to survive the wait for it.

Why these three models

A late payment looks like one problem, so founders reach for one tool: a collections process. It is three problems wearing one coat, and each needs a different model.

The first is a pricing problem. Different customers can pay at different speeds and they know their own speed better than you do. That is a screening and mechanism-design question: how do you design a contract that makes them reveal it and pays you for the wait. The second is a timing problem. Cash arrives in a queue, and a company with a healthy margin can still starve while it waits. That is queueing and congestion, a complex-systems model where the wait, not the average, decides survival. The third is a spillover problem. One buyer’s delay does not stay with one buyer. It travels down the chain to your suppliers and your staff. That is contagion, modelled the way epidemiologists model an outbreak.

Three outcome types: an equilibrium (screening), a complex system (queueing), and a cycle-regime spread (contagion). The behavioral layer, the founder’s private belief that asking for money upfront signals weakness, is real but not a fourth card. It is folded into the screening model, because that is exactly the reflex screening is built to overrule.

The framework

1. Screening: put a price on the term, and let the customer choose it

Stop treating the payment term as a courtesy you extend and start treating it as a product you sell. A customer who insists on 90 days is buying financing from you. Price it.

The instrument is a menu, not a single number. Quote two prices on the same scope. One for settlement in 14 days. A higher one for net 90, with the difference set at roughly your own cost of capital across the days you wait. Now the term becomes a line item the customer picks, priced and visible on the page, and their choice tells you what their cash actually looks like. The buyer who reaches for net 90 at the higher price has just shown you a balance sheet they would never have emailed you.

A late invoice from a big buyer is a loan you never agreed to make, on terms you never got to set.

Screening also protects you from the worst version of the trade. The customers most eager to accept a long term are, on average, the ones least able to pay on a short one. Suppliers who lend where banks will not do so because they can read a buyer cheaply and can pull the goods back.¹ You usually can do neither. So the menu is your substitute for underwriting: a customer who takes the cash-price option is telling you something an investor would pay for.

Assumes customers differ in how fast they can pay and know it better than you do.

Fits because a two-price menu makes them reveal it and pays you for the wait instead of leaving the loan free.

Breaks when you have a single buyer and no menu to offer, so there is nothing to screen.

Counteracts the reflex to chase a slow payer rather than reprice the segment.

May reinforce the fear that asking for a deposit signals weakness, the behavioral layer folded in here.

2. Queueing: your survival is the wait, not the margin

Model your cash as a queue. Invoices enter when you deliver, wait a variable number of days, then pay out. The number that decides whether you live is the length of that wait and how much it varies, not the profit on any single job.

This is why a company can die with a positive contribution margin. Every deal is profitable and the business still runs dry, because the money is real but it is in the queue and the queue is long. Queueing systems fail the way traffic fails: waiting times do not rise gently as volume grows, they blow up as arrivals approach capacity. Win three big slow-paying contracts at once and you have not tripled your strength. You have tripled the load on a cash queue that was already near its limit, and the wait explodes.

The lever queueing hands you is counter-intuitive. When the wait is killing you, the fix is rarely to chase harder. It is to change the arrival pattern: bill on delivery instead of at month-end, break one milestone into three, take a mobilization advance, and prefer fewer customers who pay fast over more who pay slow. You are shortening the queue rather than maximising the revenue line.

Assumes cash arrives as a queue: work in, a variable wait, payment out.

Fits because survival depends on the wait and its variance, and waiting times rise nonlinearly as the queue fills.

Breaks when payment timing is steady and predictable, where an average is enough.

Counteracts reading a healthy profit-and-loss as a healthy company.

May reinforce over-spending on collections when the real fix is fewer, faster-paying customers.

3. Contagion: one buyer’s delay becomes everyone’s

Arrears are infectious. A negative liquidity shock travels down the trade-credit chain until it reaches a firm with cash or a bank line to absorb it, and the default of one firm can drive its creditors into difficulty even when they were sound to begin with.⁷ Model it like an outbreak. The ministry that pays late infects you. You, short of cash, pay your own suppliers late and stretch payroll. They infect the next firm down.

