What Your Spreadsheets Actually Cost You

The Five Lines

1. Assembly Hours

Start with the obvious one, but measure the whole thing — not the time spent writing the report, the time spent getting to the point where the report can be written.

For one recurring report, count:

• Time pulling and exporting source data
• Time chasing people who haven't submitted their inputs
• Time reconciling figures that don't agree between sources
• Time formatting and distributing
• Time spent by the other people who supply inputs, not just the person who assembles

Multiply by frequency, multiply by fully loaded hourly cost — salary plus employer contributions plus overhead, not the take-home rate.

Most managers underestimate this by half, and almost always for the same reason: the chasing and reconciling doesn't feel like work on the report. It feels like admin. It is the report.

2. Rework and Error Correction

Every operational spreadsheet estate carries an error rate. The question is not whether yours has errors — it is how long they survive before someone catches them, and what happens in the meantime.

Count the incidents you can remember from the last twelve months where a number was wrong and it mattered: a figure restated after distribution, a decision made on bad data, a compliance submission corrected, an invoice queried. For each, estimate the hours spent correcting it and the downstream consequence.

The uncomfortable part of this line item is that it is systematically undercounted, because errors that were never caught cannot be counted. Whatever number you arrive at is a floor.

3. Decision Latency

This is the largest cost on most sites and the one almost nobody measures.

Ask a single question about each report: how much time passes between something happening and somebody being able to see that it happened?

If a process starts drifting on the Tuesday night shift and the report that would reveal it lands on the seventh of the following month, you did not lose a report. You lost five weeks of production running in a state you would have corrected on the Wednesday morning.

The cost here is not administrative. It is the value of the decision you could not make in time, multiplied by how often you couldn't make it. On a continuous operation that arithmetic gets ugly quickly, and it dwarfs the assembly hours in line 1 — which is why fixing the wrong report is such a common and expensive mistake. The most annoying report is rarely the most expensive one.

4. Key-Person Concentration

Identify, for each critical workbook, the people who could rebuild it from scratch if it were lost tomorrow.

If the answer is one person, you do not have a reporting process. You have an arrangement with an individual. That arrangement ends when they resign, retire, go on extended leave, or get promoted into a role where they no longer have time for it.

You can price this directly: what would it cost to reconstruct that workbook and re-establish trust in its outputs, under time pressure, without the person who built it? Include the period during which the business is running on numbers nobody fully trusts.

This is also the line item that turns from theoretical to urgent with no warning at all.

5. The Ceiling

Manual reporting scales linearly with volume. Automated reporting does not.

So the last question is forward-looking: what does this cost when you add the next site, the next contract, the next customer, or the next compliance requirement? If the answer is "another person" or "everyone works later," you are not looking at a cost — you are looking at a constraint on what the business is allowed to become.

A great deal of manual reporting quietly caps growth long before anyone identifies it as the reason.

Running the Numbers

Take your single most important recurring report and fill this in.

Two rules for doing this honestly.

Use loaded cost, not salary. A R45,000-a-month analyst does not cost R45,000 a month.

Do the whole estate, not one workbook. Most operations have somewhere between four and twenty of these. The individual numbers look tolerable. The total usually does not.

When a Spreadsheet Is the Right Answer

It would be convenient for us to argue that every spreadsheet should be replaced. It shouldn't, and pretending otherwise would waste your money.

Keep the spreadsheet when:

• The logic changes frequently. If the calculation is still being argued about, freezing it into a system is premature. Spreadsheets are excellent for thinking.
• The volume is low and the frequency is low. A quarterly report that takes two hours does not justify a project.
• It is exploratory. One-off analysis, modelling, scenario work — this is what the tool is for.
• It is genuinely temporary. Though be honest about this one. Almost nothing is.

Replace it when the logic has been stable for a year or more, the frequency is high, the latency is costing you decisions, and one person holds the knowledge. That combination is the actual signal — not the spreadsheet itself.

Where to Start

Not with the biggest system. With one report.

Pick the report with the highest frequency × latency cost — the one produced most often where the delay hurts most. Not the one people complain about most; those are usually different reports. Automating one high-value report end to end gives you a real number for what the change is worth, and that number is what justifies the next one.

The organisations that get this wrong start with a platform. The ones that get it right start with a report, prove the saving, and let the second project be funded by the first. This is the same discipline we set out in Automating Process Bottlenecks: start small, prove it, then scale.

Doing This With Us

We build the systems that replace this kind of reporting — for industrial operations, logistics and service businesses in Richards Bay, the Zululand corridor and further afield.

If you want the exercise above done properly, our discovery engagement is a fixed-fee two-week assessment: we map your reporting estate, cost it using this framework, and give you a prioritised list of what to fix in what order. You own the output whether or not you build anything with us.