Why growing businesses outgrow spreadsheets — and how to know when it has happened

Almost every growing business runs on spreadsheets at some point, and for good reason: they are free, flexible, understood by everyone and available on any device. The problem this piece addresses is what happens next — the point at which the spreadsheet that built the business starts to cost it money, and the fact that this point is usually passed months before anyone notices.

What is happening?

The research on spreadsheet errors is older than most of the businesses using them, and it has not been overturned. Raymond Panko’s review of field audits of operational spreadsheets found that 94% of the 88 spreadsheets examined contained at least one error, with an average cell error rate of about 5% across the studies that measured it. The errors were not exotic: mistakes in logic, mechanical slips, and things simply left out. The more uncomfortable finding is about detection. People are highly accurate when they enter a cell — and poor at finding the errors that slipped through, because there is nothing in a spreadsheet that tells you a number is wrong.

What the research establishes is the error rate of spreadsheets built and used by people at work. What it does not establish is the cost to any particular business — that depends on what the sheet holds and who relies on it, which is the subject of the rest of this piece.

Why it matters economically

A spreadsheet has no idea what it contains. It does not know that a row is an order, that an order needs a customer, that a customer’s balance cannot be negative, or that the delivery date has passed. It holds whatever the last person typed. While one person owns it and the volume is small, that person’s attention is the control system. The business outgrows the spreadsheet at the exact moment that attention is no longer enough.

The cost is never the licence. It is four things that grow with the business:

• Concurrency. Two people cannot safely work in one sheet, so there are two sheets, then a weekly merge, then a “master” that is already wrong by Tuesday. The labour of keeping copies aligned is a hidden headcount.
• The single owner. One person knows how the sheet works. When they are on leave, ill, or gone, the operation slows or stops. Key-person risk is an economic cost even when nothing has gone wrong yet.
• Silent errors. A five per cent cell error rate in a stock sheet is stock you think you have. In a pricing sheet it is margin you think you made. Because the error is silent, its cost is discovered late — in a customer complaint, a stock-out, or the accountant’s question.
• No history. A spreadsheet holds the current state, not how it got there. Who changed the price, when the order was confirmed, why the balance moved: unanswerable. Every dispute becomes an investigation.

Take an illustrative distributor with four staff, 40 orders a day and stock in two locations, run on three spreadsheets. If the team spends a combined two hours a day reconciling copies and checking figures, that is a quarter of a person, permanently, on work the sheet creates. Add one bad month of stock errors and the cost of the “free” tool exceeds most alternatives.

What it means for African businesses

The pattern is the same everywhere; two things make it sharper in many African markets.

First, the spreadsheet is usually the only system. Where a business has grown through WhatsApp, cash and mobile money, the spreadsheet is not one tool among several — it is the CRM, the order book, the stock record and the ledger at once. When it fails, everything fails together.

Second, growth tends to cross boundaries: a second city, a second currency, a second country. A spreadsheet that just about coped with one location is the wrong shape for two, and the merge problem becomes a management problem overnight.

On the other side of the ledger, the same conditions make the alternative more valuable than it would be elsewhere. A system that holds one record of the customer and the order across messaging, payments and locations replaces not one spreadsheet but the whole improvised stack. This applies across markets; it is not specific to any one country.

Technology and AI implications

The replacement for a business-critical spreadsheet is a record with rules: a CRM or business system that knows what an order is, who owns it, what state it is in and what should happen next — with a history of every change. For most growing businesses that is the whole answer, and it should be shaped to the business rather than chosen from a catalogue.

Where the operation is unusual — field teams collecting data offline, a marketplace, a workflow no product handles — the answer is a custom web or mobile application. Where the spreadsheet’s real job was reminding someone to do something, the answer may be automation on top of the record, not a new interface at all. AI is largely irrelevant to this decision: the value is in the structure of the record, not in intelligence applied to it.

Implementation requires the business to define its records before any software is chosen — what an order is, its states, who may change it — and then to migrate the spreadsheet’s history rather than abandon it. What technology does not fix is a business that has not made those definitions. A system built on top of an undefined process is a slower spreadsheet.

Practical business implications

The point of outgrowing a spreadsheet is visible before the damage. If your business shows two or more of these, it has passed it:

1. More than one person needs to edit the same data on the same day.
2. There is a “master” copy and at least one other copy that matters.
3. A number in the sheet is one the business must get right — stock, balances owed, prices, payroll.
4. Someone spends time every week reconciling sheets with each other or with reality.
5. A customer, an order or a payment has been lost because it was in the wrong tab, or the wrong file.
6. The business is opening a second location, product line or market.

If none of these applies — a single owner, a manageable volume, one place — keep the spreadsheet. Give it discipline instead: one file, named columns, a backup, and a rule that money and stock are never edited from memory.

If several apply, do not begin by choosing software. Begin by writing down what the sheet actually is: which records it holds, who touches them, which numbers must be right, and what happens to each record from creation to close. That document is the specification of the system you need — and it is the first stage of how we work.

RevenueStack perspective

Our view: the spreadsheet is never the problem. It is the most honest tool a business owns, because it does exactly what it is told and nothing else. The problem is the moment when “exactly what it is told” is no longer enough — when the business needs its records to know things, enforce things and remember things on their own.

Through strategic economics, that moment is a sign of success and should be treated as one: the point at which the business is worth building infrastructure for. Through technology, the answer is almost always a record with rules, not a bigger sheet or a bought product with the wrong shape. Through AI — honestly — very little, until the record exists for it to work on.

Conclusion

The businesses that recognise the moment early spend a few months on a system. The ones that do not spend those months, every year, on the reconciliation. Count the copies, count the people editing, name the number that must be right — and the decision usually makes itself.

Sources

1. What We Know About Spreadsheet Errors — Raymond R. Panko, University of Hawai‘i, 2008
2. Spreadsheet Errors: What We Know. What We Think We Can Do. — arXiv (Proceedings of EuSpRIG), 2008-02