Where Did the Cash Go? A Plain-English Guide to Cash Management for Retail SMEs

You sold well today. The shop was busy, the card machine kept beeping, the drawer felt full. Then you count it at close and you're ₦4,200 short. Not catastrophic. But you can't say where it went — and that's the part that should worry you.

This is the most common money problem in retail, and it's almost never one big theft. It's a slow leak: a few naira here, an unrecorded payout there, a float nobody reset. By the time you notice, the trail is cold.

This guide is about closing that leak. We'll keep it plain, define every term as we go, and show you a routine you can run by hand tonight — and what changes when a system runs it for you.

Why cash leaks (and why you rarely catch it in the moment)

Cash is the only thing in your shop that leaves no trace when it goes missing. A bottle off the shelf shows up as a stock gap. A bank transfer shows up in the statement. Cash just... isn't there. So the leaks tend to be quiet ones:

Miscounts. Someone counts ₦47,000 as ₦47,500 at handover. Now every number after that is wrong, and nobody did anything dishonest. Counting under pressure, with customers waiting, is genuinely hard.

No record of what came in. If a sale gets rung up wrong — or not rung up at all — the drawer holds money your books don't know about, or your books expect money the drawer never got. Either way, the two don't match.

Unrecorded payouts. This is the big one. Your shop attendant pays the bread supplier ₦8,000 from the till because that's what was on hand. Totally normal. But if nobody writes it down, the drawer is now ₦8,000 "short" against the sales — except it isn't short at all. The money did a job. It just left no note.

Float confusion. The float (also called the opening cash) is the money you put in the drawer at the start of the day so you can give change — say ₦10,000 in small notes and coins. If you don't separate the float from the day's takings, you'll either count it as profit (and overstate the day) or forget it entirely (and think you're short). Most float problems are really "we never agreed what the float is."

Notice that three of those four are not theft. They're sloppiness — honest, ordinary, daily sloppiness. And that's exactly why a leak is so hard to read: when the drawer is short, you can't tell an innocent miscount from a hand in the till. A routine fixes the sloppiness so that when something is genuinely wrong, it stands out.

The three words that make cash manageable

You only need three pieces of vocabulary. Learn these and the rest of this guide is easy.

Register session

A register session is one continuous period that a drawer is "open" — usually one shift or one trading day. It has a clear start (you put the float in and open it) and a clear end (you count it and close it). Everything that happens to the cash — sales, refunds, payouts — belongs to that one session.

Why it matters: without sessions, a single drawer absorbs days of activity and you can never isolate when a problem started. With sessions, a bad day is a bad session — a small, contained thing you can investigate.

Close-out

The close-out is the act of ending a session: you count the physical cash, write down the number, and compare it to what the session should hold. "Should hold" is simple arithmetic:

Opening float + cash sales − cash refunds − recorded cash payouts = expected cash

You count the drawer. That's your actual cash. Now you have two numbers to compare.

Variance

Variance is the gap between expected and actual:

Variance = actual cash − expected cash

If you expected ₦88,000 and counted ₦88,000, variance is zero — a clean close. Count ₦87,200 and you're ₦800 short. Count ₦88,500 and you're ₦500 over (which is also a problem — it usually means a sale or payout wasn't recorded).

Variance is the single most useful number in cash management. It turns a vague worry ("does the cash look right?") into a fact you can act on.

Setting up a daily close routine

Here's a routine any shop, kitchen, or boutique can run starting tonight. It takes five to ten minutes.

1. Agree a fixed opening float. Pick a number — say ₦10,000 — and make it the float every single day. Count it into the drawer at open. The float never changes day to day; only the takings on top of it do. A fixed float removes a whole category of confusion.

2. Write down every payout, every time. If cash leaves the drawer for anything that isn't a refund — paying a supplier, buying diesel, settling a small bill — it gets noted: amount, what for, who authorised it. A scrap of paper in the drawer is enough to start. This one habit eliminates the most common "shortfall" there is.

