Cash Flow Forecasting for Kenyan SMEs: The Two-Week Warning
Profitable Kenyan businesses run out of money regularly, and almost always with two weeks of warning that nobody was looking at. A thirteen-week rolling forecast you can actually maintain, built from records you already have.
Profit and cash are different things, and the gap between them is where otherwise healthy Kenyan businesses fail. You can have a record month, a full order book and a growing customer base, and still be unable to pay salaries on the 28th — because profit is recognised when you invoice and cash arrives when the customer decides.
The frustrating part is that this is almost always foreseeable. The information needed to see a squeeze forming two to three weeks out sits in the receivables ageing, the payables list and the payroll calendar. What is usually missing is not data but a single forward view that someone looks at every week.
Why growth causes the squeeze
This is the counter-intuitive part, and it catches good businesses precisely because they are doing well.
Growing means buying more stock before you sell it, extending more credit to more customers, and often hiring ahead of the revenue. Every one of those consumes cash now against revenue later. So the faster you grow, the wider the gap — and a business growing at 40% a year can be more fragile than the same business standing still, while every profit measure says the opposite.
The cash gap a growing business finances
This is the cash conversion cycle. At 65 days, every additional shilling of monthly sales needs roughly two months of funding before it returns — which is why growth consumes cash. The mechanics are in working capital and the cash conversion cycle.
Growth is a cash consumer, not a cash generator. A business expanding fast on thin working capital is not succeeding safely — it is succeeding at increasing speed towards a wall it cannot see.
The thirteen-week rolling forecast
Thirteen weeks is the useful horizon for an SME: long enough to see a quarter's obligations, short enough to be genuinely predictable. Rolling means it moves forward every week rather than being rebuilt each quarter, which is what stops it becoming a document.
It has four rows and one output.
| Row | Where it comes from | The discipline |
|---|---|---|
| Opening cash | Bank, reconciled | Weekly reconciliation, not monthly — otherwise week one starts wrong |
| Money in | Receivables ageing, by expected date not due date | Use how each customer actually pays, not their agreed terms |
| Money out — committed | Payables, payroll, statutory, rent, loan repayments | Include commitments, not just invoices received |
| Money out — discretionary | Stock purchases, capital spend, anything deferrable | Kept separate, because this is the row you can actually flex |
The output is closing cash per week, and the only number that matters is the lowest point across the thirteen weeks. If that trough is below your comfort threshold, you have a problem — and crucially, you have it several weeks before it arrives, when discretionary spending is still discretionary.
Expected date, not due date
This single distinction separates a forecast that works from one that is optimistic every single week.
A forecast built on invoice due dates assumes every customer pays on terms. They do not, and you already know which ones do not — that knowledge lives in payment history you can see. A customer who has taken 55 days on every invoice for two years will take 55 days on this one, and forecasting them at 30 is not optimism, it is choosing to be wrong in a predictable direction.
So forecast each significant customer at their demonstrated behaviour. The forecast becomes less pleasant and dramatically more useful. And the practice of tightening that behaviour is a separate, parallel exercise — covered in receivables and collections.
The trap of a forecast that is always right
If your forecast keeps matching reality, check whether it is actually forecasting. Many SME "forecasts" are last month repeated with a growth percentage applied, which will look accurate in a stable month and tell you nothing about the one where a large customer slips and a tax payment lands in the same week. A forecast earns its keep in the bad weeks.
What to do when the trough appears
Seeing a squeeze three weeks out is only valuable if there is a playbook, because the instinct in the moment is to do the most damaging thing available — delay salaries or statutory payments.
-
Defer discretionary outflows first
Stock purchases, capital spend, anything with no fixed date. This is why keeping discretionary separate in the forecast matters — it is the lever you can pull without consequences.
-
Accelerate specific collections
Not a general chase. Name the three largest invoices in the affected window and work those, with an agreed payment date recorded against each account.
-
Talk to suppliers before you are late, not after
A supplier told a week early will usually agree; the same supplier told a week late has already been surprised and starts from a worse position. Credibility is a cash management asset.