This is not a metaphor in this market. In South Africa, in one recent quarter, 90,856 supplier invoices sat unpaid past the legal 30-day limit, worth about 15.5 billion rand, and the state itself records that the delay forces suppliers to borrow to meet their obligations, retrench staff, or close.^{3, 8} In Kenya, state corporations alone owed suppliers 379.81 billion shillings, roughly 74 percent of a national pending-bills stack of 516.27 billion.² So treat a single systemically-late buyer as a channel rather than an account. A shock you did not cause can travel through it and land on your own payroll.

The defence contagion models point to is concentration control. An outbreak needs a susceptible, connected host. A receivables book where no single slow payer exceeds a set share of what you are owed does not transmit the same way.

Assumes arrears spread: one buyer’s delay infects you, and yours infects your suppliers and staff.

Fits because a liquidity shock travels down the trade-credit chain until it hits a firm with cash or credit.

Breaks when your customers are independent and unconnected, so no shared shock cascades.

Counteracts concentrating revenue in one large, systemically-late payer.

May reinforce a fire-sale of receivables at a steep discount during a general squeeze.

The levers, from the cheapest edit to the hardest walk-away

Order them by cost and reversibility. Start with the edit you can make on the next quote.

• Reprice the term. Put two prices on every quote. One for cash. One for time. This is a wording change and it is free.
• Ask for money at the front. A mobilization advance of 20 to 40 percent on a public contract, or an annual-prepay discount on a subscription, converts the loan you were going to make into cash you hold.
• Shorten your own queue. Invoice on delivery, not at month-end. Split one milestone into three. Every day you cut off the front of the cycle is a day of financing you no longer provide.
• Cap concentration. Set a rule: no single slow payer above a fixed share of total receivables. Enforce it by pricing the next contract with that buyer higher, or declining it.
• Arrange the facility before you need it. Line up invoice discounting or an overdraft while you are still solvent, knowing it is thin and dear here. Factoring is barely 1.3 percent of global volume in Africa and over 80 percent of that sits in South Africa alone.⁵ This is a hedge, not a plan.

What to do before you sign the next big contract

A dated portfolio. Anchors are relative to the day you read this.

Do now, by T+3 to T+14. Compute one ratio for your three largest customers: the cash they owe you at any moment (roughly their annual spend times their payment days, divided by 365) against the round you are raising. If any one of them owes you more than you are trying to raise, you already know who your lead investor is. Then add the two-price term to your next quote and stop offering net 90 as the silent default. Both moves are reversible and cost nothing but nerve.

Hedge, by T+14. Open a discounting facility or overdraft as tail insurance against a shock, and write down your concentration cap as a number. You are buying the ability to survive one buyer going quiet, not planning to live on borrowed working capital. Replacing that customer’s loan with a bank’s is anyway not on offer for most founders: private credit runs at about 24 percent of GDP across Sub-Saharan Africa against 77 percent in other developing economies, and the continent’s small-firm finance gap exceeds 330 billion dollars a year.^{4, 6}

Defer and trigger, from T+28. Walking away from a marquee public or enterprise contract is irreversible and costs you a logo, so do not do it on feeling. Pre-commit the trigger instead. Write it now: if the projected receivable from one buyer would exceed your concentration cap, or would outrun your cash buffer at their stated payment speed, then you require a mobilization advance or you decline. When the deal arrives, you are executing a rule, not losing your nerve in a room.

What usually happens to a company that funds its buyer

Match the case on structure, not on industry. The reference class is not “startups in my sector.” It is firms with a positive contribution margin and a negative cash-conversion cycle: profitable on paper, dependent on a slow anchor for the cash. Across that class the base rate is unkind. The failure is a liquidity event, not an economics one, and it arrives while the founder is still quoting a healthy gross margin to investors.

Two present-state modifiers make the African version worse than the textbook. Credit to replace the missing cash is scarce and expensive.⁴ The receivables-finance market that exists elsewhere to bridge the gap is thin here.⁵ Both push the base rate up.

Subtract the counterfactual before you over-correct. Some firms in this class would have failed anyway on weak margins, and repricing terms would not have saved them. The lever only rescues the company whose economics are sound and whose death is purely timing. Diagnose which one you are before you act, because the wrong diagnosis wastes the one asset you have left.

