When to replace spreadsheets with a custom tool

Replace a spreadsheet with a custom tool when it has quietly become the system your business runs on: several people edit it every day, customers or money depend on what it says, and mistakes keep slipping through because nothing stops them. Keep the spreadsheet when one person owns it, when it is used for analysis rather than day-to-day operations, or when the process behind it still changes every month. Most spreadsheets should stay spreadsheets. The few that run a core process are worth replacing, and they tend to show the same symptoms.

We build these replacements as part of our client portals and internal tools work, and a fair share of first conversations end with advice to keep the spreadsheet a while longer. The checks below are the ones we use to decide.

What are the signs a spreadsheet has become a system?

Size is rarely the problem. As of October 2026, Excel allows 1,048,576 rows by 16,384 columns per worksheet, according to Microsoft's specifications and limits, and Google Sheets allows up to 20 million cells per file under Google Drive's file limits. Few business sheets get near either. They fail on people, not rows.

These are the signs we look for:

• Several people edit it at once, and someone asks "which version is right?" every week.
• It is the only record of money or customers. Orders, job status, stock, commissions or renewals live there and nowhere else.
• Data is copied in and out by hand. Someone pastes rows from the website, CRM or bank statement into it, then copies results back out.
• One person understands the formulas. When they are on leave, the sheet is frozen or quietly broken.
• You need to limit who sees what. A client should see only their rows, a salesperson only their deals. Spreadsheets share files, not records.
• Nobody can answer "who changed this, and why?" Version history exists, but finding one edit in it takes an afternoon.
• Errors have cost real money. A formula dragged one row short, a sort that broke a lookup, a deleted row nobody noticed for a month.

Two or three of these are normal for any growing team. Five or more, on a sheet that runs a core process, is usually the point where a proper tool is cheaper than the mistakes.

When should you keep the spreadsheet?

More often than a software studio would like to admit.

It is for thinking, not operating. Budgets, pricing models, what-if scenarios and one-off analyses belong in a spreadsheet. Nothing beats it for working with numbers.

The process is still changing. If you changed the columns three times this quarter, wait. Software fixes a process in place. Build once the process has settled, or you pay to rebuild it.

The team is small and the stakes are low. Three people, a shared Google Sheet with data validation, protected ranges and an input form will hold for a long time and costs nothing.

A product already does the job. If the sheet is really a CRM, buy a CRM. If it is stock for an online shop, use the store's inventory. Our guide to custom software vs off-the-shelf covers that decision in detail.

Before replacing anything, tidy it: one tab per kind of record, dropdowns instead of free text, protected formula columns and a form for new entries. That alone buys many teams another year.

What are the options between a spreadsheet and custom software?

Move one column to the right only when the column you are in has clearly failed. A no-code database is a sensible step for many teams, and an off-the-shelf product beats a build for anything standard.

What does a custom replacement involve?

The spreadsheet's tabs become tables with proper relationships. The customer column with fourteen spellings of the same company becomes one customer record that orders point to. Then, around that data:

• Forms with rules. You cannot save an order without a customer, a delivery date in the past or a negative quantity.
• Roles. Staff, managers and, if needed, clients each see and change only what they should. Roles and permissions covers how to design this.
• History. Every change is recorded with who made it and when.
• Views and reports that replace the daily filter-and-pivot routine. How to build an internal dashboard for operations shows what a useful first version looks like.
• Connections to the website, accounting or WhatsApp, so rows stop being pasted by hand. API integration explained describes how that works.
• Export. One click back to CSV or Excel, because finance will still want a spreadsheet, and that is fine.

A focused tool of this kind takes eight to fourteen weeks as of October 2026. Most of the risk is in scope: a tool meant to replace one sheet grows to replace five. Replace the one that hurts most, live with the result for a month, then decide on the next.

How do you move off the spreadsheet without disrupting work?

1. Freeze the structure. No new columns while the tool is being built. Changes go on a list for version two.
2. Clean the data before importing it. Export each tab to CSV, remove duplicates and fix the obvious mess. Our CSV to JSON converter is handy for checking how an export will look to an import script. Data migration from old systems is the full checklist.
3. Run both for two to four weeks. Work happens in the new tool. The spreadsheet becomes read-only and is used only to check figures.
4. Retire the sheet. Archive a final copy, remove edit access and tell everyone where the truth now lives.

To decide whether it is worth doing at all, put a number on the manual work. The automation hours-saved calculator turns minutes per task, frequency and number of people into hours and cost per year. Compare that with the build cost over three years, and add the cost of the last two expensive mistakes.

A good first step this week: list every spreadsheet that more than two people edit, and mark the ones that touch money or customers. If one of them shows five of the signs above, that is the one to look at.