Spreadsheets do not break, they get expensive
Nobody wakes up to find the spreadsheets have stopped working. The transition happens differently: one sheet for stock, then another for orders, then a third that reconciles the first two. A year later there are twelve files, three of them are "the right one," and nobody remembers which.
Excel is an excellent tool, and most companies genuinely need nothing else up to a certain size. The problem is not Excel — it is that the boundary of its usefulness is crossed without anyone noticing. The cost rises gradually, so there is no event that forces a decision.
This piece is about spotting the boundary before it costs money. If you have already decided a system is needed, the more useful starting point is how to choose an implementation partner.
Seven symptoms that appear before the problem does
Individually none of them means much. Three or more at once and the spreadsheets already cost more than a system would.
1. "What is our stock?" requires the follow-up "according to which file?"
The most reliable single indicator. If a simple question cannot be answered without establishing whose version is newest, the data is no longer singular — and copies multiply faster than people do.
2. Somebody spends a day a month assembling a report by hand
Cost that day out. It usually comes to several months of a subscription across a year, and that is before counting the errors that manual consolidation guarantees.
3. You have sold something you do not have
Classic for any company with two sales channels. The salesperson reads a stock figure updated once a day; the warehouse shipped it yesterday. Once a month is annoying; once a week is reputation.
4. Cost of goods is "roughly"
If freight, duty and currency movement are not in your unit cost, you do not know your margin — you know your markup. On imported goods the gap between them runs into tens of percent, and companies operate at a loss on individual lines for years without noticing.
5. New joiners are trained on the files, not the job
When onboarding is two weeks of explaining which spreadsheet does what and why that column must never be edited, the knowledge lives in people rather than in the process. One resignation becomes an operational risk.
6. History is unreachable
"What did we charge this customer in the spring?" means digging through last year's files. The data technically exists but cannot be used, which means for practical purposes it does not exist.
7. The file takes longer to open than you can hold the reason for opening it
The most mundane symptom and the most common. Excel starts to struggle somewhere past a few tens of thousands of rows with formulas — precisely when the data has finally become valuable.
What the free option actually costs
Spreadsheets feel free because their cost never arrives as an invoice. It is spread across four lines, and each one can be estimated.
| Cost | How to estimate it | Typical range |
|---|---|---|
| Manual consolidation | Hours per month × loaded hourly cost | 8–24 hours every month |
| Data errors | Cost of one error × frequency | From a mispick to a lost customer |
| Capital tied up in stock | Excess inventory × your cost of money | 10–30 % of stock nobody needs |
| Decisions not made | Never measured, costs the most | Decisions wait "until we have the numbers" |
The fourth line is the important one and the least obvious. When figures take two days to assemble, managers stop asking for them and start running on instinct. That is not an accounting problem; it is a decision-quality problem.
Work it out on your own numbers
A comparison that takes five minutes and gives an honest answer.
Step 1. How many hours a month go into manual consolidation, reconciling files and chasing discrepancies? Count everyone: finance, warehouse, sales, yourself.
Step 2. Multiply by loaded hourly cost. For a company of fifteen people this usually lands between 20 and 60 hours a month.
Step 3. Compare against the monthly cost of a system. Business Central licences run $80–110 per user per month, and not everyone needs a full licence — part of the team works on cheaper roles. Model it for your own headcount in the cost calculator.
If the first number exceeds the second, the question is settled and you are already overpaying. If it does not, spreadsheets are still sufficient — and that is a legitimate answer.
What to do if it is still too early
Roughly half the companies that approach us do not need a system yet. That is not a reason to do nothing: there are things worth doing now that will make the eventual move cheaper.
- Tidy the master data. One product, one row, one code, one name. Duplicates and inconsistent naming are the single biggest driver of migration cost — see the data preparation checklist.
- Declare one file the single source of truth. Not ideal, but it is the discipline without which no system works either.
- Write your processes down in plain text. Who does what after an order arrives. This document is needed when selecting a system and saves weeks during implementation.
- Do not migrate to a second spreadsheet. The worst outcome is replacing Excel with an elaborate scripted construction in Google Sheets. It survives exactly until the person who built it leaves.
Three mistakes when moving
Migrating the mess. If the spreadsheets are chaotic, the system will be chaotic too, only more expensive. Cleaning happens before migration.
Automating everything at once. A fifteen-person company should not launch five modules in month one. Start where it hurts most — usually stock or finance — and add the rest once the first part works.
Keeping Excel "just in case." Running both beyond one reporting period reliably ends with people drifting back to the familiar option. The switch-off date belongs in the plan from the start. More on how projects go wrong in why ERP implementations fail.
Frequently asked questions
At what revenue does Excel stop working?
Revenue is a poor guide. A small company with three warehouses and imports hits the limit before a trader with one location and ten times the turnover. Look instead at the number of stock locations, sales channels and people editing the same data simultaneously.
Can we manage with accounting software instead of an ERP?
If your requirement is filing accounts, yes — and it does that well. What it does not answer is how much to order next week or which customer is actually profitable. That is exactly the point at which companies start bolting spreadsheets onto the side.
How long does moving from spreadsheets take?
Four to eight weeks for a small company with clean data. Most of that time goes into tidying master data and documenting processes rather than the technical work. Poor data quality can make preparation longer than the configuration itself.
Do we have to abandon Excel completely?
No, and you should not. Excel remains excellent for one-off analysis: export, calculate, conclude. The problem is only when it holds primary data that changes daily.
What if the team resists?
Resistance almost always means nobody explained what changes for them specifically. A warehouse supervisor does not care about consolidated reporting; they care whether there is less manual work. Have the conversation in those terms, and train on real company data rather than a demo company.