3. Count at close, then compare. At end of session, count the cash. Subtract the float to see the day's takings, then check actual against expected. Now you have your variance.

4. Set a variance threshold. Decide what gap is "normal noise" versus "look into this." For a small shop, something like ±₦500, or under 1% of the day's cash sales, is a sensible starting line. Under the line, log it and move on. Over the line, you investigate before you go home — while memories are fresh and the day is still reconstructable.

The threshold is the whole point of the routine. It tells you when to stop counting and when to start asking questions — so you're neither ignoring real problems nor losing sleep over ₦50.

The weekly reconciliation habit

Daily close catches today. Reconciliation — checking your records against an independent source of truth — catches the patterns daily close can't see.

Once a week, sit down and line up the week's sessions against your bank deposits and card-machine settlements. The questions you're asking:

• Did the cash we counted actually get banked, in full?
• Do the card-machine totals match what our sales records say we took on card?
• Are the daily variances random — sometimes over, sometimes under, all small — or do they lean one way?

That last question is the important one. Random small variances are life. A variance that's consistently short, especially on the same shift or the same person's days, is a pattern. Patterns are what you're hunting. One short day tells you nothing; four short Tuesdays tell you a lot.

When variance means theft — and when it just means sloppiness

This is the question everyone's really asking, so let's be direct.

It's probably sloppiness when: the variance is small, it goes both directions (some days over, some under), it shrinks once you tighten the payout-logging habit, and it's spread across everyone. Honest mistakes are noisy and random.

It's worth a hard look when: the variance is consistently in one direction (almost always short), it clusters around a specific person, shift, or till, it's accompanied by unusual voids or discounts or "no-sale" drawer opens, and it doesn't improve when the routine tightens. Theft is quiet, one-directional, and patterned.

A rule of thumb: don't accuse a number, investigate a pattern. A single short drawer is a counting error until proven otherwise. A trend is a conversation. The routine above is what turns isolated shortfalls into a trend you can actually see — which protects honest staff as much as it catches dishonest ones.

Doing this by hand vs. letting a system do it

Everything above works on paper. Plenty of good operators have run tight cash for years with a notebook and discipline. The cost isn't that paper can't work — it's what paper asks of you:

• You do the expected-cash arithmetic by hand, every session, and arithmetic done tired is arithmetic done wrong.
• Payouts only get logged if someone remembers to log them — and the one that's forgotten is the one that breaks your numbers.
• Variance only gets calculated if you sit and calculate it. Skip a busy night and the trail goes cold.
• Spotting a pattern means manually comparing weeks of scribbled close-outs — which almost nobody actually does.

In other words, the by-hand version depends entirely on the most tired, busiest person in the building doing fiddly maths correctly at 10pm. It works right up until the night it doesn't.

How Eleo handles this

Eleo runs the whole routine above as part of how the till works — not as extra admin on top of it.

Register sessions and close-outs are built in. A drawer opens with a recorded float and closes with a count; everything in between — sales, refunds, cash payouts — is attached to that session automatically, so each day is a clean, contained record.

Variance flagging is automatic. At close, Eleo already knows the expected cash (it's been tracking the session in real time), so the moment you enter your count it shows the variance and flags anything past the threshold you set — no arithmetic, no forgetting.

Cash confidence scoring looks across sessions for you. Instead of you manually hunting for patterns, Eleo reads the trend — direction, consistency, where variances cluster — and tells you whether your cash is healthy or whether something deserves a closer look. That's the weekly-reconciliation instinct, running continuously.

Underneath all of it is a double-entry ledger — the same accounting backbone banks use, where every figure has a matching source and nothing moves without both sides being recorded. That's why a payout logged at the till already shows up correctly in your books: the cash drawer and the accounts are never telling two different stories.

The result is the thing the notebook can't reliably give you: at the end of a long day, you know exactly where your cash is — and if it's off, you know before you've gone home.

Cash you can't account for is money already half-gone. Close the leak before it becomes a hole.

See how Eleo handles your cash →