-
Draw on facilities deliberately
If you have an overdraft or facility, planned use is normal and cheap relative to the alternatives. Unplanned use at the last moment costs more and signals worse.
-
Protect salaries and statutory obligations last of all
These should be the final things to move, not the first. Late statutory payments attract penalties and late salaries cost you people — both are far more expensive than deferring a stock order.
What we do and do not do
What AWRA OpsHub does today
- Receivables ageing by customer and invoice, with payment history visible so expected dates are grounded in behaviour.
- Payables with due dates, plus commitments visible at approval rather than only when invoiced.
- Payroll and statutory liabilities in the same ledger as everything else.
- Bank reconciliation, so opening cash is a fact rather than an estimate.
- Budget controls that show committed against available before more is spent.
- A report builder for your own forward views from live data.
What it does not do
- We do not produce a predictive cash forecast automatically. The system supplies grounded inputs; the forecast is assembled by you or your accountant.
- We do not arrange finance or introduce lenders.
- We do not advise on treasury, hedging or borrowing.
- We cannot make a customer pay. Ageing and statements are visibility, not leverage.
This is not financial advice. Decisions about borrowing, deferring obligations or prioritising payments should be taken with your accountant or financial adviser, and statutory obligations should be confirmed with the relevant authority.
Making it a habit rather than a project
The forecast that survives is the one that takes twenty minutes a week and is looked at by someone with authority to act. Same day each week, same four rows, rolled forward one week. If it takes half a day to produce, it will be produced twice and then quietly abandoned — which is the fate of most cash forecasts in growing businesses.
That is why the inputs matter more than the model. Reconciled bank weekly, ageing that reflects real behaviour, and payables including commitments — get those and the forecast is largely assembly. Skip them and you will spend the twenty minutes arguing about the opening balance. The records side is covered in financial management software in Kenya.
Our take
Build a thirteen-week rolling forecast, use each customer's demonstrated payment behaviour rather than their agreed terms, keep discretionary outflows in their own row, and look at the lowest point every week. Twenty minutes weekly is the cheapest insurance a growing Kenyan business can buy — and the businesses that fail on cash almost never had it.
See the inputs a real forecast needs
Receivables ageing with payment history, payables and commitments, payroll and statutory liabilities, and a reconciled bank position — all on one ledger.
Explore cash & ageing visibilityFrequently asked questions
How can a profitable business run out of cash?
Because profit is recognised when you invoice and cash arrives when the customer pays. If stock sits for 45 days, customers take 50 days to pay and you pay suppliers in 30, you are funding roughly 65 days of operations from your own resources — and every extra shilling of sales widens that gap. This is why growth consumes cash and why a fast-growing business on thin working capital can be more fragile than a stagnant one.
Why thirteen weeks rather than twelve months?
Because thirteen weeks is both long enough to capture a quarter's obligations and short enough to be genuinely predictable. Annual cash forecasts for SMEs are largely fiction beyond the first quarter, and their inaccuracy teaches people to ignore the whole document. A rolling thirteen-week view that moves forward each week stays close enough to reality that people trust it and act on it.
Should we forecast customer payments on due dates?
No — forecast on how each significant customer actually pays, which their payment history already tells you. A customer who has taken 55 days on every invoice for two years will take 55 days on this one, and forecasting them at 30 is not optimism, it is being wrong in a known direction every single week. The forecast becomes less comfortable and far more useful. Tightening that behaviour is a separate exercise.
Does the system produce the forecast automatically?
No, and we would rather say so than imply a predictive capability we have not built. What it provides are the grounded inputs a forecast needs — reconciled bank position, receivables ageing with actual payment history, payables with due dates, commitments visible at approval, payroll and statutory liabilities — plus a report builder for your own forward views. The assembly and the judgement are yours or your accountant's.
What should we cut first when cash is tight?
Discretionary outflows — stock purchases, capital spend, anything without a fixed date — which is exactly why they belong in their own row in the forecast. Then accelerate specific named collections, then talk to suppliers before you are late rather than after. Salaries and statutory obligations should be the last things you move, not the first: penalties and losing people cost far more than deferring a stock order. Take these decisions with your accountant.
