At the end of the day you count the money, and it comes to roughly what a good day should. Cash in the drawer, M-Pesa on the phone, added together into one number that looks about right. It balances. You lock up and go home.
That number — the single one that "balances" — is the most dangerous habit in a Kenyan shop. Not because it's wrong, but because it can be short by two thousand shillings and still look right, and you will never know, because you blended the two things that should never touch.
The two rails money arrives on
In a Kenyan shop, money comes in on two completely different rails, and they fail in completely different ways.
Cash is physical. You can count it, you can make change with it, and you can lose it — to a genuine mistake, to a wrong-change moment, or to a hand in the drawer. Cash is the rail where money actually goes missing.
M-Pesa is a record. Every payment leaves a trace on the till or the phone. It is much harder to steal an M-Pesa payment — but it is easy for one to simply not get recorded as a sale, so the stock leaves and the money sits in a number nobody reconciled.
Different rails, different risks. The moment you add them into one total, you lose the ability to tell which rail a problem is on — and a problem you can't locate is a problem you can't fix.
A single total that balances doesn't mean nothing is wrong. It means you can no longer see what is.
Why "it balances" hides the leak
Say your true day was Ksh 18,000 — Ksh 10,000 cash, Ksh 8,000 M-Pesa. Now say Ksh 2,000 of cash quietly went missing, and Ksh 2,000 of M-Pesa came in that never got rung up as a sale. Your drawer is Ksh 2,000 short; your M-Pesa is Ksh 2,000 over. Add them together and you get… Ksh 18,000. It balances. Perfectly. And two separate problems just cancelled each other out in front of your eyes.
This is not a rare trick. It is the ordinary way a busy shop leaks — one rail short, the other fuzzy, the blended total covering for both. The owner sees a normal day and sleeps fine.
How to actually close the day
Closing properly is not more work. It is the same count, kept on two separate lines instead of one.
- Record the opening float. The change you started the day with. If you don't know what went in, you can't know what should come out.
- Count the cash on its own. Physically. Then compare it to what the cash should be: cash sales, plus the float, minus anything paid out of the drawer (a supplier, an errand, a cashout). Drawer versus expected — that gap, on its own, is real information.
- Check the M-Pesa total on its own. What the till or statement says came in, against what was rung up as M-Pesa sales. That gap, on its own, is different information.
- Two checks, two answers. A short drawer points to the cash rail. A mismatched M-Pesa points to unrecorded sales. Each gap tells you where to look. "It balances overall" tells you nothing and lets you look nowhere.
Do this at the close of every shift — not once a week, not "when it feels off." A gap you catch tonight has a name and a shift attached to it. A gap you notice at month-end is just a number, cold, with no trail.
Why it has to be per shift, and per person
Here is the part most owners miss. Even if you separate the rails, closing the whole day as one still hides who. Two, sometimes three people worked that counter. If the day is short, which shift? Which person? You can't say, so you say nothing, and nothing changes — until it happens again.
Close each shift as its own reconciliation, tied to the person who worked it. Not to accuse anyone — most gaps are honest mistakes — but because a gap you can locate to a shift and a name is a gap you can actually talk about and fix. A gap you can't locate just becomes the background cost of doing business, and it grows.
This is one of the specific jobs Stoka does. Every sale is recorded on the rail it came in on and against the staff member on shift. At close, Stoka shows expected cash versus counted cash, and M-Pesa on its own line — and flags the discrepancy, per shift, per name — so the drawer and the phone are each checked on their own. The owner doesn't have to be there counting; the books close themselves, honestly. It works the same whether you run one counter or two shops at once.
The habit that pays for itself
Separating your rails and closing every shift feels like a small, boring discipline. It is also the single cheapest theft-and-error control a Kenyan shop has. No cameras, no accusations, no drama — just two lines instead of one, checked every night, tied to a name.
The shop that closes on one blended number will keep having "about right" days and quietly bleeding. The shop that keeps the rails apart catches the two-thousand-shilling gap the night it happens — and after a while, stops having them at all, because everyone knows the books can see.