The matrix breaks if payment rails change under you. Instant settlement, enforceable statutory payment terms with real penalties, or a working public-sector e-invoicing mandate would shrink the involuntary loan toward zero and demote this whole argument to a footnote. Watch for it. It is not here yet.

What this reading leaves out

The models price the loan. They do not price the relationship. A marque customer can be worth carrying at a financing loss because the logo underwrites your next raise and pulls three more buyers behind it. That is a signaling value the queue cannot see, and sometimes it is the right reason to lend to your customer on purpose. The models also miss the politics of refusing a government buyer in a market where the state is the largest client in the economy, and they miss the founder’s private read that a deposit request will cost the deal. Sometimes it will.

None of that changes the one action that survives the uncertainty. Before you sign, write down the receivable the deal creates and the price of carrying it to the customer’s real payment speed. If you cannot name both numbers, you are not pricing the contract. You are the bank.

Joshua Agonya Pi’Rwot, Founder.

Sources and notes

1. Petersen, M. A., and Rajan, R. G. “Trade Credit: Theories and Evidence.” NBER Working Paper 5602, published in The Review of Financial Studies 10(3), 1997. A firm’s accounts receivable proxy how much it lends to its customers; firms use trade credit relatively more when credit from financial institutions is not available; and suppliers lend to firms banks will not because “they may have a comparative advantage in getting information about buyers cheaply, they have a better ability to liquidate goods, and they have a greater implicit equity stake in the firm’s long term survival.” Abstract and paper.
2. Capital FM Business, “Pending bills drop by Sh106.6bn as State moves to clear debt,” September 2024, reporting Kenya Controller of Budget figures. National government pending bills stood at Sh516.27 billion as of 30 June 2024, down from Sh622.82 billion; state corporations owed suppliers Sh379.81 billion, approximately 74 percent of the national total, with ministries, departments and agencies owing Sh136.45 billion. Article.
3. South African Government News Agency (SAnews), “Non-payment of invoices within 30 days persists in government,” reporting the National Treasury and Public Service Commission quarterly non-compliance statistics. At the end of the third quarter of 2025/26, 90,856 invoices older than 30 days were unpaid by national and provincial departments, worth R15.5 billion, with provinces accounting for 98 percent; the report states “the cash flow of Small, Medium and Micro Enterprises (SMMEs) is compromised, forcing suppliers to borrow money to meet contractual obligations and, in some cases, retrench staff or close their businesses.” Article.
4. Demirgüç-Kunt, A., and Klapper, L. “Financial Inclusion in Africa: An Overview.” World Bank Policy Research Working Paper 6088, 2012. “The ratio of private credit to GDP averaged 24% of GDP in Sub-Saharan Africa in 2010 and 39% in North Africa, compared with 77% for all other developing economies, and 172% for high income economies.” Working paper PDF (text layer present, claim checked in the document).
5. FCI, “2024 Regional Updates: Africa.” Africa represents a 1.3 percent share of total world factoring volume, a 13.5 percent increase on 2022; South Africa dominates, representing over 80 percent of Africa’s total factoring volume. Regional update.
6. TechAfrica News, “Afreximbank and FCI Lead Factoring Conference to Boost SME Finance and Intra-African Trade,” April 2024, reporting figures presented at the Afreximbank and FCI regional factoring conference. Africa carries “over US$330 billion SME finance gap per annum” while factoring remains “a paltry 1.3% of the global factoring volume.” Article.
7. Jacobson, T., and von Schedvin, E. “Trade Credit and the Propagation of Corporate Failure: An Empirical Analysis.” Sveriges Riksbank Working Paper (published in Econometrica 83(4), 2015), using Swedish firm-level trade-credit default data. “A negative liquidity shock is transmitted along the trade credit chain until it reaches a trade creditor with access to external financing or sufficient cash,” and “the default of one firm may cause a chain reaction driving its creditors into difficulty.” Working paper PDF.
8. South African Government News Agency (SAnews), “Non-payment of invoices within 30 days a violation of the public.” Government departments are required to “pay all legitimate invoices from suppliers timeously or within 30 days as required by the Public Finance Management Act,” with National Treasury Instruction No. 34 governing the exception reporting; the Public Service Commission states the practice “is a violation of the public.” Article.